PDF (1966) - Sugar Industry Collection

Proceedings of the
FORTIETH
ANNUAL CONGRESS
of the
South African Sugar
Technologists'
Association
HELD AT MOUNT EDGECOMBE
21 st-25th MARCH, 1966
The Copyright of these papers is the property of the Association
The Association does not hold itself responsible for any of the
opinions expressed in papers published herein
Published by
South African Sugar Technologists'
Association
SOUTH AFRICAN SUGAR ASSOCIATION EXPERIMENT STATION
MOUNT EDGECOMBE, NATAL
PRICE FOR EXTRA COPIES R5.00
Printed in the Republic of South Africa by Brown Davis and Piatt Ltd
Proc. Annual Cong. S. Afr. Sugar Tech. Ass. No. 40, pp. 1-351, Durban, 1966

Proceedings',*'The South African Sugar Technologists' Association—March 1966
THE SOUTH AFRICAN
SUGAR TECHNOLOGISTS'
ASSOCIATION
The South African Sugar Technologists' Association was
founded in 1926. It is an organisation of technical
workers and others directly interested in the technical
aspect of the South African Sugar Industry. It operates
under the aegis of the South African Sugar Association,
but is governed under its own constitution by a Council
elected by its members.
The office of the Association is situated on premises
kindly made available to it by the South African Sugar
Association at the latter's Experiment Station at Mount
Edgecombe.

THE SOUTH AFRICAN SUGAR TECHNOLOGISTS' ASSOCIATION
OFFICERS OF THE SOUTH AFRICAN SUGAR TECHNOLOGISTS'
ASSOCIATION
LIST OF MEMBERS AND GUESTS
OPENING ADDRESS, by DR. A. J. A. Roux, Director General,
Atomic Energy Board
REPLY by MR. J. P. N. BENTLEY
PRESIDENTIAL ADDRESS
REPLY bv MR. L. F. CHIAZZARI
FORTY-FIRST ANNUAL SUMMARY OF LABORATORY REPORTS
by C. G. M. PERK
BOILER DESIGN AND SELECTION IN THE CANE SUGAR
INDUSTRY by N. MAGASINER
FUELS AND FURNACES by P. R. A. GLENNIE . . . .
STEAM AND VAPOUR DISTRIBUTION by R. E. MARSH .
CONDITIONING BOILER FEEDWATER FOR THE SUGAR MILL
by G. E. ANGUS
ECONOMIC DESIGN AND OPERATION OF PROCESS HEAT
EXCHANGE EQUIPMENT by E. J. BUCHANAN . .
PROCESS STEAM PRODUCTION USE AND CONTROL by D. T.
O. GRIFFITH
BOILER OPERATION, MAINTENANCE AND TESTING by S. G.
HOLTON
STEAM TURBINES — THEIR CONSTRUCTION, SELECTION AND
OPERATION by W. B. JACHENS
STEAM TRAPPING AND CONDENSATE CONSERVATION by
J. M. CARGILL
RESIDUAL FUEL OIL AS A SUPPLEMENTARY FUEL by J.
GUDMANZ
STEAM ECONOMIES AT DARNALL by D. J. L. HULETT .
How TO MEASURE AND EXPRESS SUGAR MILLS' EFFICIEN­
CIES bv T. H. FOURMOND
ALTERATIONS AND IMPROVEMENTS TO MOUNT EDGECOMBE
MILLING TANDEM by R. C TURNER
THE ELECTRICAL SUPPLY SYSTEM OF SUGAR FACTORIES by
A. GRADENER
PHOSPHORIC ACID AS AN AID TO CLARIFICATION, AND
OBSERVATIONS ON LIMING TECHNIQUES AND MUD
VOLUMES by G. G. CARTER
VACUUM PAN CONTROL — PROGRESS REPORT NO. 1 bv
D. H. JONES and D. E. WARNE
IMPROVEMENTS IN RAW SUGAR QUALITY by R. P. JENNINGS
THE QUALITY OF IMPORTED RAW SUGARS by R. P. JENNINGS
and RITA VAN KEPPEL
A MODIFIED METHOD FOR DETERMINING FILTERABILITY
by R. P. JENNINGS
SOME EFFECTS OF BORAX ON THE POLARISATION OF SUGAR
SOLUTIONS by D. ADAM and R. P. JENNINGS
Proceedings of The South African Sugar Technologists' Association" March 1966
CONTENTS
29
64
69
79
89
102
108
113
132
135
142
149
152
162
171
181
192
196
199
206
REFINED SUGAR CONDITIONING AND STORAGE by A - M-
HOWES . . 214
WEATHER REPORT FOR THE YEAR 1ST JUNE 1965 'i'° 31ST
MAY 1966 by K. E. F. ALEXANDER . . . . . . 220
AUTOMATION OF A ROUTINE LABORATORY bv R. T'- BISHOP
and T. D. ALGIE '. . . . . 226
DETERMINATION OF COPPER AND ZINC IN SLK>ARCANE
LEAVES BY ATOMIC ABSORPTION by P. DU PREEZ . . 234
NITRIFICATION IN RELATION TO pH — ITS IMPORTANCE: IN
FERTILIZER NITROGEN UTILIZATION BY CANE IN SOME
SUGAR BELT SOILS by R. A. WOOD . . . . . . 241
ABNORMAL NITROGEN REQUIREMENTS OF SUGAR CANE: ON
THE MONTMORILLONITIC BLACK CLAY SOIL OK THE
KAFUE FLATS IN ZAMBIA bv D. S. HUGHAN iind D.
R. C BOOTH . . . 247
THE INFLUENCE OF TRASH ON NITROGEN MINERALISATION
— IMMOBILIZATION RELATIONSHIPS IN SUGAR BELT
SOILS by R. A. WOOD 253
SOME CHARACTERISTICS OF THE SOILS OF THE Sui« ARCANE
GROWING AREAS AROUND MALELANE-KOMATIPOORT,
EASTERN TRANSVAAL by R. R. MAUD and E. A. VON
DER MEDEN 263
NOTE ON SALINITY LIMITS FOR SUGARCANE IN NATAL by
E. A. VON DER MEDEN 273
AVAILABILITY OF SOIL WATER TO SUGARCANE IN NATAL
by J. N. S. HILL 276
THE USE OF PERFORATED PIPES FOR IRRIGATION EXPERI­
MENTS by P. J. M. DE ROBILLARD and M. J. STEWART 283
SOME FACTORS AFFECTING THE TRANSLOCATION OF RADIO­
ACTIVE PARAQUAT IN CYPERUS SPECIES by G. H. WOOD
and J. M. GOSNELL '. . . . 286
SUMMARY OF AGRICULTURAL DATA: SUGARCANE CROP
1965 by J. L. DU TOIT and M. G. MURDOCH . . 293
SOME FACTORS INFLUENCING SUCROSE % CANE AT HIPPO
VALLEY ESTATES by C. A. JOHNSON 299
THE RESULTS OF HERBICIDE SCREENING TRIALS IN SUGAR­
CANE DURING 1965 by J. M. GOSNELL and G. D.
THOMPSON 304
WEED CONTROL ON A NEWLY DEVELOPING ESTATE AT
MAZABUKA, ZAMBIA by D. S. HUGHAN and D. R. C
BOOTH . [ 312
PROBLEMS IN WEED CONTROL ON AN ESTATE hy E. C
GILFILLAN 315
PEST CONTROL PROBLEMS AT MAZABUKA, ZAMBIA hy D. S.
HUGHAN and D. R. C. BOOTH 317
THE PROGRESS OF AN UNTREATED OUTBREAK OF A/mnicia
Viridis, MUIR by A. J. M. CARNEGIE 319
THE SUGARCANE NEMATODE PROBLEM hy J. DICK . . . 328
THE PRODUCTION OF TRASH AND ITS EFFECTS AS A MULCH
ON THE SOIL AND ON SUGARCANE NUTRITION h v G. D.
THOMPSON ". , . 333
STUDIES OF THE EFFECT ON SUGARCANE OF DAMAGI;
CAUSED BY FROST AND ASSOCIATED MICRO-ORXJ ANISMS
bv G. ROTH
INSTRUCTIONS TO AUTHORS
343
351

Proceedings of The South African Sugar Technologists' Association—March 1966
L.
1926-27
1927-28
1928-29
1929-30
1930-31
1931-32
1932-33
1933-34
1934-35
4935-36
1936-37
1937-38
1938-39
1926-27
1927-28
1928-29
1929-30
1930-31
1931-32
1932-33
President
F. CH1AZZARI
Hon-Secretary
(Mrs.) M. WELLS
ML MCMASTER
M. McMASTER
H. H. DODDS
H. H. DODDS
G. S. MOBERLY
G. C. DYMOND
G. C. DYMOND
B. E. D. PEARCE
E. CAMDEN-SMITH
G. C. WILSON
G. C. WILSON
J. RAULT
P. MURRAY
L. E. ROUILLARD
H. H. DODDS
G. S. MOBERLY
G. S. MOBERLY
G. C. DYMOND
A. C. WATSON
A. C. WATSON
icm 14/G. C. DYMOND
I^JJ J4 \E. CAMDEN-SMITH
1934-35
1935-36
1936-37
1937-38
B. E. D. PEARCE
E. CAMDEN-SMITH
J. RAULT
P. MURRAY
J. B. ALEXANDER
W. J. G. BARNES
G. S. BARTLETT
J. P. N. BENTLEY
1939-40
1940-41
1941-42
1942-43
1943-44
1944-45
1945-46
1946-47
1947-48
1948-49
1949-50
1950-51
1951-52
1938-39
1939-40
1940-41
1941-42
1942-43
1943-44
1944-45
1945-46
1946-47
1947-48
1948-49
1949-50
1950-51
1951-52
OFFICERS
1966-1967
Life Patron
DR. H. H. DODDS
F ormer Presidents
P. MURRAY
E. P. HEDLF.Y
F. W. HAYES
A. MCMARTIN
G. BOOTH
G. S. MOBERLY
G. S. MOBERLY
W. BUCHANAN
W. BUCHANAN
J. L. DU TOIT
H. H. DODDS
A. MCMARTIN
G. C. DYMOND
Former Vice-Presidents
E. P. HEDLEY
E. P. HEDLEY
F. W. HAYES
A. MCMARTIN
G. BOOTH
F. B. MACBETH
G. BOOTH
W. BUCHANAN
G. C DYMOND
G. C. DYMOND
G. C DYMOND
J. L. DU TOIT
O. W. M. PEARCE
O. W. M. PEARCE
Council of the Association
L. F. CHIAZZARI K. DOUWES-DEKKER
T. G. CLEASBY J. L. DU TOIT
D. J. COLLINGWOOD J. R. GUNN
J. DICK D. J. L. HULETT
Vice-President
T. G. CLEASBY
Hon. Technical Secretary
D. J. COLL1NGWOOD
1952-53
1953-54
1954-55
1955-56
1956-57
1957-58
1958-59
1959-60
1960-61
1961-62
1962-63
1963-64
1964-65
1965-66
G. C. DYMOND
G. C. DYMOND
G. C. DYMOND
J. B. GRANT
J. B. GRANT
J. P. N. BENTLEY
J. P. N. BENTLEY
J. P. N. BENTLEY
J. L. DU TOIT
J. L. Du TOIT
J. L. Du TOIT
J. R. GUNN
J. R. GUNN
J. R. GUNN
1952-53 K. DOUWES-DEKKER
1953-54 J. B. GRANT
1954-55 K. DOUWES-DEKKER
,/G.C. DYMOND
iy:o-oo^w_ G GALBRAITH
1956-57 W. G. GALBRAITH
1957-58 J. L. DU TOIT
1958-59 J. L. DU TOIT
1959-60 J. L. DU TOIT
1960-61 J. DICK
1961-62 J. P. N. BENTLEY
1962-63 J. P. N. BENTLEY
1963-64 L. F. CHIAZZARI
1964-65 L. F. CHIAZZARI
1965-66 L. F. CHIAZZARI
T. R. LOUDON
G. D. THOMPSON
J. WILSON

Proceedings of The South African Sugar Technologists' Association—March 1966
South African Sugar Technologists' Association
ADLER, E.
ALEXANDER, J. B.
ALEXANDER, K. E. F.
ALGIE, T. D.
ALLAN, G. N.
ALLSOPP, L.
ANNETTS, J. R.
ARCHIBALD, T. D.
AUCOCK, H. W.
BA LEY, E. D.
BA INES, A. C.
BARNES, W. J. G.
BARTLETT, G. S.
BEATER, B. E.
BENTLEY, J. P. N.
BAX, G.
BOOTH, A. C.
BOURNE, J.
BOWES, N.
BOWLES, MISS R. E.
Bo YES, P. N.
BOULLE, P. F.
BRASH, F. A.
BRETT, P. G. C.
BREWTTT, A. D.
BROKENSHA, M. A.
BROWN, J. W.
BRUIJN, J.
BRUNIQUEL, J.
BUCHANAN, E. J.
CALDER, D. B.
CARGILL, J. M.
CARNEGIE, A. J. M.
CARNOCHAN, R.
CHIAZZARI, L. F.
CHRISTIANSON, W. O.
CLEASBY, T. G.
COLLINGWOOD, R. J.
COLLINGWOOD, D. J.
COOPER, G.
CORTS, B.
CUGUET, H.
DARLING, H.
DAVIES, W. F.
DAWES, C. H.
DEDEKIND, E.
DELGADO, N. H. C.
DICK, J,
DICK, J. Mc.D.
DODDS, H. H.
DOWSON, C,
Fortieth Annual Conference
The Fortieth Annual Congress of the South African Sugar Technologists'
Association was held at the S.A. Sugar Association's Experiment Station,
Mount Edgecombe, on 21st to 25th March, 1966
The following members and guests were present:
L. F. CHIAZZARI (President)
DENT, C.
DREW, B. L.
DE ROBILLARD, P. J. M.
DU PREEZ, P.
DU TOIT, J. L.
ELLIS, MRS. E.
ELSDON-DEW, R.
FENWICK, J. H.
FOURMOND, T.
FRANCIS, D. W.
FRASER, J. C.
FRANGS, G. B.
FROESLER, H. P.
GILLATT, J. F. G.
GlLFILLAN, E. C.
GINSBERG, A. M.
GIRLING, R. A.
GLENNIE, P. R. A.
GLOVER, J.
GOURLAY, I.
GRAHAM, W. S.
GRADENER, S.
GRIFFITH, D.
GRIFFITHS, E. K.
GRINDLEY, L. R.
HACKLAND, D.
HALSE, R. H.
HAMMOND, A. G.
HALL, T.
HALL, J. R.
HARDMAN, R.
HARRIS, K.
HARRIS, N. J.
HALL, D. M.
HEMPSON, W.
HENNING, D.
HEIN, J. C.
HARMSWORTH, J. R.
HILL, J. N. S.
HILLIARD, J.
HICKS, M. K.
HOLTON, S.
HOWES, A. M.
HUGO, J. R.
HULETT, D. J. L.
HUNTER, J.
HURTER, A.
JACHENS, W.
JEHRING, G.
JENNINGS, R.
JOHNSON, C. A.
JONES, D. H.
KING, N. C.
KLOMFASS, J.
KLUSENER, W. B.
KRAMER, F. A.
KRUUSE, R.
LANDREY, P.
LANDSBERG, J.
LAX, R.
LEE, GEO.
LINTNER, J.
LLOYD, A. A.
LOWNIE, J.
MACGREGOR, W.
MAGASINER, N.
MARSHALL, D.
MASSON, D. R.
MAUD, R. R.
MAY, W. B.
MlLNER, M.
MILLAR, J.
MOBERLEY, P. K.
MOOR, B. ST. C.
MORAN, J. F.
MORRISON, E.
MCCARTHY, J. C. H.
MCKENNA, A.
MCCLUNAN, E.
NEWTON, D.
O'CONNOR, H. A.
PACKHAM, G.
PEARSON, C H. O.
PEMBERTON, N. G.
PERK, C. G. M.
PFOTENHAUER, V. O.
POLE, R.
POYNTON, R. G.
PRESTON, W. H.
PRINGLE, D. H.
RAULT, J.
REID, T. G.
RENAUD, C. L.
RENNIE, L.
RENTON, R. H.
RIC-HANSEN, R. W.
RlSHWORTH, A.
ROTH, G.
ROYSTON, J. H.
ROUTLEDGE, D. A.
SARGENT, N. V.
SAUNDERS, K.
SAUNDERS, E. K.
SAVILLE, R. F.
SAVILLE, G. W.
SHARVELL, N. A. D.
SIMPSON, G.
STEAD, A.
STEIZAKER, P. H.
STEWART, M.J.
STRACHAN, MRS. E.
SWART, MRS. E.
TIJL, J.
TOMLINSON, K.
THOMSON, G. M.
THOMPSON, G. D.
TOURNOIS, R.
TOTI, R.
TUCKER, A. B.
TURCK, G.
TURNER, L. E.
TURNER, R. C.
TYSON, D.
VAN ECK, V.
VAN HENGEL, A.
VAN KEPPEL, MISS H.
VON DER MEDEN, E.
VAN DER SPUY, L. B.
VAN DER RIET, C. B.
WAGNER, C. L.
WARNE, D. E.
WIEHE, H. F.
WHITLEY, W.
WHITEHEAD, C.
WHITMARSH, A.
WILLIAMSON, B. B.
WINCKLER, W. G.
WILSON, J.
WILSON, J. S.
WILKES, D. S.
WAT, W. F.
WALSH, W. F.
WOOD, G. H.
WOOD, R. A.
WYATT, R.
YOUNG, C. M.

Proceedings of The South African Sugar Technologists' 1 Association—March 1966 1
FORTIETH ANNUAL CONGRESS
Proceedings of the Fortieth Annual Congress of the South African Sugar
Technologists' Association, held at the South African Sugar Association's
Experiment Station, Mount Edgecombe on the 21st to 25th March, 1966.
The President: Ladies and Gentlemen—I have great
pleasure in asking Dr. A. J. A. Roux, Director General,
Atomic Energy Board, to open our Fortieth Annual
Congress.
OPENING ADDRESS
Dr. Roux: Mr. President, ladies and gentlemen, I
was especially pleased to be invited to deliver the
opening address at the Fortieth Annual Congress of
your Association, because although it has been my
privilege to address many scientific gatherings, I find
it most refreshing to speak to experts in a field in
which I am a complete layman. In common with the
general public, we in the physical and engineering
sciences tend to take the consumable commodities of
life very much for granted. We spare no thought for
the scientists, the technologists, the farmers and others
who conduct the chain of events which brings these
necessities, for example sugar, to our tables. We are
also inclined to ignore the very significant contribution
which such commodities make to the national
economy. I was surprised to find out that, based on
gross output, the sugar industry is the seventh largest
manufacturing industry in South Africa, and that
sugar production has doubled in the last fifteen years.
I can assure you that I now regard sugar with greater
respect.
On reading through the Proceedings of some of
your earlier Congresses, I began to realise that research
in the sugar industry has a great deal in common
with my own field of nuclear research. Having discovered
this, and having learned a good deal more
about your Association and its work, I feel that although
I am a layman here, 1 am no stranger, and
that at this Congress of the South African Sugar
Technologists' Association, I am among colleagues,
and among friends.
The most striking example of the points of similarity
between our organisations is breadth of field. I was
most impressed by the scope and wide diversity of
your research programme, so aptly described by Mr.
Chiazzari at last year's Congress as covering all aspects
of the sugar industry "from the soil to the bag". By
comparison the Atomic Energy Board's programme
covers all aspects of nuclear research from uranium
ore treatment to power generation; one might almost
say "from the mine to the megawatt".
In addition, we are trying to produce uranium of
the highest possible purity to ensure ready acceptance
on world markets. It is mentioned several times in
J. R. GUNN (President) was in the Chair.
your Proceedings that your Association aims to produce
the best sugar in the world, because future
market success will depend on quality, not quantity.
Another less obvious point of similarity is the international
recognition of work done in South Africa,
and the very good relations with international organisations
in the same field.
1 have no doubt that in the years to come, the links
between the sugar and nuclear industries will become
stronger as more and more use is made of the products
of the nuclear age. Radioactive isotopes are being increasingly
applied in plant nutrition studies, soil research,
and in a multitude of practical and inexpensive
devices in industry. I venture to predict that the
electrical power used in your mills and refineries will
one day be derived in part from nuclear power stations,
and it is even possible that process steam, which
forms the subject of your Symposium at this Congress,
will at some stage in the future be generated by nuclear
reactors instead of coal-fired boilers.
The fact that two research organisations have a
similarity of approach to their work is perhaps not
as much of a coincidence as we think, and I would
like to examine this topic a little more closely.
In an age when worthwhile research, employing
modern techniques, tends to become more expensive
by the day and absorbs an increasing proportion of
national and industrial revenue, it is becoming more
and more necessary to be extremely selective in the
choice of the direction of one's research programme,
and even the choice of individual projects. The time
when a country — even a very large one — could
spread its research programmes over the whole field
of natural science, has passed. Dr. Glen Seaborg.
Chairman of the United States Atomic Energy Commission,
expressed himself as follows on research
effort in a large country such as the U.S.A.: "The
time has passed when every competent scientist can
have all the money he wants for any reasonably
worthwhile project".
A matter of equal importance is the calibre of the
research worker. To support a research worker of
average ability, however important the project on
which he works, is a serious risk to say the least,
unless he works under excellent guidance. In a report
to the President of the United States, his Scientific
Committee put strong emphasis on this aspect of research
and said: "In science, the excellent is not just
better than the ordinary; it is almost all that matters".

2 Proceedings of The South African Sugar Technologists' Association March 1966
A research programme in any particular field of
science should therefore be carefully selected according
to national needs, should be no larger than can
be justified on the basis of such needs, and should at
any rate never be larger than procurable staff, with
exceptional ability for research, can handle.
In this context, 1 wish to elaborate on three basic
criteria which are involved. The project must have
scientific merit; it must have technological merit; it
must have social merit.
Scientific merit implies that the work will contribute
to man's overall inventory of knowledge. Knowledge
leads to the amplification and diversification of
human capabilities; it extends the power of our
natural senses; it creates new senses and it widens the
boundaries of the human mind. That knowledge is
the father of progress is a truism that hardly merits
repetition, yet we must not forget that the invisible
plasma of basic knowledge from which our technological
advances arise, does not come from nothing
— it has to be created.
The search for knowledge has technological merit
if it can be applied in a practical way to the solution
of problems which beset us in our everyday life and
work. In John 8, verse 32, we read "And ye shall
know the truth, and the truth shall make you free".
This is as true in the scientific context as it is in the
spiritual. The freedoms we acquire are freedom from
want, freedom from disease, and freedom from
drudgery. Freedom is not, however, an automatic
consequence of truth. It is the task of the technologist
to translate the truth into transistors, nuclear reactors,
drugs, soil husbandry and processing techniques; it
is their task, in short, to translate truth into freedom.
Social merit is perhaps a little harder to define and
an easy way out would be to say that anything which
improves our "standard of living" — if that well-worn
phrase has any meaning — has social merit. However,
there is more to it than that. All around us, we see
evidence that Man is slowly but surely destroying his
own environment. We see precious topsoil stripped
of its very life and washed out to sea, we see pollution
of the atmosphere and of natural waters by industrial
activities, we see the balance of nature upset by the
injudicious use of insecticides, denudation of forests
and excessive cropping. It is the responsibility of the
scientist and technologist not only to help us to live
more comfortably in our environment, but to ensure
that the environment is preserved for future generations.
We also see Man destroying himself by what appears
to be a complete departure from moral and spiritual
integrity, and who is to say that science and technology
cannot help to provide the answer to this situation as
well. Here we have a common meeting place, without
barriers of language or ideology, where international
understanding can be cultivated. I see this as the most
promising way of breaking down misunderstanding
and suspicion between nations, and thereby bringing
about the most precious of all freedoms, the freedom
from fear.
Once a research organisation has decided to embark
on a particular programme, there are a number of
internal requirements to ensure its effective execution.
These include proper training of staff, the provision
of an efficient organisational framework, and programme
co-ordination.
Training is probably the most vexing internal
problem facing science and technology today. Human
knowledge and its application to new techniques are
evolving so rapidly, that there is a danger of what
one might call human obsolescence in the ranks of our
scientists and technologists. This can lead to technical
stagnation, and every research organisation must make
sure that its members have sufficient technical versatility
to respond continuously to the changing
scientific scene.
The matter of training is tied up very closely with
the provision of an efficient organisational framework.
The creative researcher is human, with all the (bibles
of humanity. Above all, he is an individualist, and
yet he cannot thrive in isolation. The organisational
framework within which he works must provide him
with an environment and an atmosphere in which he
will give of his best. It must ensure that he is content
as a member of a team with an organised, common
purpose, and yet retains his creative individuality.
Edison defined genius as "one per cent inspiration and
ninety-nine per cent perspiration", and one might
sum up this aspect of research organisation by saying
that you must provide your researcher with time to
think, and incentive to perspire.
Programme co-ordination goes much further than
its obvious function of ensuring that maximum benefit
is derived from available resources. Programme coordination
requires a continuous, overall assessment
of the practical needs of the industry, and a continuous
assessment of how effectively the current programme
is filling these needs. This will ensure that the emphasis
is always in the right place and that there is no unnecessary
duplication of effort. It will also ensure that
the very important factor of time is given the attention
it deserves. Those of you who have read "Gulliver's
Travels" will recall that at the Grand Academy of
Lagado, Gulliver met a scientist who "had been eight
years upon a project for extracting sunbeams out of
cucumbers . . . ", which would be bottled for use on
cold days. "He did not doubt that in eight years or
more, he should be able to supply the governor's
gardens with sunshine, at a reasonable rate". I think
you will agree that time has assumed considerably
greater importance in our modern world, both in the
acquisition of knowledge and its application in practice.
The latter is a very real problem in many research
projects; any delay in putting newly acquired information
to work considerably reduces its value, and
it is this aspect which merits particular consideration
in a well co-ordinated programme.
The place occupied in the overall research picture
by Associations such as your own, whose objectives
are centred around research and development, is a
very significant and important one. The value of your
contribution lies in your ability to view the overall
picture from the outside, but with eyes and minds
sharpened and trained by experience gained on (he
inside. As an Association, your view is objective, uncluttered
by mundane considerations such as budgets

Proceedings of The South African Sugar Technologists' Association- March 1966 3
and staff problems. You are therefore more conscious
of the present, can probably see more clearly into the
future, and can make better use of the experiences
of the past.
One of the more important functions of Associations
is education. You have no geographical or
technical horizons, and can therefore undertake the
continuous education of your members, whose knowledge
can so easily become obsolete. By means of
publications, conferences and symposia, the communication
and interchange of knowledge and achievement,
which is the basis of mutual education, is
assured. Up on the Reef, a Johannesburg snob is
defined as one who buys retail, and I believe that a
Natal snob is one who uses imported cube sugar. If
this situation still exists I think your Association
would do well to consider a programme of educating
the public, as a rider to your function of educating
your members.
As an Association composed of technologists from
all branches of the sugar industry, and representing
the separate companies and organisations, the exchange
of knowledge mentioned earlier makes the
Association a natural vehicle for co-ordinating research
programmes over the whole industry. This is,
in fact, almost an automatic process, in which the
Association acts as a feed-back path from the pooled
intellect to the separate programmes. It is abundantly
clear from your Proceedings, that your own Association
is very conscious of your responsibility in this
respect, and gives very positive guidance to the sugar
industry in the most efficient solution of its problems.
1 spoke earlier of the importance of the proper
environment and atmosphere in research. Now these
are not just vague abstractions, they are positive and
tangible necessities, depending as much on the colour
of the walls and the view from the window as they
do on the technical significance of the project being
tackled and the human relationships within the research
organisation, 1 believe very firmly that there
is no more vital ingredient — except perhaps money
— in any research programme, than atmosphere and
environment, because the success of the programme
depends on the enthusiasm and efficiency of human
beings. The human being, like any other animal, is
more susceptible to atmosphere and environment
than we sometimes take the trouble to realise. The
influence of a technical Association on the atmosphere
of a man's work-a-day world must not be underestimated.
Within the Association, he feels at home,
his technical curiosity is stimulated, his professional
pride is fed, and he also feels that he is making a
contribution. Within this group of colleagues, each
man retains his individuality, and any contribution
which he makes is as an individual with a name and
with a position in society. Foster such an atmosphere
within your Association, and its influence on the
industry which you serve will be as beneficial as your
technological contribution.
Your own Association, Mr. President, has a long
and proud tradition behind it. 1 was very interested
to learn that the Persians were the first to do research
into sugar refining, as a result of which, by 700 A.D.,
Nestorian monks in the Euphrates region were pro­
ducing white sugar by the purifying action of ashes.
You are therefore part of over 1,200 years of technological
history, and bear the heavy responsibility of
providing us with one of the prime necessities of life.
I referred to sugar as a necessity, although Shakespeare
obviously equated it to the good things in life,
for Falstaff says in Henry IV, "If sack and sugar be
a fault, God help the wicked." We all need the
essentials, and enjoy the good things in life, Mr.
President, and we owe much to Associations such as
yours, who ensure that we receive them.
I wish you good fortune in your endeavours to
maintain this tradition, to bring to your industry
fresh knowledge and new technology, and to encourage
a standard of excellence befitting your profession.
May this Congress be successful and fruitful,
and may the high standard established for your
discussions be maintained. I now have great pleasure
in declaring this Congress open.
Mr. J. P. N. Bentley in reply to Dr. Roux's opening
address.
Mr. President, Dr. Roux, Ladies and Gentlemen,
In reminding us of Edison's remark that "genius
is one per cent inspiration and ninety-nine per cent
perspiration" you have done the Natal coastal belt
a great service, for we do quite often have our share
of inspiration and we certainly have our ninety-nine
per cent perspiration in the sugar industry.
When in future our friends from the highveld
complain about Natal Fever we will know that they
are really jealous of Natal Genius.
Of particular interest to us in the sugar industry is
your prediction that one day we will be generating
power and raising process steam using nuclear reactors.
This may not be as far off as we think once
the new prototype Fast Reactor now under construction
in the far north of Scotland, at Dounreay, has
proved itself. In this reactor the power of the atomic
bomb is controlled and the heat produced is taken
up by a primary liquid sodium circuit within the reactor,
transferred to a secondary liquid sodium circuit
which, in turn, transfers the heat to the steam circuit.
This type of reactor is sometimes referred to as the
third generation reactor if we consider Calder Hall
with its magnox unit as the first generation and the
advanced gas-cooled reactor as the second. It is called
a Fast Breeder because the speed of the neutrons
during the reaction is so high that some escape from
the core and they bombard a surrounding shield of
uranium 238, which is natural uranium and nonfissile,
and convert it to plutonium which is then used
with uranium 235 as the fuel for the reaction.
In this way the reactor can be made to breed more
fuel while it is operating, and it is said that it is expected
to extract 75% of the available energy of the
fuel compared with 2% achieved in existing nuclear
power stations using so-called "slow" reactors.
Under these circumstances, it is expected that the
cost of producing electricity will be below 0.25c. per
unit.
At present we are buying electricity at just over
Ic. per unit in the sugar industry. We can generate

4 Proceedings of The South African Sugar Technologists* Association-March 1966
it ourselves using coal as a fuel at between 0.75 and
0.95c. per unit depending on the efficiency of the
equipment we have available and, when using bagasse
as a fuel and setting a realistic value to bagasse, we
can generate at about .5c. per unit. If a nuclear power
station could supply us with electricity at, say, .3c.
per unit, the feasibility of buying all our power requirements
immediately moves into the realms of
practical economics.
Dounreay prototype Fast Breeder Reactor is expected
to be generating 250 MW by 1971, by which
time we should know how close to their estimated
costs they have come and how soon the nuclear
power station will become a serious competitor of the
conventional thermal power station.
As in the nuclear field, so also in the sugar technology
world we are making significant progress.
After more than 10 years of research the traditional
sugar milling process now looks as if it will have to
make way for the simpler, more efficient and more
economic diffusion process. Here, in Natal, Rabe has
now developed a juice clarification process which may
well revolutionise our ideas and set us on the road to
producing the best sugar in the world. This is what
Mr. Gunn, our President, set out as one of our aims
in his inaugural address some three years ago.
May I say, Dr. Roux, that we as sugar technologists'
are both proud and pleased that you feel you are
among colleagues and among friends. We are proud
that you, as Director-General of the Atomic Energy
Board, should consider us as colleagues and pleased
that you should so rightly feel that you are among
friends.
On behalf of the members of our Association and
our guests, may I say that we have found your address
most stimulating and thought-provoking and I would
like to thank you for coming all this way to address
us and for giving up so much of your valuable time
to us. I hope you have also found your trip rewarding
and worthwhile.
PRESIDENTIAL ADDRESS
Dr. Roux, Ladies and Gentlemen, it gives me pleasure
to deliver my presidential address to this, the
fortieth annual congress.
It is not my intention to delve into the past as that
is history and can be followed by reading the Proceedings
of our Association. Regarding the past, it is
sufficient to say that our Association has been built
on solid foundations, has enjoyed a healthy growth
rate and today is one of the largest sugar technologists'
associations in the world. It is also not my
intention to report on the immediate past activities
of our association as these have been adequately
described in my report to the Annual General Meeting.
Instead I propose to move into the future, the
immediate future, and to do a bit of crystal ball
gazing.
The Challenges of the Future
It is my opinion, and this is shared by many, that
during the next ten years but probably sooner, the
sugar industry will undergo radical changes. These
changes will be partly voluntary due to improved
methods and partly involuntary due to labour shortages.
I have no doubt that our industry will not suffer
as a result of these changes as 1 have every confidence
in our technologists to be ahead and on top of any
problems which will arise. However, I do predict
that the future holds many challenges for the ingenuity
and technical skills of our members.
The most obvious obstacle in the immediate
future is the shortage of field labour for cutting and
loading of sugar cane. This problem is urgent and
requires immediate attention. It is often said that the
best form of defence is attack and here I submit that
our only defence against a harvesting labour shortage
is attack, not next year or the year after but today,
right now. I believe that the labour shortage is beginning
now; it is inevitable that we shall be short of
labour in our fields. There does not appear to be
anything that we can do about it except to face up
to the fact that we are rapidly running into this problem.
Many sugar producing countries have experienced
the metamorphosis from hand cutting to mechanical
harvesting and have emerged with what appears to be
a reasonably satisfactory solution. I say "appears"
because often if one looks deeply into it, one finds
that given time, the system could have been improved.
However, the pattern is the same in most countries,
they were not given time to convert efficiently. We
find the Hawaiians bulldozing, push raking and tearing
their cane fields asunder and then having to resort
to laundries, huge installations using oceans of water,
to make the cane anything like millable at the factories.
In Florida, where the terrain is so flat that there are
no distinctive land marks and one can easily get lost,
the solution has been a half manual half mechanised
monster that creeps slowly along creating noise and
dust in its action of picking up hand cut cane and
transferring it into in-field trailers from which it is
subsequently transferred into huge haulage units
which deliver it to the mill. But a large portion of the
work is still manual, the cane is cut by hand.
That delightful tropical island in the Caribbean,
Puerto Rico, which was the venue for the last international
congress, is undergoing an enforced change
from manual cutting to mechanisation and it has been
found that neither the Louisiana nor the Florida
method is successful in Puerto Rico. Many of us
visited this island last year and we could see the
trouble that industry was experiencing due to shortages
of labour.
I have mentioned other lands for a very real purpose
and that is to illustrate what happens when
conditions are forced on one without any time for
experimentation. I believe that if we in South Africa
do not start taking a very serious look at mechanised
harvesting we shall be too late ,by tomorrow. I
believe that this problem is so serious that our defence
against it must be a vigorous attack on all aspects of
mechanical cutting and loading, either in one unit or
separately. We must invent, experiment and circumvent
so that before the evil hour descends on the
industry, we shall have all the answers. I am here
repeating the urgent appeals that have been made from
time to time by the Mechanisation Committee.

The success of mechanised harvesting will depend
to a large extent on revising certain agricultural practices
and possibly altering some of the off-loading
appliances at the mills. One thing is certain, the mechanical
harvester must be designed to be the best
suited to our lands and then everything else must be
tailored to suit the harvester. 1 feel that this is the
only correct approach to the problem. Any other way
will stifle the design of the most difficult machinery
due to lack of scope and imagination. Altering
unloading facilities at the mill may be costly but will
not be difficult. I believe that this lesson has been
learnt in other countries but could not be put into
practice due to lack of time.
Revised Agricultural Practices
It is quite obvious that no mechanised harvesting
will be successful when the cane setts are planted in
furrows and are protected by ridges on either side
of the furrow. It will be impossible to cut the cane at
ground level without cutting into the ridges as well.
This will cause damage to the cutting mechanism and
frequent breakdown of the plant. One obvious change
in agricultural practices will, of necessity, have to be
the method of preparing the land for planting cane
setts. Here the Mechanisation Committee has very
firm and fixed views regarding setting the implements
on the tractor tool bar to "hill up" the soil rather than
leave the usual furrow for planting. To add voice to
their appeal, we must use this new method of planting
now because, who knows, the fields may have to be
mechanically harvested before they are replanted the
next time. It will take a long time to have all our fields
planted to suit mechanical harvesting and every delay
now will mean a corresponding delay in the final
completion of the programme.
It has been found in many countries that burning
is the only solution to reducing the trash content of
the mechanically harvested cane. This will be a complete
departure from present practices and we will
lose the benefit of the trash blanket or mulch. However,
mechanically harvested cane should be delivered
to the factories in no worse condition than it is at
present (preferably in better condition) and it may well
be that mechanical trashing of the cane will prove to
be too costly and complicated. We must investigate
this and have the answers before it is too late.
It does appear that the industry is already tending
towards a shorter growing period between cuttings.
I believe that it has progressively reduced from 21
months to 17 months. I suggest that mechanical
harvesting may well cause us to reduce this period
still further. Whether this will be beneficial to the
factories is not known but what is known is that
mechanised harvesting cannot deal with heavy long
cane that has lodged. We must find out whether the
tendency will be for cane to be cut before it reaches
full maturity. The purity of the cane juice may be
affected and this could have very detrimental effects
on sugar production.
It has been found in Australia and in Florida that
cutting the cane into lengths of 12 to 18 inches has
greatly facilitated the loading operation. As the ideal
machine should both cut and load, these two coun­
tries have solved the latter problem by ejecting short
lengths of cane into basket-like carriers which are
driven alongside the harvesting machines. Whilst this
solves one problem, it creates another. Every cut
through the cane opens an area of potential infection
and of enzymatic attack causing inversion. It also
makes the drying-out of the stalk~ very much quicker
and if there is a delay in milling the cane there will be
large losses in weight and sucrose. The economics of
this loss have to be assessed accurately and we must
have this information before it is too late.
In the long run I visualise the plant breeders specialising
in canes which have as one of their attributes
the suitability for mechanical harvesting. I see that
the immediate future has interesting challenges for
our agricultural technologists and there is no time
like the present to commence the attack.
Future Trends in Factories
I foresee that the traditional brute force sucrose
extraction methods at present used in South Africa
are going to make way for the more subtle powersaving
diffusion. In fact in the not too distant future
we shall have at least three diffusion plants in our
industry. I am quite confident that these three diffusion
installations will yield results superior to most of our
existing milling tendems. It would seem that now
that we know more about milling than most other
countries; (this is a result of the recent mutual milling
control experiments) we are to branch out into a new
sphere of lixiviation or diffusion. I am confident that
our technologists will rapidly become experts in
this new field.
Diffusion opens up really exciting new possibilities.
Any diffusion plant should reduce the quantity of
starch in the extracted juice and as starch is one of
the major ingredients about which raw sugar refiners
comment, this in itself is a step forward. It does
not end there because the latest trends are to combine
the diffusion process with the clarification process
and to decant clarified juice from the diffuser. It
is claimed that this major break-through has been
achieved in Hawaii. I am confident that if it can be
done successfully elsewhere, we shall find a way of
doing it here. This combined diffusion-cum-clarincation
process has the added possibility that with
adequate control, an almost complete removal of
starch may be possible. There is plenty of meat into
which our mill technologists will be able to get their
teeth.
Automation
The advent of diffusion, reducing the amount of
brute force power used in the extraction process,
should make the automation of this process a greater
possibility. It has been done elsewhere. Thus we hope
to see a factory that can control its throughput of
cane to very close limits so that all the peaks and
troughs of production are eliminated right at the
beginning of the process. This will lead to steady
conditions throughout the whole process and will
radically improve the possibility of the pansman producing
a better quality sugar. Automation of each
phase of production, being easier with diffusion.
5

6 Proceedings of The South AJrican Sugar Technologists' Association—March 1966
should then lead to the next logical step and that is
integrated control and finally computer control. I
know that our industry, at present, views instrumentation
and automation with suspicion because of
the difficulty of maintaining and servicing delicate
instruments. I believe that the sooner this prejudice
is removed the better. In the first place the care of
delicate instruments must be carried out by competent
instrument mechanics and every large factory should
have a place on the staff for at least one such employee.
1 believe that due to the large throughput at many of
our factories we cannot afford any longer to have
most functions manually controlled. A proper automatically
controlled system does not suffer from human
errors and fatigue and will give a more constant
quality article. Having achieved this state of affairs,
an integration of the automatic systems either by
interlocking or computer control will see our industry
on the same lines as the petroleum or other highly
automated industries.
Before I conclude my address, I foresee a shortage
of sufficiently highly trained men in the technician
level in our industry. 1 acknowledge that the South
African Sugar Millers Association Limited is financing
an excellent course for factory technologists. 1 agree
that this course is urgently required and is filling a gap
in this level of employment. What 1 am concerned
with is the wastage on this course. We must not lower
the standard of the course but rather we must make
the course and profession of a factory sugar technologist
more attractive in order to entice more applicants
of the right character and qualifications. An
automated process requires technicians who fully
understand the process itself and also the functions
of the automatic controllers.
In conclusion I wish to state that I have enjoyed the
high honour of being your president for the last three
years and I must thank you for having been so tolerant
with me. I wish to thank the Council who have so
diligently applied themselves to the task at hand and
who have assisted me greatly and also I wish to thank
Mr. Chiazzari, our vice-president who has always
been available with his excellent advice when I have
needed it. I wish to congratulate Mr. Chiazzari on his
appointment as president and I wish him the best of
luck in this high office.
Mr. L. F. Chiazzari in reply to the President's
address.
Mr. President, Ladies and Gentlemen.
Before replying to our President's address I am sure
you would like me to express our appreciation for
what he has done for our Association. As most of
you know, he has held this office for three years and
during that whole period I have been his Vice and in
that position have got to know him rather well. His
application to his duties and his untiring efforts on
our behalf are most manifest, but 1 think it was his
quiet manner of leadership when he was head of our
delegation a year ago at the International Society of
Sugar Cane Technologists' Congress in Puerto Rico
that impressed me so much. I can assure you he gained
the esteem and respect of all the other delegates as
well. A final tribute. He is one of a small band who
do not wish to remain parochial in this widely diversified
industry of our. Primarily an engineer, he has
made it is his business to study all facets and today
is a very knowledgeable person.
Our President's address commences by issuing
challenges to all of us to face up squarely to the radical
changes that must inevitably take place. He particularly
refers to improved methods and labour
shortages. In his own inimitable style, full-of enthusiasm,
he tells us attack is the keynote and not to wait
for next year, or the year after, but today, right now.
We really must "give it the gun" (Gunn).
Summarising the main points of his address very
briefly, he mentions mechanical substitution of field
labour, diffusion as opposed to intensive mechanical
force in juice extraction, instrumentation to supplant
factory labour shortages, automation and computer
control, to improve efficiency and higher standards of
training for our technologists. All these emanate
from positive thoughts and are materialisticmechanics,
electrics, electronics and education, consequently
an industrial revolution in the field of
labour is fast approaching. The raising of the standard
of living and work must inevitably follow and lead
to the non-European taking over the tasks of the
lower order now being undertaken by the European.
Already the demand for workers with a technical
training is rapidly increasing.
Now to implement these ideas. How best to do it?
There are already certain moves under way at the
moment, particularly in our factories and the Mechanisation
Committee, but in order to avoid duplication,
which can be unnecessarily expensive, it is essential
to pool all ideas and have co-ordination. The S.A.S.A.
has the Technologists, this Experiment Station and
the Mechanisation Committee while the.Millers have
the S.M.R.I, all of whom are already integrated under
a Research Committee recently formed. So it would
appear that the necessary machinery is already in
existance and only requires to be geared to action.
Ladies and Gentlemen, I conclude by once again
thanking you, Mr. Gunn, for giving us such a stimulating
address.

Proceedings of The South African Sugar Technologists'' Association—March 1966 7
FORTY-FIRST ANNUAL SUMMARY
OF LABORATORY REPORTS
OF SUGAR FACTORIES IN SOUTHERN AFRICA FOR THE 1965/66 SEASON
by CHARLES G. M. PERK
MB.—The data and figures in this summary are as declared by the Mills in their Final Laboratory Reports.
THE SOUTH AFRICAN CROP 1965/1966
The sugar belt in South Africa experienced the
worst summer drought ever recorded. In the period
from November 1964 till April 1965 only 13.36 inches
of rain fell compared with a computed mean rainfall
of 23.17 inches for the same period. It hit the crop
at the worst possible time, i.e. in the period of optimum
growth. Therefore any rain falling afterwards could
not make up for the considerable loss of growth
experienced in the optimum growth period. Exceptional
cold weather following the drought did not
help either to restore the damage done to the crop.
Owing to these adverse weather conditions the cane
ripened prematurely; the cane harvested in the first
months of the 1965/66 season showing a higher
sucrose % cane than usual for these months. However,
though the sucrose content was higher in the
first months, it failed to rise further, i.e. to the usual
level for the middle of the season. In the contrary
the sucrose content started also dropping prematurely;
proving that the cane had been actually ripe in the
first months of the season. It is obvious that with such
a sequence of adverse phenomena the mean sucrose
content for the whole season could be low only.
To illustrate the abnormal trend of the sucrose
content of the cane in the 1965/66 season the following
table shows:
(i) the average trend, i.e. a ten year mean of the
sucrose content of the cane by months, for the
S.A. sugar belt.
(ii) the sucrose % cane by months for Pongola's
1965/66 crop, and
(iii) the average trend by months for the other 17
mills (except Pongola), also for the 1965/66
crop.
The second column of the table shows the usual
trend, i.e. a gradual increase of the sucrose content
from May onwards, reaching its maximum in September,
followed by a gradual decrease, ending in
February with the same sucrose content as it had in
the beginning of the season, in May.
Pongola showed in 1965/66 a similar trend, only
the top is here reached a month later than in Natal.
The last column of the table shows the 1965 trend
for all factories with exception of Pongola. It reveals
how in 1965 owing to "noodrijping", i.e. premature
ripening of the cane owing to need, the sucrose '%
cane was higher than usual in the first months of the
season. However, it did not increase any further in
the following months because the cane was already
mature. The assumption of "noodrijping" explains
too why the drop in sucrose content also started
prematurely.
Comparison of the Sucrose "„ Cane by Month
Month
Mav
June
July
August . . . .
September .
October. . .
November .
December .
January.
February
Average
previous
10 years
12.35
13.00
13.57
14.20
14.44
14.21
13.64
13.11
12.59
12.35
Pongola
1965/66
13.55
14.20
14.16
14.50
14.34
14.16
12.46
10.97
Average all
other Mills
1965/66
13.49
13.47
13.43
13.50
13.06
12.91
12.22
12.25
12.18
1.1.69
In September, when the cane should have shown
its maximum sucrose content, it was already on its
way down. The result of this abnormal trend was that
the average sucrose % cane of the Optimum Period
was lower than the average of the months preceeding
this period. As the higher sucrose content in the first
months could not make up for the loss in the Optimum
Period, the season ended with a low average sucrose
% cane for the whole crop.
The low sucrose % cane combined with the low
yield of cane owing to stunted growth resulted in a
sugar production for the 1965/66 season which was
28'% lower than the record tonnage obtained in the
1964/65 season.
Tons of 2,000 lbs.
Season
1961/62 . . .
1962/63 . . .
1963/64 . . .
1964/65 . . .
1965/66 . . .
Tons Sugar
1,098,781
1,193,279
1,264,704
1,395,446
1,001,784
N. .B. -Pongola has been singled out, because as all cane has
been en irrigu_ irrigated here, the cane was not seriously alTected by
the e drought.
Tons Cane
9,390,544
10,731,263
10,970,338
11,752,031
9,266,324
Cane/Sugar
•Ratio
8.51
9.01
8.66
8.42
9.21

8 Proceedings of The South African Sugar Technologists' Association March 1966
Tons of 1,000 kg.
Season Tons Sugar
1961/62 . . . 996,926
1962/63 . . . 1,082,525
1963/64 . . . 1,147,321
1964/65 . . . 1,265,921
1965/66 . . . 908,803
Tons Cane
8,513,085
9,751,707
9,939,529
10,661,207
8,406.269
Cane/Sugar
Ratio
8.51
9.01
8.66
8.42
9.21
N.B. -The above recorded tonnages of sugar are the official
tonnages as supplied by the S.A.S.A. They dilfcr from the
tonnages according to the Laboratory Reports of the Mills
because Glcdliow and Sezela do not declare their actual produced
sugars, but only the sugar tonnages passing from their
rawhouses to their refinery departments.
OPTIMUM PERIOD RESULTS
It is routine in South Africa when comparing different
seasons, to compare not the results of the whole
season, but the result of the Optimum Period of each
season. This has two advantages. Firstly it cuts out
differences in length of the seasons and secondly it
restricts the comparison to the same cane areas. For
example in the past season there were several Mills
which started late, while there were other mills which
ended their crushing seasons much later than usual
for those mills. From mid-August onwards, however,
all mills were crushing. Therefore the Optimum
Period of the past season covers approximately the
same harvesting area as in the previous seasons.
N.B.—The cane/sugar ratios of the Optimum Periods as well
as of the whole season are based on sugar tonnages as declared
by the Mills (and NOT on official sugar tonnages).
Season 1961/62
Optimum Period .
Balance of Crop
Total Crop . .
Season 1962/63
Optimum Period .
Balance of Crop .
Total Crop . .
Season 1963/64
Optimum Period .
Balance of Crop .
Total Crop
Season 1964/65
Optimum Period .
Balance of Crop .
Total Crop
Season 1965/66
Optimum Period .
Balance of Crop .
Total Crop
of
Crop
69
31
100
56
44
100
59
41
100
60
40
100
67
33
100
Sucrose
Cane
14.11
12.98
13.75
13.77
12.65
13.30
13.91
13.02
13.55
14.41
13.17
13.90
13.10
12.76
12.99
Fibre
Cane
14.46
14.63
14.52
15.32
15.73
15.50
15.38
15.66
15.50
15.20
15.62
15.38
15.44
15.83
15.57
Cane
to
Sugar
Ratio
8.23
9.18
8.51
8.58
9.63
9.01
8.36
9.06
8.63
8.06
9.01
8.38
9.06
9.50
9.20
Purity
of
Mixed
Juice
86.69
84.52
86.04
83.51
83.15
83.36
86.09
84.10
85.30
86.01
84.74
85.52
84.53
83.50
84.22
N.B. ••-For Comparison of optimum period results of seasons
before 1960, we refer to the 36th Annual Summitry (Season
1960/61) where a review of results from 1928 till I960 is shown.
Our present table shows that the cane/sugar ratio
of the Optimum Period for the past crop was more
than 9.00 to 1, such a high ratio being experienced
only before 1937. However, before 1937 the Overall
Recovery was approximately 10% lower than at
present and the high ratio due more to unsatisfactory
overall recovery than to a low sucrose content of
the cane.
As already stated, the most remarkable (and regrettable)
fact of the past season was that (he cane
harvested in the Optimum Period was less good than
that crushed in the preceeding months. It can only
be satisfactorily explained by assuming that owing to
premature ripening the peak sucrose content of the
cane came some months earlier than usual and subsequently
the drop in sucrose %, cane started also
earlier than usual.
VARIETIES PERCENTAGES
Another item of interest concerning the South
African Cane Crop is the composition of the crop
with regard to the different cane varieties. The following
table shows the percentages of cane varieties
harvested in the last five seasons. It shows how the
variety N:Co.376 is gaining more and more momentum,
which is particularly true for the South and
North Coast areas of the S.A. Sugar Belt, where
50% and more of all cane crushed consisted of
N:Co.376. The crops of the Zululand Mills: Umfolozi,
Empangeni and Amatikulu and of the Transvaal
Factory Pongola still consist of 3/4 or more of
N:Co.310, but Felixton and Entumeni, although
being also Zululand Mills, are exceptions in this
respect. Entumeni's main variety is N:Co.293, followed
by N:Co.376 and N:50/2U; N:Co.310 taking
fourth place. Felixton is replacing its N :Co.310 more
and more by N:Co.376.
Season
Uba . . .
Co.281 . .
Co.290 . .
Co.301 . .
Co.33l . .
P.O.J.'s . .
N:Co.310 .
N :Co.292 .
N :Co.293 .
N:Co.334 .
N:Co.339 .
N:Co.376 .
N :Co.382 .
N :50/211
1961/62
0.01
0.01
0.01
0.56
8.97
0.01
55.65
2.36
5.23
0.42
4.75
17.03
I.I 1
0.01
1962/63
0.01
0.01
0.01
0.24
8.89
0.02
54.00
2.28
4.62
0.33
3.67
18.04
1.92
0.22
1963/64
0.01
0.01
0.01
0.12
6.32
0.01
50.75
2.03
4.93
0.44
3.23
21.45
1.81
1.23
1964/65
0.005
0.O1
Nil
0.07
4.41
O.Ol
46. 9 1
1 .32
3.72
0.42
2.57
23 . 36
2.87
2.84
1965/6
._
2.70
40. 15
0.H9
4.51
1.76
32.19
3.35
.3.52
The Summary of Agriculture Data: Sugarcane
Crop 1965 by J. L. du^Toit and M. G. Murdoch
(also published in the 1966 Proceedings) allows us to
have a look into the future. This paper mentions that
only 16% of the plant cane consist of N:C'o.31(),
while 46% is made up by N:Co.376. It shows too
that only in the Midlands the variety Co.331 is still

Proceedings of The South African Sugar Technologists' Association—March 1966 9
a main variety with 13% plant cane; in the other
areas it has dwindled down to 1% or less. The newer
variety N:50/211 covers 7% of the 1965 plant cane,
approximately the double of the percentage crushed
in the past season.
N:50/211 was released in 1959; of the other varieties
(also bred in Natal but) released later, no records
regarding their extension are available. However, as
soon as their percentages exceed 1% of all cane
crushed, they will be promptly recorded in the Annual
Summaries.
OVERALL IMPRESSION OF THE
SOUTH AFRICAN SEASON 1965/66
The S.A. Crushing Season 1965/66 is not one to
look back on with satisfaction. This does not refer
only to the disappointing sugar tonnage obtained,
but also to the operational results of the mills.
It started in the beginning of the season, when a
number of factories had to postpone their starting
date because the reconstructions of their factories
were not yet completed. The main reason for the
delay was often a too late delivery of machinery
ordered from overseas; lack of sufficient artisans
played its part too. In this connection it should be
mentioned that during the inter-season 1965 not only
were extensions carried out at existing factories, but
in addition three new factories were under construction,
i.e. New Amatikulu, Jaagbaan and Dalton. In
addition to the three complete factory plants for
these new mills, boilers, mills, heaters, evaporators,
vacuum pans, centrifugals, turbo-alternators, etc.,
had to be manufactured and installed in the existing
mills.
A further and really very unwelcome delay was
experienced at those mills where the new plant
showed shortcomings.
The milling tandems maintained, in general, their
high standard of performance obtained in previous
years, but the performance of the boiling house of
the factories was, in general, disappointing. Owing
to a bigger quantity of final molasses than commensurate
with the purity of the juice and moreover a
final molasses with a high purity, the boiling house
performance was low, in general.
TIME ACCOUNT OF THE
SOUTH AFRICAN MILLS
In the following table different details regarding
length of the season, time efficiency and the mean
crushing rate are compared for the last three seasons:
Of more interest than a comparison of the results
of the last three years is a review including the years
before the expansion caused by the Sugar Production
Programme 1949 and an estimate of the future combined
crushing rates of the S.A. Mills. So far in all
tables (also in the Time Account table) the Mean
Crushing Rate of all mills has been shown, however,
in the following table (not the weighed average) but,
the sum of all individual crushing rates will be used
for comparison purposes.
Season
Number of Factories .
Tons of Cane Crushed
Total hours Mills open
Total hours actual
Crushing. . . .
Average number of
weeks of Mills open,
per Mill . . . .
Average number of
days of actual Crushing,
per Mill. . .
MEAN CRUSHING
RATE (tch) . . .
MEAN OVERALL
TIME EFFICI­
ENCY . . . .
Hours Cane Shortage
% hours Mills open
1963/64
17
10,956,448
91,038
82,111
37i-
201
133
90%
6%
1964/65
17
11,752,031
92,457
85,266
37|
209
138
92%
T 1 0/
J 1 /()
1965/66
17*
9,266,324
76,751
67,090
311
164
138
87 !•%
* Since New and Old Amatikulu did not crush simultaneously
(the former crushed after the latter had stopped), they are
accounted for as "one" Mill.
Season
Increase in Combined Crushing Capacity of the
South African Mills from 1950 onwards
1950/51 . . . .
1951/52 . . . .
1952/53 . . . .
1953/54 . . . .
1954/55 . . . .
1955/56 . . . .
1956/57 . . . .
1957/58 . . . .
1958/59 . . . .
1959/60 . . . .
1960/61 . . . .
1961/62 . . . .
1962/63 . . . .
1963/64 . . . .
1964/65 . . . .
1965/66 . . . .
1966/67 . . . .
1967/68 . . . .
Sum of
Crushing
Rates
1,300
1,294
1,338
1,364
1,528
1,665
1,693
1,766
1,868
1,883
1,962
2,033
1,946
2,147
2,228
2,414
2,925
3,120
Percent of
1950/51
100.0
99.6
102.9
104.9
117.6
128.1
130.2
135.8
143.7
144.8
150.9
156.4
149.6
165.1
171.4
185.7
225.0
240.0
6%
Percent
Increase
0.0
-0.4
+ 2.9
+4.9
+ 17.6
+ 28.1
+ 30.2
+ 35.8
+43.7
+44.8
+ 50.9
+ 56.4
+49.6
+ 65.1
+ 71.4
+ 85.7
+ 125.0
+ 140.0
N.B.—The sum of the crushing rates obtained in the 1950/51
season-100.0%. The figures for the 1966 and 1967 seasons
are estimated.
The table of the combined crushing rates reveals
that as a result of the launching of the Sugar Production
Programme 1949, the combined crushing rate
of the S.A. Mills increased from 1300 to 2000 tch,
or by approximately 50%.
The announcement of the Minister of Economic
Affairs on January 24th, 1964 — after consultation
with the S.A. Sugar Association — that the restriction
on the allocation of new sugar quotas was removed,
gave another impulse to extension and a far bigger
one! It did not lead only to a general expansion of
the crushing capacities of the existing mills, but it
led moreover to the planning and erection of new
sugar factories. The table shows that when the latter
extension programme will have been completed, the
combined crushing rate will be nearly 2\ times the
initial one.

10
Before starting the general discussions regarding
the results of all factories regularly reporting to the
S.M.R.I., some information and some details should
be given about the new factories and other new
features.
ABOUT THE NEW SUGAR FACTORIES
IN SOUTH AFRICA
The old Amatikulu Mill closed down for the last
time (at the present site) on the 12th November 1965.
Shortly afterwards its task was taken over by the
new Amatikulu Mill, also of Hulett's Sugar Mills
& Estates Ltd., a mill with a capacity of 250 tch or
more than twice that of the old mill.
In the 1966/67 season two other new mills will
come into operation in the Noodsberg area. One
called Jaagbaan and belonging to the Noodsberg
Sugar Company will also have a capacity of 250 tch,
while the other named Dalton Mill and belonging to
the Union Co-operative Bark & Sugar Company
will crush initially at a rate of 60 tch.
In the 1967/68 season another 250 tch mill will be
commissioned, situated between Malelane and Hectorspruit
and belonging to "Die Transvaalse Suikerkorporasie
Beperk". In addition farmers in the Hoedspruit
area (also in the Eastern Transvaal) intend to
erect another 250 tch mill; the association of farmers
concerned being "Die Blijde-Letaba Suikermaatschappij
Beperk".
The new Amatikulu Mill is equipped with seven
84 inch mills driven by single-stage steam turbines,
while the Jaagbaan tandem comprises six 84 inch
"self-setting" mills driven by electric cascade drives.
The Malelane Mill will recover the sucrose from the
cane by the milling-cum-di(fusion process; the diffusion
plant being preceeded and followed by one
84 inch "pressure-feeder" mill. Another new milling
tandem which will also come into operation in the
1966/67 season is the five 84 inch "pressure-feeder"
mills of Sezela.
In the 1967 68 season there will be four tandems
of different composition in operation in South Africa,
all four handling more than 200 tons of cane per hour
(from 5500 to 6000 tons per day). One tandem will
comprise seven mills, the second six mills, the third
five mills and the fourth installation two mills only;
a continuous diffusion plant of the percolation type
doing the job of the five replaced mills.
We are sure that sugar technologists the world
over will follow the results of these four installations
with the greatest interest.
Before leaving the subject of "mills", it should be
mentioned that the New Amatikulu units have been
fitted with feeder rollers which are directly driven by
pinions from the top roller shaft. This arrangement
avoids driving chain breakages. Jaagbaan has gone
a step further with independently driven feeder rollers;
the rollers being driven by variable speed hydraulic
motors. This will enable the mill engineer to run the
feeder rollers, each at its optimum speed relative to
the circumferential speed of the mill rollers.
Proceedings of The South African Sugar Techno log is Is' Association March 1966
Malelane will neither be the first nor the last
factory in South Africa to be equipped with millingcum-dilfusion.
Not the last, because the Hoedspruit
factory will also apply diffusion and existing factories
in Natal are contemplating to extend their capacities
by introducing diffusion. Not the first, because the
"primeur" for Natal will be the Entumeni and Dalton
factories which will start applying diffusion in the
1966 67 season.
In Natal, Dalton will also inaugurate the doublecuring
of C-massecuites with the aid of continuous
centrifugals, while Malelane will extend the use of
continuous centrifugals to the curing of the B-massecuites,
the next season.
Entumeni and Jaagbaan are extending the number
of electrically-driven mills in Southern Africa with
three and six units respectively; both factories with
a reduction of heat requirements in mind.
Umfolozi, Entumeni and Dalton will increase the
number of "climbing film" apparatuses in Southern
Africa from one to four: extending the one of 20,000
sq. ft. at Ubombo Ranches with three of 9,000 sq. ft.
heating surface each.* The quadruple effects of the
new factories: Amatikulu. Jaagbaan and Malelane
are of the conventional "Natal" type, i.e. the large
heating surface required in the first effect being divided
over more than one "short-tubed" vessel.
N.B. - -The climbing film apparatuses were introduced to reduce
the residence time of the juice, counteracting in this manner
the higher temperatures accompanying the introduction of
vapour bleeding (Quarterly Bulletin No. 3, 4 & 6, 1957/58).
The system of steam or vapour distribution by a
great number of angled slots around the top of the
calandria of pan or evaporator vessel may nowadays
be called "standard practice" for Natal. The original
idea dates from the beginning of this century when the
so-called "Puunene Evaporator" was provided with
cast-iron belts or bulges around its calandrias. It was
introduced as a vacuum pan application in Natal in
1957 with Stork Pans supplied to Umfolozi. We would
not have referred to this old "invention" if Jaagbaan
had not extended the application to juice heaters:
Jaagbaan's vertical juice heaters being provided with
an annular bulge or steam belt at the top part of their
bodies.
With regard to condensers there is, fortunately, a
general return to the counter-current condenser. The
disadvantages of the parallel-flow condenser, i.e.
higher water consumption and bigger volume of air
to be extracted make the latter type unsuitable for the
cane sugar industry where no cold injection water is
available.
The first of a set steam-jet thermo-compressors was
installed at Darnall during the 1965'66 season, while
Jaagbaan is equipped with a set of three steam-jet
compressors to re-compress part of the 5 p.s.i.g.
vapour to 15 p.s.i.g. process steam.
Fluidised bed cooling-cum-drying of white sugar
introduced at Gledhow in 1964 will be applied to
raw sugar at Jaagbaan; new Amatikulu uses the wellproven
system of rotary Louvre driers to dry its sugars.
*During the 1966 -'67 season another 20,000 sq. fl. Scmi-Kcstner
will be also installed at Umfolozi.

Proceedings of The South African Sugar Technologists' Association—March 1966 11
This chapter would not be complete without mentioning
that Pongola is installing a refinery department
annex to the raw sugar factory. It will be a meltcarbonatation
refinery using a lime kiln for the production
of CaO and COa like Sezela and Gledhow.
Via Pongola's railway head at Piet Retief the refined
sugar can be distributed in the Transvaal.
DISCUSSION OF THE RESULTS OF ALL MILLS
REGULARLY REPORTING TO THE S.M.R.I.
Up to this juncture the discussions have been
restricted to general remarks regarding the South
African Sugar Crop and its prospects; from here on
more detailed discussions will follow concerning all
Mills, i.e. all sugar factories which regularly supply the
S.M.R.I. with their monthly and final laboratory
reports.
The following items will be discussed successively:
I. Operation of the Milling Tandems
II. Boiling House Operations
The latter subject will be subdivided into:
Boiling House Performance, Non-sucrose Accounts,
Reducing Sugars Accounts, Non-sucrose Circulation
and Exhaustion of C-strikes.
As a separate subject will be treated:
III. Factory Control
I. Operation of the Milling Tandems
Dr. K. Douwes Dekker commences the "Preface"
of "A Review of Terms used for Indicating Milling
Results''' (issued in 1951 as Communication of the
S.M.R.I. No. 7) with the statement that, so far the
ideal criteria forjudging the performance of our mills
has not been found, however, a figure denoting the
loss expressed in terms of juice is much more preferable
than indicating the performance as a percentage
ratio of the sucrose recovered to the sucrose present
in cane.
In the article itself, it is explained that a mill does
the only operation it can do, i.e. squeezing liquid out
from the cane or the intermediate bagasse. It is therefore
that its performance should be based on this feat,
i.e. squeezing out of liquid, quite apart whether this
liquid contained sucrose or not, or any other solute.
As it is always more educational to express a performance
in the form of what is still lost than what
we have already gained, the performance of the mills
is indicated as parts of liquid lost per 100 parts fibre
instead of as parts of juice recovered per 100 parts
fibre.
Sucrose recovered and sucrose lost are items for
financial reports, not suitable for judging what the
mills really have been doing.
Using a term as "Lost Absolute Juice % Fibre" as
yardstick for evaluating the effect achieved by a
milling tandem does not give us the complete picture.
We also want to know the different circumstances
which were favourable and unfavourable to obtain
the performance concerned. To obtain an impression
of the circumstances or conditions we added four
more columns to the table showing the figures of lost
juice obtained, viz:
(a) the loading of the tandem indicated as lbs. of
fibre milled per hour and per cu. ft. Total
Roller Volume;
(b) the amount of water used to obtain the result,
indicated per 100 parts of fibre;
(c) the number of imbibition steps indicating how
many times the water and subsequently the
diluted juice was sprayed on the intermediate
bagasses;
(d) the Dilution Ratio indicating the effectiveness
of the water applied in its task to dilute the
residual juice in bagasse.
With regard to the last item, it is the task of the
imbibition water to replace the juice in the cane or at
least to dilute it. It is obvious that the less juice
remains in the bagasse of the first unit, i.e. in the
bagasse before any imbibition is applied, the better
the dilution effect of the imbibition will be. A good
Mill
PG
UF
EM
FX
EN
AK(o)
AK(n)
DK
GD
DL
GH
MV
TS
ME
1L
RN
SZ
UK
MH
UR
LB
MR
LostAbs.
Juice
% Fibre
40
33-i
3Si
34-jr
42
44
33-i
45
44
29
40i
42
30
25
34
44
57
45
46
43
44
53
Specific
Feed
Rate
36
29
59
37
50
44
38
51
40
50
52
47
43
47
46
59
58
37
40
68
57
67
Imbibition
%
Fibre
265
270
325
271
223
273
319
245
287
375
195
238
212
271
292
187
203
227
204
186
216
165
Number
of Imb.
Steps
5
5 + 6
5
5 + 5
4
5
6
5
4
5
5
4
5 + 6
6
5
4
4 + 2
5
5
5
5
4
Dilution
Ratio
75
81
78
82
68 .
69
87
72
66
80
74
80
82
87
80
71
56
73
70
66
70
58

12
and properly squeezing first unit is, therefore, of
primary importance for a high imbibition effect and
subsequently for a good final result of the milling
tandem.
However, what has been explained with regard to
the first unit, holds also for the subsequent mills,
each mill should squeeze out as much (diluted) juice
as possible in order to promote the effect of the next
application of imbibition liquid.
Diffusion
In the case of diffusion we cannot apply a repeated
squeezing action every time water or diluted juice
is sprayed on the bagasse. To obtain the same dilution
ratio as with mills, it is therefore necessary to apply
more "imbibition steps". If we want to exceed the
dilution ratio obtained with mills, we will have to
apply a greater number of steps than with milling. If
two mills precede the diffusion plant, five steps are
used and where only one crusher or one mill precedes,
nine steps are necessary. Finally, if the cane enters
the diffusion plant prepared only by cutting instruments,
and no squeezing action has been applied
beforehand, eighteen "imbibition steps" are required
to excel pure milling work. These eighteen steps
should be compared with the six steps which can be
applied in a 21-roller milling tandem.
Returning to our table, it reveals that the percentages
of Lost Absolute Juice range from 25 to 58 per
cent, the specific feed rates from 30 to 75 lbs. fibre
per hour, per cu. ft. Total Roller Volume and the
imbibition rates from 170 to 380 per cent on fibre.
We repeat: Though Lost Absolute Juice % Fibre
may indicate the effect, we can only gauge the efficiency
of the mill operation if we also know the "circumstances"
availing when the lost juice percentage was
obtained.
The following factories recorded the lowest, i.e. the
best Absolute Juice percentages this season:
Mill
ME
DL
TS
UF
New AK
IL
FX
Lost Absolute
Juice % Fibre
25
29
29|
33i
33+
34
34}
Specific Feed
Rate
47
50
43
29
38
46
37
Proceedings of The South African Sugar Technologists' Association—March 1966
Imbibition
% Fibre
271
375
212
270
319
292
271
Mount Edgecombe's tandem did not only obtain
the best lost juice percentage, but it also achieved the
biggest improvement compared with the previous
season, viz. 25 per cent this year against 39 per cent
lost juice in the previous year.
A worthwhile improvement was shown by Glendale's
tandem; reducing lost juice from 51 to 44 per
cent, after replacing the two-roller crusher by a
shredder and a three-roller crusher; changing the last
mill for a bigger one.
Melville's change of milling train brought a drop of
7| units in the lost juice percentage, which improvement
would have been bigger if the crusher had not
for part of the season to be by-passed in connection
with the repair of the mill turbine.
Tongaat also showed an appreciable drop in lost
juice, viz. from 35 to 30 per cent lost juice. The same
can be said about Felixton where the lost juice percentage
improved from 39 to 35, and also Empangeni
Mill where the drop was from 43 to 39 per cent.
In order to illustrate the progress made in recent
years by the S.A. Mills with regard to milling performance,
the following table shows the average Lost
Absolute Juice percentages and the circumstances
under which they were achieved for the last seven
seasons:
Season Averages
for
South African Mills
1965/66
1964/65
1963/64
1962/63
1961/62
1960/61
1959/60
Lost Abs.
Juice
% Fibre
37.58
36.98
37.47
37.36
38.96
42.03
43.00
Specific
Feed
Rate
46
47
46
42
41
45
46
Imbibition
V
So
Fibre
261
256
288
266
253
238
218
There were in the past season nine Mills with a
higher, six with a lower and two Mills with the same
lost juice percentage as the previous season. This
explains why, notwithstanding that some Mills showed
a material improvement, the weighed average is higher
than that of the 1964/65 season.
Performance of the First Squeezing Unit
Returning to the subject of the importance of a high
extraction of juice by the first squeezing unit, i.e. the
crusher mill, it should be said that the amount of
juice the first unit can squeeze out depends not only
on such items as feeding, mill setting, hydraulic pressure,
etc., but also on the cane composition, i.e. the
ratio between fibre and juice.
The following table shows the minimum juice extraction
to be obtained by the first unit for different
compositions of the cane.
Note: "Ratio" indicates "parts of juice per 100 parts of fibre".
Cane Composition
Fibre
12%
13%
14%
15%
16%
17%
Ratio
733
669
614
567
525
488
Ratio
After
1st Unit
257
246
238
227
220
209
Parts of Juice squeezed out:
per 100 parts
juice in cane
65
63
61
60
58
57
per 100 parts
of cane
57
55
53
51
49
47
These minimum juice extraction figures show how
the fibre content of the cane affects the extraction and
the amount of residual juice left in the bagasse after

Proceedings of The South African Sugar Technologists' Association—March 1966 13
Proceedings of The South African Sugar Technologists' Association—J i—March 1966 13
the three-roller crusher. The longer the tandem is,
i.e. the more units it comprises, the smaller the influence
of the cane composition will be on the percentage
lost absolute juice in final bagasse. For
example for a 18-roiler tandem, the effect of the
difference between 12 and 17 per cent fibre in the cane,
is a difference of only 1.1 per cent. Here again, we
see an advantage of the use of "Lost Juice % Fibre in
Final Bagasse" as a yardstick for evaluating milling
train performances, viz. that it is only slightly effected
by the composition of the cane.
II. Regarding Boiling House Operations
(a) Purity
Sugar technologists the world over learnt by experience
how to use the purity of the juice as a guide
when trying to predict "recovery". We emphasize
"trying to predict" because though purity is the best
guide we have, it is far from a perfect guide. After all,
what is purity? Purity is the percentage ratio between
the sucrose content and the specific gravity; the latter
indicated as degrees Brix. Not a very precise indicator
for purity! However, there is no better one, but that
does not imply that we should close our eyes to its
shortcomings.
(b) The Winter Formula
Experience taught the sugar technologist that the
purity of the juice had a great influence on the quantity
of sugar to be recovered from a certain amount of
sucrose in juice. As it is the non-sucrose in the juice
which causes that part of the sucrose which is not
recoverable. Winter (1897) derived the following formula
to predict the recoverable sugar:
S — 0.4 (B — S)
in which formula S stands for the sucrose and B for
the Brix in juice. The formula indicated the amount
of Standard Muscovado to be expected; Standard
Muscovado being an assumed sugar with 95.1 per
cent crystal.
As processing technique improved, the Mills started
to make more sugar than indicated by the Winter
formula. It was then decided that the formula should
from then on not indicate the amount of Standard
Muscovado, but the quantity of crystal in sugar to
be expected.
Winter derived his formula statistically from hundreds
of returns of Java Mills. Actually the formula
says that each part of non-sucrose (B — S) in mixed
juice makes 0.4 part sucrose not crystallisable. We
draw attention to this fact because authors — not
conversant with the background of Winter's formula—
have stated that the constant 0.4 pointed to
an assumption of a final molasses purity of 28.57° by
Winter. This is incorrect, in two ways. Firstly Winter
did not arrive at the factor 0.4 by assuming a molasses
purity of 28.57° and secondly the assumption that
the factor is so strongly correlated as the formula:
Final Molasses Purity
Factor =10o_Final Molasses Purity
would indicate, is incorrect.
The latter formula assumes either that all nonsucrose
in mixed juice will end up in the final molasses,
or that all non-sucrose present in mixed juice will
leave the factory in the form of products of 28.57°
purity, viz. either as filter cake, final molasses, or
undetermined. Both presumptions are incorrect. In this
context we refer to what will be said about the nonsucrose
present in mixed juice under the heading of
"Non-sucrose Account". Here it will be explained
that during processing non-sucrose is removed, nonsucrose
is added and non-sucrose is formed and that
the non-sucrose in final molasses has a quite different
composition than the original non-sucrose present in
mixed juice. An indirect proof is that those Java defecation
factories which achieved 99^- and 100 per cent
"Winter", had final molasses (gravity) purities far
above 28.57°, to wit from 32.8° to 34.7°.
(c) Boiling House Performance
When Natal introduced in 1950 the Winter formula
for calculating the to be expected recoverable crystal in
sugar, it was felt that with a view to Natal conditions
a more lenient factor than 0.4 should be applied; for
example 0.5. In addition it was preferred not to use a
constant factor, but a factor adjusted to juice purity.
This preference was based on the experience that lower
juice purities were attended by lower final molasses
purities and therefore a lower factor could be applied
than where higher juice purities were concerned.
(d) Practical Application
However when applying the B.H.P. formula with
its variable factor in practice, it should never be forgotten
that the original Winter formula was based on
results with juices obtained from healthy and mature
cane, stripped from tops and trash and processed
within a day after cutting. When the composition of
the juice deviates from the normal one, previous
experience based on normal juices is not applicable
any more.
This now was the case in the past season: the
abnormal non-sucrose composition of the juice made
it impossible to predict "recovery" based on previous
experience with normal juices.
The abnormal composition of the cane juices was
attributed to the detrimental effect of the prolonged
drought. However, there is still another circumstance
which affected the composition of the non-sucrose
adversely: Green tops left on short cane have a
greater detrimental effect on the mixed juice quality
than where stalks of normal length but with the tops
left on, are milled. The prolonged drought caused the
cane to be short and because the cane was short there
was a tendency not to shorten it any further by removing
the top. In some cases the cane was so short
that the tops had to be left on, otherwise it could not
have been handled.
After these introductory remarks with regard to the
abnormal non-sucrose composition of the past season's
juices, the results achieved in the boiling houses
of the different mills will be discussed at the hand of
a number of tables. The main table shows in addition

14
to the Boiling House Performance figures the following
parameters:
(i) the purity of the final molasses,
(ii) the sucrose lost in final molasses,
(iii) the undetermined lost sucrose,
(iv) the mixed juice purity, and
(v) the reducing sugars/sucrose ratio of mixed
juice.
The purities of mixed juice and final molasses determine
the quantity of final molasses, while the final
molasses purity also determines the sucrose lost in
final molasses. The final molasses purity has therefore
a two-fold effect on the sucrose lost in final molasses
and subsequently on the magnitude of the B.H.P.
Name of
Mill
PG.
UF.
EM
FX
EN.
AK (o)
AK (n)
DK
GD
DL.
GH
MV
TS .
ME
IL .
RN
SZ .
UK
Mean
MH
LTR
LB
MR
Boiling
House
Perfor.
96.44
95.15
95.58
94.49
93.79
95.47
94.31
95.58
97.49
96.16
94.82
93.21
95.35
96.70
96.10
95.10
95.55
96.30
95.65
95.39
95.44
95.13
95.88
Final Molasses
Purity
38.79*
40.96
40.48
39.41
40.60
41.18
38.98
41.93
36.80 s
40.14
40.60
41.41
41.13
36.69
39.57
38.34*
39.44
39.90
39.91
39.83*
39.80
38.94
40.75
Sucrose
Lost
9.29
10.01
11.04
9.96
10.09
10.48
12.01
8.39
9.25
9.72
11.44
11.40
10.52
10.49
8.39
8.85
8.83
8.42
9.98
10.32
11.16
8.78
8.65
* Apparent Purity
Undeterm.
lost
sucrose
0.91
1.63
1.02
1.80
2.64
1.88
2.24
2.24
1.47
1.75
1.76
1.84
1.26
1.23
2.14
3.24
1.98
1.95
1.63
1.86
0.51
1.93
2 02
Proceedings of The South African Sugar Technologists' Association—March 1966
Mixed Juice
Purity
84.63
83.95
83.91
84.40
86.29
83.32
81.09
85.05
84.00
83.51
83.65
84.36
83.97
82.63
85.08
85.05
85.45
86.73
84.22
84.40
83.25
85.79
85.38
'Ratio
3.11
3.04
4.02
4.17
2.89
3.70
5.07
3.46
3.80
3.69
4.09
3.70
3.97
4.32
4.51
3.00
3.48
—
3.73
—
5.31
3.57
3.85
N.B.—The sucrose losses are not per 100 sucrose in cane, but
per 100 sucrose in mixed juice, in order to eliminate any differences
in extraction.
The sucrose lost in final molasses of SZ and UK is estimated
by assuming a non-sucrose account factor equal to 0.85
(SZ and UK do not declare their final molasses weights).
When the complete analysis of the final molasses
is at hand, we can calculate with the aid of the Douwes
Dekker formula the target purity. Comparing the
target with the actual purity will reveal if the final
molasses was properly exhausted. However, even
when we obtain or even exceed the target purity, this
does not always imply that we could not have obtained
a lower final molasses purity. We are referring here to
the fact that when reducing sugars are destroyed by too
high alkalinities, we raise both, viz. target and actual
purity. We should therefore always draw up a reducing
sugar balance in addition to calculating the target
purity.
If no complete final molasses analysis is at hand,
also the reducing sugars/ash quotient of the molasses
will give us an indication of the exhaustability of the
molasses.
The main table shows that the B.H.P. figures range
from 93.21 to 97.49 per cent, with a mean for the
Natal factories of 95.65 per cent. Latter percentage
we want to compare with that of the 1957/58 season
when the highest mean B.H.P. was obtained, attended
by the lowest average final molasses purity, on record.
In the following table the relevant data of these
two seasons are compared. In order to make the comparison
easier the third line of the table contains the
figures which would have been achieved in 1957/58,
if the mixed juice purity had been 84.2° and the undetermined
loss 1.63 per cent, as in the past season.
The fourth line finally shows the conversion completed,
as here the B.H.P. and the sucrose lost in
final molasses are shown when mixed juice and
molasses purities and the undetermined sucrose loss
had been the same in 1957/58 as they were in 1965/66.
Season
1957/58 .
1965/66 .
Convert. 1
Convert. II
B.H.P.
98.45
95.65
97.45
96.94
Final
Molas.
Purity
38.52
39.91
38.52
39.91
Sucrose lost in:
Final Mol.
7.97
9.98
8.45
8.96
Undeter.
1.11
1.63
1.63
1.63
Mixed
Juice
Purity
85.1
84.2
84.2
84.2
This gradual conversion revealed that it was not
only the higher undetermined losses and molasses
purity which caused the lower B.H.P., but also the
quantity of molasses, which was in 1965/66 relatively
higher than in 1957/58, i.e. higher than commensurate
with the lower mixed juice purity.
In order to be able to discuss the main table, it
is necessary to throw more light on the different processing
stages. It is therefore that a number of "ancillary"
tables have been drawn up. The first table to
be shown and discussed is the non-sucrose account
table, followed by the reducing sugar account table,
the non-sucrose circulation table and two tables
showing the exhaustion of the different strikes.
Non-sucrose Account
In the previous section it has been stated that the
non-sucrose account factor was higher in 1965/66

Proceedings uf The South African Sugar Technologists' Association—March 1966 15
than in 1957/58. We will therefore start the present
section with a review of these factors for recent years,
after having given a definition of this factor.
The factor or the non-sucrose ratio is the ratio
between "the non-sucrose in total final molasses"
and "the non-sucrose originally present in mixed
juice". Total final molasses is the final molasses as
weighed plus the (final) molasses film around the
crystals of the bagged sugars.
Non-sucrose Account for Recent Years for S.A. Factories
1965/66 0.85
1964/65 0.83
1963/64 0.79
1962/63 0.85
1961/62 0.81
1960/61 0.81
1959/60 0.81
1958/59 0.81
1957/58 0.79
1956/57 0.80
The table shows that the ratio was in 1965/66
7.6 per cent higher than in 1957/58 which implies
that in the past season relatively 7.6 per cent more
final molasses was obtained than commensurate with
the juice purity, than in 1957/58.
The following table shows the individual nonsucrose
ratios of all factories as obtained in the
1965/66 season.
Mill
PG
UF
EM
FX
EN
AK(o)
AK(n)
DK
Ratio
0.84
0.83
0.89
0.87
0.89
0.79
0.83
0.79
NON-SUCROSE ACCOUNT
Mill
GD
DL
GH
MV
TS
ME
IL
RN
Ratio
0.88
0.81
0 87
0.92
0.85
0.91
0.82
0.84
Mill
sz
UK
MEAN
MH
UR
LB
MR
N.B. The ratios of SZ and UK arc assumed to be 0.85
Ratio
0.85
0.85
0.85
0.89
0.92
0.93
0.78
The table reveals that in 1965/66 the "ratios"
ranged between 0.78 (MR) and 0.93 (LB). This does
not imply that at MR 22 per cent and at LB 7 per cent
of the non-sucrose in mixed juice has been removed
by the clarification process. Firstly there is not only
non-sucrose removed but also non-sucrose formed and
non-sucrose added during the clarification process.
Secondly the non-sucrose substances in mixed juice
are of a quite different composition than those present
in total final molasses. Therefore they have a different
effect on the Brix spindle readings and subsequently
on the calculation of non-sucrose in final molasses
compared with that in mixed juice.
With regard to the non-sucrose added, lime, sulphur,
phosphoric paste and Separan should be mentioned
and regarding non-sucrose formation we point
to inversion and destruction of reducing sugars.
Returning to the abnormal high individual ratios,
these can be caused by too high final molasses weights,
formation of non-sucrose (for example owing to inversion),
or a combination of both. The other ancilliary
tables will help to throw a light on this question.
MILL
PG
UF
EM
FX
EN
AK(o) . . . .
AK(n) . . . .
DIC
CD
DL
GH
MV
TS
ME
IL
UN
SZ
UK
MEAN . . .
MH . . . .
UR
LB
MR
REDUCING SUGARS ACCOUNT
Clear Juice
102
98
95
88
111
96
94
—
92
95
88
96
—
84
102
84
98
—
95
116
96
99
PERCENTAGE AVAILABLE IN
Syrup
78
103
90
85
66
88
78
89
75i
91
80
100
80
81
81
J 04
84
—
In the Reducing Sugar Account table the reducing
sugars in mixed juice are taken as 100 and the reducing
sugars present in the other products are expressed as
percentages of those in mixed juice.
The Reducing Sugars Account Table is actually a
more interesting table than the previous one, because
here we can follow the process stage by stage. However,
this is only then possible when the analysis of
the different products are accurate. Perusing the present
table one can only arrive at the conclusion that
some of the analysis are not quite correct.
With a modern juice liming plant, the milk of lime
will be mixed thoroughly with the juice in the shortest
possible time, and no pockets of high alkalinity will
be formed. Accordingly the percentage of reducing
sugars in clear juice should be close to 100 per cent,
provided the juice is not over-limed.
The percentage of reducing sugars in syrup depends
on several factors as pH of the clear juice, temperature
and retention time of the juice in the evaporator, more
in particular time and temperature in the first vessel,
pre-evaporator or vapour cell. In general the time
the juice requires to pass these first vessels is essential,
next is the temperature and the pH. In the preevaporator,
vapour cell or first vessel inversion and
destruction of reducing sugars can take place simultaneously,
even when the pH is correct, but either
the temperature is too high or the residence time too
long, or a combination of these two. Such a phenome-
86
—
81
87
80
Total Final
Molasses
105
106
87
85
98
109
96+
—
98
81
114
105
97
102
—
102
—
99
—
133
43

16
non occurred at MV where a new vapour cell with
automatic pressure control had been installed. Samples
of juice taken after the vapour cell showed discoloration
and drop in purity.
The percentage of reducing sugars in total final
molasses should be compared with the percentage
present in syrup before a conclusion can be drawn.
An exceptional high percentage points to a too high
molasses weight, which should be confirmed by a too
high non-sucrose account factor and a too low nonsucrose
circulation factor. However, it can also be a
result of inversion which should be confirmed by a
higher than usual reducing sugars content of the
molasses or a too high reducing sugars/ash quotient.
A too low percentage for total final molasses in the
reducing sugars account table points to a too low
determination of the reducing sugar content of the
molasses, which can be checked by comparison with
figures of the previous years.
Mill
PG
UF
EM
FX
EN
AK(o)
AK(n)
DK
Percent
100
125
119
124
120
105
143
142
NON-SUCROSE CIRCULATION
Mill
GD
DL
GH
MV
TS
ME
IL
RN
Percent
132
131
119
124
138
117
115
115
Mill
SZ
UK
MEAN
MH
UR
LB
MR
Proceedings of The South African Sugar Technologists' Association—March 1966
Percent
N.B. "Percent" in the table above means the quantity of nonsucrose
in C-massecuite, expressed as a percentage ratio
of the quantity of non-sucrose present in total final
molasses.
Provided the volume of C-massecuite and the final
molasses weight is correct, the percentage of nonsucrose
circulation gives a good indication of the
quality of the (pre-cured) C-sugar. To explain this
statement the following examples are given:
A C-strike, containing 35 per cent crystal, is cured
and turns out a C-sugar of 85°Pol, which corresponds
with a crystal content of the C-sugar of 80 per cent,
i.e. 20 per cent molasses adhere still to the crystal.
To simplify calculations we assume that the bagged
sugar is 100 per cent crystal, i.e. all non-sucrose is
removed from the factory with the weighed final
molasses. The C-m.c. (before curing) comprises 35
parts crystal and 65 parts of final molasses per 100
parts of massecuite. After curing there will be 100 x
35/80=43.75 parts of C-sugar of 85°Pol and 100 —
43.75=56.25 parts of final molasses per 100 parts
of cured C-m.c. The circulation ratio is therefore
100 x 65/56.25 = 116%.
If the same C-m.c. had produced a C-sugar of 86^
per cent crystal (90°Pol), there would be after curing
100 x 35/86^=40.46 parts pf C-sugar and 100 —
—
—
123
—
128
165
135
40.46-- 59.54 parts of final molasses, per 100 parts
of C-m.c. The circulation rate in this case would have
been 100x65/59.54=109%.
These examples should be a guide when perusing
the non-sucrose circulation table. A percentage less
than 100 per cent points to a too low volume of
C-massecuite or a too high molasses weight. Percentages
far above 100 per cent can be caused by poor
curing, a too high C-m.c. volume, a too low final
molasses weight, or a combination of two or more of
these possibilities.
So far we have considered only the magnitude of the
non-sucrose circulation, another question is the
extent of the circulation. If the C-sugar is mingled
into a magma and the magma used as seed for the Aand
B-strikes, the circulation will extend as far as the
A-strikes. If the C-sugar is double cured (and really
cured well!), the non-sucrose circulation will be
limited to the system C-m.c. pan and C-m.c. centrifugals,
provided the wash of the after-curers is returned
to the C-massecuite.
If we return the wash to the B-strike, the nonsucrose
circulation will be extended to the B-strike.
Every increase in the magnitude and every increase
in the extent of the non-sucrose circulation will
increase the "stickiness" of the last strike. We have
to repeat this warning again, because in the past
season we saw again how the natural stickiness of the
C-massecuites was in many cases increased by
violating this rule. We saw a really very poorly cured
C-sugar being mingled into a magma with the aid of
syrup; the wash of the after-curers which handled this
magma being returned to the B-strike. The big quantity
of final molasses which was in this manner recirculated
to the B-strike made the B-massecuite in
its turn "sticky" too.
Some mills stated that they could not see much
improvement in double curing, compared with single
curing of the C-massecuites. In such cases the same
incidental circumstances could be discovered which
explained the lack of improvement. These circumstances
are:
The same melter was used for remelting the C- and
the B-sugar. In addition juice or syrup was used for
dissolving the sugars instead of water. This made
checking on the quality of the double-cured C-sugar
by checking the purity of the melt impossible.
The pre-cured C-sugar was mingled into a magma
with circulating wash, cold and "saturated" with final
molasses. A heater in the mixer above the after-curers
was lacking.
There was not sufficient space between screw conveyor
and discharge opening of the after-curers to
collect a sample in order to check on the quality of
the double-cured C-sugar.
Note: Latter shortcoming is also often noticed in the case of
the fore-curers.
In connection with the exhaustion of the final
molasses the following table is shown which renders
details about the final strike as well as on the final
molasses.

Proceedings of The South African Sugar Technologists' Association—March 1966 17
The annual summaries are not the place to indulge
in rules to be followed and measures to be taken to
obtain low final molasses purities. However, excepttion
should be made for a couple of hints; one about
the purity of the C-massecuites, the other on the
disposal of pan steamings.
It cannot be called a coincidence that the two factories
which recorded the lowest molasses purities
this season, show also the lowest C-m.c. purities.
Neither is it accidental that the mill with the highest
C-m.c. purity has also the highest molasses purity.
This confirms the rule that to obtain a low final
molasses purity we have to start with a low C-m.c.
purity, i.e. a purity well below 60°.
We all know that we should drop the C-m.c. with
the highest practicable Brix, but why then do we dilute
the massecuite immediately afterwards with hot pan
steamings? All pan steamings (and we are not only
referring to those of the C-m.c. pans) should be collected
separately and returned to the weighed mixed
juice tank. To be able to separate the steamings from
the massecuites, opening in the form of a slit should
be cut crosswise in the gutters under the pan discharge
valves. The slit is closed by a cover in the form of
a trapdoor as long as the massecuite passes through
the gutter, but when the steamings start pouring out
of the pan, the opening is uncovered in order that the
steamings can be collected in a separate set of small
gutters, directly beneath the massecuite gutters and
leading to a collecting tank with stirrer and pump.
The table shows the "exhaustion" of the A-, the
B and C-massecuites; "exhaustion" indicating the
parts of crystal recovered in sugar per 100 parts of
MILL
PG
UF
EM
FX
EN
AK(o)
AK(n)
DK
GD
DL
GH
MV
TS
ME
IL
RN
SZ
UK
MEAN
MH
UR
LB
MR
MILL
PG
UF
EM
FX
EN
AK(o)
AK(n)
DK
GD
DL
GH
MV
TS .
ME
IL .
RN
SZ .
UK
RECOVERED CRYSTAL PER 100 SUCROSE
IN MASSECUITE
MEAN . .
MH . . .
UR . . .
LB . . .
MR . . .
DATA REGARDING THE C-MASSECUITES
°Brix
97.8
99.3
99.0
97.2
98.2
97.8
98.2
97.7
98.0
98.6
95.4
96.5
97.8
99.8
95.7
97.2
99.7
97.3
97.8
98.1
99.3
99.7
A-m.c.
60
62
60
62
60
66
69
68
67
62
69
61
64
61
56
63
59
61
63
62
60
62
B-m.c.
62
56
63
57
59
65
65
57
62
59
58
63
58
57
59
60
52
59
60
63
53
60
C-m.c.
59
53
58
54
49
55
61
59
59
55
55
50
56
54
63
60
57
58
56
56
55
59
MEAN
60
57
60
58
56
62
65
61
62
58
61
58
59
57
60
61
56
59
60
60
56
60
sucrose in massecuite. The advantage of expressing
the recovered crystal per 100 parts of sucrose instead
of per 100 parts of massecuite, is that a percentage of
60 per cent or more indicates a good and percentages
below 60 per cent an unsatisfactory exhaustion.
With regard to UF and UK the following should be
mentioned: Both factories follow a four-boiling
"Purity
60.9
59.3
60.2
59.5
58.8
60.1
60.1
63.2
56.9
60.0
60.2
58.6
59.5
55.9
61.2
60.7
59.4
59.2
59.7
59.8
58.2
58.9
C-massecuite
% Crystal
35.3
31.1
34.3
31.2
28.3
32.7
36.2
36.2
31.1
32.3
31.6
28.3
32.4
30.4
37.7
35.4
33.6
33.6
33.9
32.6
31.7
34.8
°Brix
93.4
93.8
93.6
91.9
92.0
90.8
88.6
91.4
93.7
92.5
89.7
90.6
90.8
93.8
92.2
92.7
91.5
90.0
91.7
92.2
91.0
92.5
90.7
-inal Molasses
"Purity
38.8
41.0
40.5
39.4
40.6
41.2
39.0
41.9
36.8
40.1
40.6
41.4
41.1
36.7
39.6
38.3
39.4
39.9
39.9
39.8
39.8
38.9
40.8
R.S./Ash
Quotient
0.68
0.87
0.83
0.86
1.18
1.03
0.98
1.45
0.84
1.37
1.16
1.02
—
1.32
—

18
system, bagging the sugars of the first two strikes only.
In the event of UF these two strikes are a remelt- and
an A-strike, in the case of UK are an A- and a Bstrike,
both using C-sugar as pied-de-cuite. (The Cstrike
is boiled on a magma of D-sugar according to
the double-magma system applied at U.K.) The percentage
shown under "A-m.c." is the weighted average
of the exhaustions of the first two strikes.
It should be mentioned too, that an exhaustion of
60 per cent or more in the event of a C-massecuite
is not always attended by a low molasses purity. To
obtain the latter the purity of the C-m.c. should be
low too.
New Amatikulu had a good start with a mean
exhaustion of 65 per cent achieved mainly owing to
the high exhaustion of the A-massecuite. That it was
not only due to the new installation, i.e. better pans
and modern centrifugals is confirmed by the fact that
old Amatikulu obtained also a high mean. It would
have been even higher when the exhaustion of the
C-massecuite had not been so very low (low gravity
centrifugals).
Glendale, another mill with a mean = 62 per cent
shows a better exhaustion in the C-massecuite than
old Amatikulu and good exhaustions in the A- and
the B-massecuites. However, it should be stated that
owing to lack of pan capacity, the A- and B-boilings
were very finely grained.
All ancillary tables being shown, we can return now
to the main table at the beginning of the section
regarding boiling house performance.
Opening the discussion with the mill with the highest
boiling house performance, viz. GD with a B.H.P. of
97.5 per cent, the main table reveals a low loss of
sucrose in the final molasses; low with regard to the
purity of the juice. The undetermined sucrose loss is
"average". The sucrose loss in molasses is low, not
because the quantity of molasses is low, but because
the purity is low. Actually the amount of final molasses
is high with regard to the juice purity, according to
the non-sucrose account factor (0.88!).
The molasses purity was low because the C-m.c.
purity was low. The molasses purity could have been
even lower if the massecuite was dropped tighter.
Notwithstanding the rather low Brix of the C-m.c. the
curing could have been better according to the nonsucrose
circulation percentage (132!). If the heating
system of the m.c. had been more efficient the m.c.
could have been boiled tighter and the curing been
better.
The mill following GD in B.H.P. is ME, also with
a low molasses purity. The sucrose loss in molasses
is higher than that of GD because the juice purity is
lower and the non-sucrose account factor is even
higher than that of GD, i.e. 0.91 against 0.88. The
high factor made ME check the molasses scale, however
it appeared that the molasses weight was correct.
This is also confirmed by the reducing sugars account
(97 per cent for total final molasses) and the nonsucrose
circulation (117 per cent).
As GD and ME use the same type of C-m.c. centrifugals,
the lower non-sucrose circulation percentage
Proceedings of The South African Sugar Technologists' Association—March 1966
of ME (notwithstanding the higher Brix of the C-m.c.)
should be attributed to the better (improved) heating
system applied to the C-m.c. at ME.
SZ with 96.55 per cent B.H.P., Pongola with 96.44
per cent B.H.P. and UK with 96.30 per cent B.H.P.
recorded final molasses purities below 40", viz. 39.44°,
38.79° and 39.90° respectively.
Another mill with a B.H.P. above the average is
DL with 96.16 per cent B.H.P. Final Molasses purity
is nothing to write home about, but notwithstanding
this fact, the sucrose loss in final molasses compares
favourably with the loss of GD, when the lower mixed
juice purity is taken into account. Why is this? The
non-sucrose account reveals a low factor (0.81) compared
with the (high) factors which the other mills
show this season. The reducing sugars account points
in the same direction; no non-sucrose formation
during processing. With regard to the final molasses
purity, if the C-m.c. purity had been lower, we might
have expected a lower molasses purity.
In looking for an explanation for the low B.H.P.
of EN, the main table shows a high undetermined
loss in addition to a sucrose loss in final molasses
which is high with respect to the high (the highest)
purity of mixed juice. The reason for the latter loss
being high, are the facts that as well as the quantity
the purity of the final molasses is high. That the
quantity is high is revealed by the high non-sucrose
account (Factor=0.89) and that something abnormal
took place is revealed by the reducing sugars account.
With regard to the high molasses purity the ancillary
table reveals a very low crystal content of the C-m.c.
A high non-sucrose account is shown by LB too,
i.e. a factor of 0.93, which is accompanied by the
highest reducing sugars account, i.e. 133 per cent
recorded. This points either to molasses formation
owing to inversion, or to a too high final molasses
weight. Confirmation of inversion is to be found in
the high reducing sugars content (18.24 per cent) of
the final molasses, rendering a glucose/ash quotient of
18.24/13.77-1.32.
We could go on and on looking at figures in the
main table and trying to explain them with data
collected in the ancillary tables. Actually, the ancillary
tables should be extended with another table revealing
the overall time efficiency, because a poor time efficiency
reduces recovery and the time efficiency was —
in general — low this season. A further extension
should be a Brix Balance to be compared with the
Sucrose Balance, the Non-sucrose Account, and the
Reducing Sugars Account. However, each Chief
Chemist can do this for his own factory.
III. Factory Control
It was felt that no definite and far-reaching conclusions
could be drawn from the single series of figures
obtained at Entumeni during the 1964/65 season
when the bagasse was weighed as well as calculated
from the Fundamental Equation:
Weight Cane + Weight Imb. Water
=Weight Mixed Juice -f- Weight Bagasse

Proceedings of The South African Sugar Technologists' Association—March 1966 19
It was therefore suggested — after consultation
with Entumeni's Management — that the test should
be repeated in the 1965/66 season after special measures
had been taken. These measures comprised the
replacement of steam ejectors by centrifugal pumps
to transport imbibition liquids and introduction of a
tagging system for cane consignments for more accurate
stocktaking. In addition the Sugar Industry
Central Board would issue weekly bulletins, not only
showing the daily differences in weighed and calculated
bagasse, but also such useful items as:
(i) The number of cane bundles which were
crushed the same day, stored one day or stored
two days between weighing and crushing.
(ii) Daily records of the rainfall, the temperature,
the humidity and the hours of sunshine.
It is obvious that these special records about storing
time combined with the information about weather
conditions made it possible to acquire an insight into
the loss of cane weight in the millyard due to evaporation.
The new test covered 22 weeks or 110 crushing days.
With the exception of those days when the cane
had become wet after weighing, owing to rain, each
day showed the same result, viz. the weighed bagasse
tonnage was lower than the bagasse weight by inference.
(With the exception of a couple of days when
the stocktaking of the cane in the millyard had been
less accurate than usual.)
The final result of the 110 days long test was that
the weighed bagasse weight was 7.3 per cent lower
than the bagasse weight by calculation.
In the first instance it was feared that the cane weighbridge
was out of order causing a 2.4 per cent too
high cane weight. However, a check by the S.A. Scale
Company showed that the weighbridge was in order.
Re-weighing of cane bundles after 24 hours in the millyard
revealed that they lost 1£ per cent in weight due
to evaporation. Other losses are to be attributed to
loss of mud, dead sticks, etc., in the millyard and
evaporation of water during the milling process.
Though we may conclude from the second Entumeni
Test that bagasse weight by inference is in general
too high, we may not conclude that it is everywhere
7.3 per cent too high. The percentage will differ
according to circumstances, viz. time elapsing between
weighing and crushing, method of storage (in trucks
or on the ground), weather conditions, and number
of units in the milling train.
It has also been proved without doubt that Pennink
was correct when he doubted the correctness of the
results derived from the Fundamental Equation in
practice and wanted the cane weight to be determined
by inference from bagasse, juice and water weights
(H.G. PENNINK "Control Methods"; ARCHIEF
1893; Vol. I, pp. 68/80).
Comparing boiler efficiency percentages derived
from weighed bagasse with B.E. results, where the
bagasse weight was determined by inference had
shown in Java that the former were always about 10
per cent higher than the latter. Consequently doubt
about the practical results of the Fundamental
Equation arose again. We know now that this doubt
was correctly founded.
Bagasse weighing will be applied in the case of
diffusion as owing to excessive evaporation, the
bagasse weight cannot be determined by inference.
However, as in the days of the Petree and Dorr process
mud juice will have to be measured, sampled and
analysed, where mud from the diffusion-clarifier circumvents
the mixed juice scales on its way to the
boiling house rotary filters.
This is not as easy as it sounds. The mud is hot and
is a diluted sucrose solution; therefore care should
be taken by sampling to prevent evaporation and
deterioration. Taking snap samples seems to be the
best solution.
However, the mixed juice sample will also be of elevated
temperature, which reminds us of the difficulties
encountered at Z.S.M. when the mixed juice still was
heated before weighing. A dosing or metering pump
to take the juice sample and store it in a closed, watercooled
container would be a solution.
21st April, 1966

20
MILL
Darnall . .
Amatikulu .
Felixton
Empangeni
Mt. Edgecombe
Tongaat . .
Melville. . .
Illovo . . .
Umfolozi .
Glendale . .
Reynolds .
Crookes
Pongola
Gledhow . .
Umzimkulu
Doornkop .
Entumeni .
TOTAL
REPORTED
PRODUCTION .
White
—
—
—
—
—
—
190.0000
—
—
—
64,288.0000
19,049.4000
32,116.5000
75,849.3000
—
—
9,099.9000
200,593.1000
Proceedings of The South African Sugar Technologists' Association - March 1966
TABLE I
FINAL PRODUCTION 1965-1966 SEASON
(AS REPORTED TO SA.S.A. BY THE MILLS)
(ALL MILL WEIGHTS)
RAWS FOR REFINING
To Refinery
29,517.1600
37,691.2000
63,924.0000
62,861.0000
—
85,219.1970
11,634.6750
9,450.7250
—
7,657.5210
18,268.0000
10,008.0800
—
503.8500
960.0000
—
—
337,695.4080
To Terminal
—
4,771.0000
8,694.0000
7,098.0000
—
1,176.0715
—
—
—
—
—
—
—
—
—
—
—
21,739.0715
Short Tons
RAWS FOR EXPORT
General Export
and
Jap. Assort.
62,755.6400
—
—
—
—
13,484.6160
—
51,791.8350
103,729.8575
35.7700
—
—
—
—
—
22,593.6500
—
254,391.3685
Canadian
Assortment
—
—
—
—
72,219.4000
—
6,232.5250
—
—
—
—
_
—
—
—
—
1,775.2000
80,227.1250
Golden Brown
462.6000
11,603.5000
980.0000
6,113.0000
._
414.6840
70.5000
3,476.9700
258.9314
448.3500
3,765.0000
7,124.6000
32,208.1500
80.0000
36,072.3500
2,668.5000
1,390.5000
107,137.6354
TOTAI,
92,735.4000
54,065.7000
73,598.0000
76,072.0000
72,219.4000
100,294.5685
18,127.7000
64,719.5300
103,988.7889
8,141.6410
86,321.0000
36,182.0800
64,324.6500
76,433.1500
37,032.3500
25,262.1500
12,265.6000
,..„„„..,
1,001,783.7084

SYMBOLS
INDICATING FACTORIES
TONS OF CANE CRUSHED .
CANE COMPOSITION
Tons Cane per Ton Sugar
Tons Cane per Ton 96° Sugar
CANE VARIETIES
Co.33l
N:Co.3IO
N:Co.292
N:Co.293
N:Co.339
N:Co.376
N:Co.382
N:50'2U
Remainder
TOTAL RAINFALL (or 1%5 (ins ) 15.84
TONS OF SUGAR MADE .
Percentage of Whites made
Average Pol of All Sugars made ,
TIME ACCOUNT
Time Crushed % Time Mills Ope i 91.71
Cane Shortage % Time Mills Ope i 1.77
THROUGHPUTS
Tons Sugar per hour (act. cr.)
Table 2—CANE CRUSHED, SUGAR MADE, CANE VARIETIES AND THROUGHPUTS
PG
8.7.65
26.2.66
561.855
13.62
14.42
78.73
8.73
8.46
0.04
91.14
0.13
0.91
0.12
6.94
nil
0.42
0.30
64,325
50%
99.13=
PG Pongola S.M.C. Ltd. DL
UF Umfolozi Co-op. S.P. Ltd. ME
EM Hulett's Empangeni Mill EN
FX Hulett's Felixton Mill DK
AK(o) Hulett's Old Amatikulu Mill GD
AK(n) Hulett's New Amatikulu Mill GH
131.60
18.97
19.98
15.07
UF
1.5 65
16 12.65
924.410
13.20
13.87
79.80
8 88
8.73
1.58
83.67
0.35
0.03
1.65
8.41
3.84
0.31
0.16
23.02
104,065
nil
97.71'
89.28
4.55
209.41
29.04
31.18
23.57
EM
21 5 65
22.12.65
686.397
13.41
16.36
76.30
9.02
8.77
0.38
77.82
0.13
0.10
1.95
14.46
1.91
3.18
0.07
36.44
76,072
nil
98.75°
88.20
3.90
179.40
29.35
26.79
19.88
FX
21 5 65
15.1.66
687,489
12.74
15.80
78.72
9.34
9.08
1.17
38.95
0.39
nil
1.12
41.04
8.20
1.92
7.21
52.05
73,598
nil
98.80°
84.81
1.56
169.74
26.82
24.22
18.17
Hulett's Darnall Mill MV
Hulett's Mount Edgecombe Mill TS
Entumeni S.M. Co. (Pty) Ltd. IL
Doornkop S. Co. (Pty) Ltd. RN
Glendale Sugar Millers SZ
Gledhow Sugar Co. Ltd. UK
EN
19 5 65
6.12.65
114,486
13.17
14.23
77.35
9.33
8.99
4.77
9.35
0.40
40.68
1.95
29.01
1.18
12.53
0.07
34.15
12,266
74%
99.61°
94.96
3.82
28.09
4.00
4.02
3.01
AK(o)
4 5 65
12.11.65
394.544
12.98
15.62
77.40
9.43
9.18
0.24
82.39
0.15
0.10
0.10
12.63
0.27
4.10
0.02
30.07
41,855
nil
98.54°
93.72
2.54
105.78
16.52
15.25
11.22
AK(n)
19 II 65
25.1.66
129,398
11.75
17.43
77.92
10.60
10.28
0.11
77.63
nil
nil
nil
18.14
0.13
3.99
nil
30.07
12.211
nil
98.94°
62.42
4.82
159.52
27.81
21.64
15.05
DK
14 5 65
1.12.65
222,691
13.47
15.36
77.61
8.79
8.64
2.05
21.86
1.89
22.89
0.70
41.79
1.52
6.34
0.96
29.70
25,326
nil
97.73°
90.91
8.49
68.00
10.44
9.94
7.73
GD
28 6 65
24.12.65
78.299
12.42
15.90
77.53
9.62
9.35
12.92
26.01
nil
3.27
1.04
52.36
1.02
3.32
0.06
25.23
8,142
nil
98.72°
77.46
13.10
31.32
4.98
4.30
3.26
DL
14 5 65
31.12.65
817 661
13.19
15.69
77.91
8.82
8.66
3.25
55.46
0.26
1.09
3.04
25.79
1.38
3.79
5.94
33.86
92,735
nil
97.79°
91.14
5.51
192.07
30.14
28,93
21.28
GH
26 5 65
9.12.65
762,695
12.68
16.17
77.41
9.68
9.40
1.34
19.27
0.55
1.86
2.65
52.49
1.80
4.28
15.76
30.67
78,776
99%Ref.
99.94°
93.18
2.80
202.36
32.72
28.67
20.90
Identity of the Symbols used for identification of the Factories
Melville Sugar Estates MH
Tongaat Sugar Co. Ltd. UR
Iliovo Sugar Estates Ltd. LB
Crookes Bros. Ltd. Renishaw MR
Reynolds Bros. Ltd. Sezela TR
Umzimkulu Sugar Co. Ltd.
MV
14 8 65
27.1.66
184.763
12.13
16.26
77.%
10.19
9.94
7.80
22.73
0.16
0.14
5.35
47.63
1.25
14.81
0.13
30.90
18,128
nil
98.38=
Mhlume (Swaziland) Sugar Co. Ltd.
Ubombo Ranches Ltd.
Sena Sugar Estates, Luabo
Sena Sugar Estates, Marromeu
Hulett's (Rhodesia) Triangle Ltd.
75.78
11.32
73.00
11.88
9.76
7.17
TS
11 5 65
4.12.65
940,663
12.63
16.20
77.06
9.38
9.17
0.96
13.54
0.90
1.85
2.80
27.00
7.90
10.16
34.89
31.62
100,220
nil
93.19=
88.45
7.38
235.65
38.18
33.59
25.10
ME
12 5 65
15.12.65
676,516
12.50
16.13
77.78
9.37
9.12
4.17
9.68
1.63
4.60
1.43
28.53
4.80
2.25
42.91
37.80
72,219
nil
98.65°
89.59
6.00
175.67
28.33
25.51
18.75
IL
12 5 65
29.1.66
553,223
13.49
14.89
79.09
9.55
9.38
12.08
14.07
0.37
32.78
1.15
29.97
6.01
0.99
2.61
41.05
64,680
nil
97.80°
85.68
9.31
115.94
17.27
17.53
13.56
RN
1 7 65
2.4.66
357,007
12.60
17.42
77.49
9.87
9.55
1.92
11.44
3.35
1.25
1.90
60.68
8.85
7.08
3.45
43.49
36,182
53%Wh.
99.19°
85.35
6.86
74.89
13.04
10.14
7.59
SZ
25.3.66
830,751
13.17
14.%
78.82
9.35
9.08
3.60
11.40
3.18
4.38
1.44
59.82
nil
nil
16.18
48.42
88,834
74%Ref.
98.89°
84.27
5.62
186.68
27.93
26.21
19.96
UK
? 8 65
2.3.66
343,476
12.97
15.62
78.54
9.28
8.96
3.50
24.68
0.78
5.05
1.80
63.54
0.05
n I
0.54
38.35
37,032
nil
99.40°
86.63
6.80
91.41
14.28
12.67
9.86
Totals and
Weighed
Means
1 5 65
2.4.66
9,266,324
12.99
15.57
78.07
9.20
8.97
2.70
40.15
0.89
4.51
1.76
32.19
3.35
3.52
10.93
33.64
1.006,665
20%
98.49°
87.41
5.84
138.11
21.51
20.02
15.00
NOTE. The official sugar production of the 1965/66 season, according to the S.A.
Sugar Association's Records is equal to 1,001,784 tons fel quel.
The difference between this tonnage and the tonnage shown in this Table (No. 2) is
caused by the fact that GH and SZ do not declare the actual bagged sugars, but
make known the rawsugars passing from the rawhouse — to the refineries —
departments of their factories-cum-refineries.
For the actually produced sugar tonnages of SZ and GH we should look at Table 1.

Table 3—SUCROSE BALANCE, ANALYSIS OF BAGASSE, JUICES, CAKE AND SYRUP
SYMBOL INDICATING FACTORY PG UF EM FX EN AK(o) AK(n) DK GD DL GH MV TS ME RN SZ UK
SUCROSE BALANCE
Lost in BAGASSE (A) 5.70 4.72 6.60 5.51 6.32 7.51 6.35 7.01 7.22 4.63 6.52 7.02 5.20 3.98 4.58 8.60 8.92 7.30
Lost in FILTER CAKE (B) 1.38 0.85 0.55 0.38 0.74 0.50 0.83 0.58 0.20 0.35 0.55 1.10 0.79 0.52 0.59 0.57 0.98 0.68
Lost in FINAL MOLASSES 1(D) 16.70 16.66 18.41 17.00 18.98 19.44 20.53 17.47 17.36 15.92 19.41 20.43 17.16 15.75 15.22 20.22 19.73 17.35
OVERALL RECOVERY 83.30 83.34 81.59 83.00 81.02 80.56 79.47 82.53 82.64 84.08 80.59 79.57 82.84 84.25 84.78 79.78 80.27 82.65
BOILING HOUSE PERFORMANCE 96.44 95.15 95.58 95.76 93.79 95.47 94.31 95.58 97.49 96.16 94.82 93.21 95.35 96.70 96.10 95.10 95.55 96.30
Boiling House Recovery 88.33 87.47 87.35 87.84 86.95 87.10 84.86 88.75 89.07 88.17 86.21 85.58 87.39 87.74 88.45 87.28 88.12 89.16
LOST ABSOLUTE JUICE % FIBRE 39.78 33.52 38.67 34.60 41.86 44.41 33.56 44.98 43.62 29.39 40.63 41.75 29.70 25.12 34.15 44.46 57.30 44.64
Imbibition % Fibre 265 270 325 271 223 273 319 245 287 375 195 238 212 271 292 IB6 203 227
Specific Feed Rate 36 29 59 37 50 44 38 51 40 50 52 42 43 47 46 59 58 37
SUCROSE EXTRACTION 94.30 95.28 93.40 94.49 93.68 92.49 93.65 92 99 92.78 95.37 93.48 92.97 94.80 96.02 95.42 91.40 91.08 92.70
Imbibition % Cane 38.15 37.46 53.23 42.77 31.75 42.61 55.63 37.60 45.60 58.88 31.48 38.67 34.44 43.73 43.54 32.48 30.31 35.46
FINAL BAGASSE
Sucrose % Bagasse 2.35 1.97 2.32 1.90 2.71 2.73 1.96 2.62 2.58 1.77 2.23 2.26 1.87 1.42 1.81 2.68 3.33 2.55
Moisture % Bagasse 53.11 53.37 54.02 54.60 50.23 52.63 51.42 53.76 50.69 51.% 53.19 53.77 51.31 51.81 53.48 53.49 53.15 54.50
Fibre % Bagasse 43.60 43.92 42.78 42.72 46 29 43 75 45 86 42.63 45.76 45 51 43 57 43.11 47.21 46.08 43.70 43.05 42.39 42.13
L.C.V. of Bagasse (btu, lb) 3019 3003 2941 Z898 3Z61 3054 3172 2958 3224 3129 3014 2964 3183 3148 2997 2980 2998 2895
Bagasse%Canc 33.06 31.58 38.24 36.98 30.75 35.70 38.01 35.93 34.75 34.49 37.11 37.73 35 07 35 00 34 08 40.45 35.29 37.08
Imbibition Efficiency 48 44 48 42 30 46 42 46 41 34 52 42 52 41 47 49 68 52
BRIX-FREE WATER % FIBRE 29 25 27 20 46 24 8 29 26 21 21 22 28 18 20 16 21 25
FIRST EXPRESSED JUICE
Degree Brix
Degree Purity (Apparent)
LAST EXPRESSED JUICE
Degree Brix
Degree Purity (Apparent)
19.98
86.59
2.79
71.33
MIXED JUICE
Degree Brix 14.44
Degree Purity (Apparent) 84.70
Degree Purity (Gravity) 84.63
Red. Sugars Sucrose Ratio 3.11
19.14
86.40
2.10
72.89
14.15
83.95
3.04
20.40
86.17
2.69
72.53
18.81
86.06
1.97
70.86
19.41
87.90
1.91
78.05
19.47
86.08
2.96
75.43
12.99 13.48 14.16 13.49
18.00
83.76
2.10
72.06
CLARIFIED JUICE
Degree Brix 15.10 13.88 12.43 12.37 14.51 13.06 10.85
Degree Purity (Apparent) 85.63 85.19 84.73 85.06 86.20 83.97 81.96
Red. Sugars Sucrose Ratio 3.22 3.02 3.34 3.69 3.22 3.56 4.83
Average PH 7.20 7.10 7.20 7.00 — 7.40 7.30
FILTER CAKE
Sucrose % Cake
Filter Cake % Cane
3.77
5.01
2.24
5.00
83.91
3.50
1.32
5.58
84.40
4.17
0.75
6.50
86.29
2.89
1.96
5.00
SYRUP
Degree Brix 62.32 60.06 64.89 55.98 60.57
Degree Purity (Apparent) 84.40 85.48 84.86 85.26 86.10
Red. Sugars Sucrose Ratio 2.46 3.17 3.17 3.56 1.93
Average pH 6.90 6.60 6.40 6.50 —
83.22
3.70
0.79
8.26
54.43
84.58
3.26
6.90
11.54
81.09
5.00
1.54
6.33
57.50
82.50
4.01.
6.40
19.85
87.41
2.90
72.62
14.40
85.72
85.88
3.46
14.90
86.63
7.10
1.56
5.00
58.32
86.71
3.11
6.50
18.71
85.60
2.67
72.70
12.38
83.99
3.80
12.09
84.60
3.51
7.20
0.85
3.00
62.07
84.90
2.88
6.90
19.68
86.04
1.58
70.06
12.11
83.51
83.51
3.69
11.57
84.25
3.52
7.50
0.82
5.72
60.58
84.48
3.38
6.80
19.10
85.50
2.97
68.70
15.01
83.65
4.09
14.21
84.50
3.62
7.40
1.14
6.11
54.31
84.70
3.31
7.20
18.17
85.60
2.31
72.30
19.10
85.79
2.38
75.69
13.24 14.34
84.36
3.70
12.99
85 00
3 60
7.30
2.68
5.00
56.75
84 60
3.74
6.70
83.97
3.97
13.70
84.60
7.20
1.99
5.00
60.80
85.00
3.21
6.20
18.84
85.30
1.62
67.40
13.36
82.63
4.32
12.35
83.39
3.66
7.60
0.83
7.89
60.42
83.14
3.51
7.00
19.59
87.03
2.36
64.33
13.82
84.73
85.08
4.51
13.49
86.51
4.61
7.50
1.87
4.24
63.01
85.57
3.67
6.60
18.72
86.80 87.58
3.00
77.00
14.71
85.05
3.00
16.08
86.20
2.53
1.49
4.84
53.35
86.20
3.14
5.22
75.10
14.78
85.12
85.45
3.48
14.98
85.71
3.44
7.20
18.77
87.96
3.00
75.67
14.09
86.52\
86.73/
13.99
87.92
7.30
2.33 1.65
4.00
59.93
86.03
2.97
6.70
55.95
87.94

SYMBOLS INDICATING MILLS
Tons Brix in Mixed Juice
100
A-MASSECUITE
B-MASSECUITE
Puritv of Massecuite
C-MASSECUITE
Cu. ft. per Ion Brixi ,
Brix of Massecuite
Purity of Massecuite
Purity of C-Moiasses
Drop in Purity
EXHAUSTION
CRYSTAL % MASSECUITE
TOTAL Cu. ft. of MASSECUITES
FINAL MOLASSES
CLARIFYING AGENTS
Per 1,000 tons Cane:
Per ton Cane:
Table 4—DATA REGARDING MASSECUITES AND RUNOFFS, AND CONSUMPTION OF CLARIFYING AGENTS
PG
15.17
. . 27 00
92.08
85.12
69.35
. . . 15 77
. . . . 60.44
9 27
. . . . 94.80
73.68
. . . 51.76
. . 21 92
61 67
8.56
97.85
60.90
38.79
22.11
59.31
35.34
44 84
. . . . 93.37
38.79
3.62
—
. . . . . 1.39
nil
nil
0.40
UFS
14.98
33.18
93.27
88.08
73.88
14.20
61.72
9.85
96.69
71.42
52.50
18.92
55.77
8.79
99.26
59.28
40.71
18.57
52.83
31.09
68.97
51.82
93.85
40.71
40.96
10.89
16.02
0.68
3.62
-
1.18
nil
nil
0.71
EM
14.93
28.54
93.00
85.86
70.84
15.02
59.99
12.21
95.40
74.07
51.28
22.79
63.15
9.15
99.01
60.23
39,13
21.10
57.55
34.32
67.22
49.90
93.55
39.13
40.48
12.13
13.96
0.87
4.02
1.12
nil
nil
0.63
EXHAUSTION = Parts of Recovered Crystal per 100 Parts of Massecuite.
i Per ton of Brix present in Mixed Juice.
a Al AK(n), DK, MV and IL all commercial sugar was provided by the A-strikes.
3 Indus. White Sugar Massecuite 89,24 cu. ft. per ton Sugar.
4Indus. White Sugar Massecuite 66.85 cu. ft. Der Ion Brix.
FX
14.27
28.23
92.66
85.60
69.53
16.07
61.62
8.63
94.22
70.77
51.02
19.75
56.98
9.26
97.21
59.46
40.26
19.20
54.05
31.24
61.46
46.11
91.90
40.26
39.41
13.16
15.82
0.83
3.58
-
0.79
nil
nil
1.64
EN
14.30
24.76
91.28
85.30
69.90
15.40
60.00
10.35
94.55
71.40
50.60
20.80
59.00
7.63
98.24
58.80
42.10
16.70
49.00
28.30
57.06,
42.74,
92.04
42.13
40.60
9.48
3.48
3.76
1.61
nil
nil
AK(o)
14.42
20.08
93.24
84.52
65.29
19.23
65.55
11.15
95.97
73.45
49.44
24.01
64.65
8.07
97.75
60.91
41.26
19.65
54.92
32.70
53.43
39.30
90.83
40.74
41.18
12.09
14.03
0.86
3.59
-
1.00
nil
nil
1.42
AK(n)2
13.57
28.65
93.44
86.44
66.66
19.78
68.63
11.06
96.48
71.05
46.33
24.72
64.83
10.64
98.17
60.13
36.80
23.33
61.39
36.24
72.64
50.52
88.58
36.59
38.98
15.48
13.09
1.18
3.99
,—
1.30
nil
nil
nil
DKa
14.62
28.40
93.81
88.52
71.45
17.07
67.54
10.33
94.04
74.05
55.19
18.86
56.84
9.71
97.66
63.22
41.60
21.62
58.56
36.15
62.25
48.42
91.43
41.60
41.93
13.57
13.23
1.03
2.97
—
1.08
nil
19.26
2.81
GD
13.72
21.92
93.25
83.51
62.80
20.71
66.66
13.55
95.88
71.54
48.57
22.97
62.43
9.42
97.98
56.90
36.78
20.12
55.92
31.18
59.25
44.90
93.66
36.78
3.41
_
1.09
nil
nil
1.73
DL
15.06
31.06
93.30
85.80
69.90
15.90
61.60
10.04
95.60
71.60
51.00
20.60
58.70
9.64
98.60
60.00
40.50
19.50
54.60
32.30
67.39
50.74
92.51
40.14
12.58
12.86
0.98
3.58
-
1.13
nil
nil
ni
GH
14.17
23.13
93.40
87.80
69.00
18.80
69.10
13.27
94.53
72.40
52.40
20.00
58.00
9.73
95.37
60.20
40.50
19.70
55.00
31.60
63.29
46.14
89.67
40.60
10.54
3.72
9.24
0.94
(5.24)
0.97
nil
nil
MV.
13.37
32.43
91.96
86.20
70.90
15.30
61.00
13.37
94.65
72.10
48.90
23.20
63.00
9.66
96.52
58.60
41.40
17.20
50.10
28.30
75.58
55.47
90.58
41.41
12.49
3.66
-
2.18
0.58
nil
7.04
TS
14.26
20.27
93.40
84.30
65.80
18.50
64.20
11.41
95.20
72.20
52.10
20.10
58.10
10.11
97.80
59.50
39.40
20.10
55.70
32.40
55.91
41.79
90.76
39.40
41.13
14.50
10.00
1.45
3.60
-
1.27
nil
nil
0.85
ME
14.52
30.81
94.50
84.90
68.60
16.30
61.10
15.40
97.30
69.10
49.20
19.90
56.70
8.86
99.80
55.90
36.60
19.30
54.40
30.40
74.92
55.07
93.79
36.65
36.69
13.08
15.56
0.84
4.05
4.17
1.30
nil
1.83
1L,
15.12
42.75
90.58
87.92
76.04
11.88
56.39
13.86
92.25
76.28
56 66
19 62
59.34
8.79
95.73
61.21
36.05
25.16
64.28
37.66
81.05
63.01
92.20
36.16
39.57
16.74
12.20
1.37
3.21
—
1.59
nil
nil
8.36
RN
13.54
24.63
91.21
85.70
69.10
16.60
62.70
13.39
94.91
72.10
51.10
21.00
59.60
8.23
97.19
60.70
38.20
22.50
60.00
35.40
62.11
46.27
92.74
38.34
3.13
3.69
1.64
IBS
nil
SZ
14.04
22.65
92.94
84.33
68.77
15.56
59.04
13.70
96.20
70.06
52.75
17 31
52.29
10.23
99.72
59.44
38.85
20.59
56.65
33.58
61.17
46.58
91.51
39.17
39.44
15 01
12.94
1.16
3.16.
6.14
0.73
(6.14)
1.30
nil
nil
UKS
13.86
33.56
92.48
86.44
71.53
14.91
60.90
11.18
94.65
72.68
52.10
20.58
59.11
8.0 5
97.34
59.21
37.73
21.48
58.26
33.56
67.88
52.79
89.98
37.73
39.90
2.96,
5UF as well as UK applied a four'boiling system; bagging the sugars from the first two
strikes. The data shown under "A-MASSECUITE" are the weighed averages of these
first two strikes.
At UF a remelt strike preceded the regular A-m.c, but at UK the double-magma system
was applied.
e Estimated.
0.57
nil
126
131
Mean
14.49
27.89
92.77
85.91
69.41
16.50
62.78
11.78
95.18
72.22
51.27
20.95
59.53
9.14
97.84
59.70
39.26
20.44
56.37
33.92
65.12
48.81
91.72
"1-39.91
13.57
13.61
1 00
3 59
Defe-
1.18
nil
1 53

24 Proceedings of The South African Sugar Technologists'' Association—March 1966
Table 5—DATA OF LUABO AND MARROMEU, MHLUME AND UBOMBO RANCHES
NAME OF FACTORY Luabo
25.5.1965
19.10.1965
TONS CANE CRUSHED 462,228
TONS SUGAR MANUFACTURED . . . 56,313
Tons Cane crushed per hour 158
Overall Time Efficiency 95
Percentage White Sugar made 86.00
SUCROSE % CANE 14.53
FIBRE % CANE 13.90
Tons Cane per Ton Sugar 8.21
Brix of First Expressed Juice 20.89
Purity of First Expressed Juice 87.84
LOST ABSOLUTE JUICE % FIBRE . . 44.13
Specific Feed Rate* 47.33
Imbibition % Fibre 216.00
SUCROSE EXTRACTION 94.08
Imbibition % Cane 30.07
Sucrose % Bagasse 2.72
Moisture % Bagasse 52.30
Bagasse % Cane 31.67
Lower Calorific Value (btu/Ib)t . . . . 3,083
SUCROSE BALANCE
Lost in Bagasse
Lost in Filter Cake
Lost in Final Molasses
UNDETERMINED LOSSES . . .
Total of All Sucrose Losses . .
OVERALL RECOVERY
BOILING HOUSE PERFORMANCE
Boiling House Recovery
PURITY OF MIXED JUICE . . .
Reducing Sugars/Sucrose Ratios:
in Mixed Juice
in Clear Juice
in Syrup
in Final Molasses
Filter Cake % Cane
Sucrose % Filter Cake
DENSITY OF SYRUP (°Brix) . . .
FINAL MOLASSES
Degrees Brix
Gravity Purity
Apparent Purity
Molasses (85°Brix) % Cane . . .
Exhaustion of A-Massecuite . . . .
Exhaustion of B-Massecuite . . . .
Exhaustion of C-Massecuite . . . .
CU- FT. M.C. PER TON BRIX
A-Massecuite
B-Massecuite
C-Massecuite
5.92
0.71
8.26
1.82
16.71
81.04
95.13
88.54
85.79
3.57
3.13
18.24
4.87
2.11
Marromeu
17.5.1965
20.10.1965
575,531
69,205
154
94
40.00
14.64
13.85
8.32
20.18
87.00
42.78
67.00
165.00
92.07
22.81
3.67
51.56
31.63
3,129
7.93
0.97
7.96
1.86
18.72
83.29
95.88
88.28
85.38
3.85
3.21
6.68
4.00
3.54
Mhlume
6.5.1965
2.1.1966
474,741
55,015
104
89
4.00
13.96
13.32
8.63
20.33
86.65
45.53
39.50
204.00
93.59
27.23
2.92
52.78
30.68
3,037
6.41
0.34
9.66
1.74
18.15
81.85
95.39
87.45
Ubombo R.
1.5.1965
14.1.1966
553,960
59,968
127
89
8.00
13.20
15.40
9.24
19.56
85.38
50.73
68.00
186.00
93.23
28.68
2.64
50.73
33.79
3,219
6.77
0.84
10.87
0.48
18.96
81.28
95.44
86.92
84.40 83.25
2.94
1.63
5.31
6.21
4.33
5.00
2.20
60.10 64.20 49.19 64.20
92.46
38.94
38.42
3.63
59.90
53.20
54.90
32.10
16.30
11.10
90.69
40.75
38.07
3.37
62.30
59.90
59.30
23.77+
12.23
7.86
*Lbs of fibre milled per hour and per cu. ft. TOTAL ROLLER VOLUME
fL.C.V. = 7650 - 18.0 S - 86.4 M; S = Sucrose % Bagasse and M = Moisture % Bagasse.
f-Exclusive Millwhite Massecuites.
92.37
39.83
3.99
91.00
39.80
36.50
3.26
61.50
62.60
55.60
22.10J
11.30
10.60

Table 6—AVERAGE MANUFACTURING RETURNS BY MONTHLY PERIODS FOR
SOUTH AFRICAN CANE SUGAR FACTORIES (SEASON 1965/1966)
END OF MONTHLY PERIOD May 29
1965
TONS CANE CRUSHED
TONS SUGAR MADE AND ESTIMATED
TONS CANE PER HOUR
SUCROSE % CANE
FIBRE % CANE
TONS CANE PER TON SUGAR
LOST ABSOLUTE JUICE % FIBRE
IMBIBITION % FIBRE
SUCROSE EXTRACTION
SUCROSE % BAGASSE
MOISTURE % BAGASSE
BOILING HOUSE PERFORMANCE
BOILING HOUSE RECOVERY
OVERALL RECOVERY
PURITY OF MIXED JUICE
REDUCING SUGARS SUCROSE RATIO
SUCROSE IN FINAL MOLASSES % SUCROSE IN CANE , . .
UNDETERMINED LOST SUCROSE % SUCROSE IN CANE . .
GRAVITY PURITY OF FINAL MOLASSES
Month
To-date :
Month :
To-date :
Month
To-date :
Month :
To-date :
Month :
To-date :
Month
To-date :
Month
To-date :
Month
To-date :
Month
To-date :
Month
To-date :
Month :
To-date :
Month
To-date :
Month :
To-date i
Month
To-date :
Month
To-date :
Month
To-date :
Month :
To-date :
Month :
To-date :
Month
To-date :
431,671
48,504
MOLASSES (85° Brix) % CANE Month : 3.95
To-date : —
MONTHLY RAINFALL IN INCHES 2.10
TOTAL RAINFALL FROM JANUARY 1ST 9.93
166
13.49
15.88
8.90
36.30
262.00
93.92
2.25
52.82
95.29
86.99
81.70
83.06
4.61
10.01
1.55
38.86
June 26
1965
867,142
1,308,589
98,295
147,702
129
141
13.47
13.48
15.78
15.81
8.82
8.86
34.28
35.14
268.00
266.00
94.46
94.28
2.18
2.23
52.57
52.65
95.22
95.24
87.29
87.19
82.45
82.20
83.96
83.66
3.92
4.05
9.33
9.76
2.14
1.76
39.78
39.57
3.88
3.90
4.31
14.33
July 31
1965
1,301,506
2,610,095
147,938
295,640
134
137
13.43
13.45
15.54
15.68
8.80
8.83
34.52
34.84
268.00
267.00
94.52
94.42
2.01
2.13
52.95
52.80
95.72
95.48
88.06
87.60
83.23
82.71
84.61
84.13
3.52
3.79
9.22
9.49
1.50
1.66
40.57
40.07
3.55
3.71
1.40
15.90
August 28
1965
1,144,231
3,754,326
129,701
425,341
142
139
13.54
13.48
15.52
15.63
8.82
8.83
36.09
35.22
261.00
265.00
94.34
94.38
2.16
2.14
52.77
52.79
95.69
95.54
87.87
87.70
82.90
82.77
84.56
84.26
3.52
3.71
9.41
9.47
1.47
1.58
40.24
40.12
3.68
3.70
3.28
18.84
October 2 October 30
1965 1965
1,392,277
5,146,603
154,470
579,811
137
138
13.13
13.38
15.52
15.60
9.01
8.88
38.10
36.00
274.00
267.00
94.10
94.30
2.19
2.15
52.79
52.79
95.48
95.52
87.84
87.74
82.66
82.75
84.90
84.43
3.22
3.40
8.86
9.31
1.95
1.66
40.42
40.20
3.40
3.59
2.82
21.68
1,158,197
6,304,800
126,595
706,406
138
137
13.01
13.32
15.37
15.55
9.15
8.93
36.%
36.10
256.00
264.00
94.14
94.27
2.17
2.15
52.91
52.81
96.10
95.62
88.16
87.81
83.00
82.78
84.52
84.44
3.40
3.40
8.72
9.20
1.55
1.66
39.99
39.86
3.21
3.52
4.68
26.34
Nov. 27
1965
1.188,812
7,493,612
123,580
829,986
138
137
12.36
13.16
15.24
15.51
9.62
9.03
38.08
36.30
252.00
262.00
94.69
94.34
2.12
2.15
53.09
52.86
95.69
95.64
87.61
87.78
82.96
82.81
83.93
84.37
3.90
3.47
9.50
9.41
1.63
1.49
39.94
40.10
3.34
3.49
4.05
30.39
January 1 January 29
1966 1966
1,002.866
8,496,478
102,969
932,955
139
137
12.41
13.07
15.69
15.53
9.74
9.11
37.%
36.50
277.00
263.00
93.62
94.27
2.10
2.14
52.44
52.81
95.75
95.65
88.59
87.56
83.00
82.88
83.66
84.29
4.27
3.63
9.30
1.46
39.21
40.00
4.00
3.55
3.19
33.64
436,851
8,933.329
43,753
976,708
151
138
12.23
13.04
16.06
15.55
9.98
9.14
45.60
37.00
243.00
263.00
92.39
94.10
2.34
2.17
53.28
52.83
95.59
95.65.
87.75
87.91
81.08
82.72
83.87
84.31
4.56
3.69
9.01
9.34
1.45
1.37
39.35
39.96
3.55
3.57
6.66
6.66
Feb. 26
1966
229,013
9,162,342
21,168
997,876
122
138
11.52
13.00
15.85
15.56
10.82
9.18
50.58
37.37
222.00
261.00
91.12
94.03
2.72
2.19
54.20
52.97
95.84
95.71
86.86
87.70
79.14
82.47
83.40
84.25
4.53
3.72
9.78
9.35
1.11
1.53
38.36
39.92
3.58
3.24
9.91
April 2
1966
103,981
9.266,324
8,788
1,006,665
116
138
11.29
12.99
16.49
15.57
11.83
9.20
57.00
37.58
202.00
261.00
89.70
93.99
2.90
2.20
53.67
52.98
93.30
95.65
83.03
87.67
74.46
82.40
80.69
84.22
6.02
3.73
9.38
1.53
37.48
39.91
3.59
0.68
10.54

Table 7—COMPARISON OF FINAL RESULTS FOR S.A. SUGAR FACTORIES
(Season 1955 to Season 1965 inclusive)
SEASON 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
CANE
Sucrose per cent 13.87 13.35 13.11 13.12 13.66 13.69 13.75 13.29 13.55 13.90 12.99
Fibre per cent 15.74 15.81 15.38 15.92 15.92 15.22 14.52 15.50 15.50 15.38 15.57
IUICE
Brix per cent First Expressed Juice 19.80 20.30 19.60 19.20 19.60 20.10 19.60 19.70 19.80 20.30 19.27
Purity of First Expressed Juice 88.00 87.30 87.30 86.70 87.70 87.80 88.00 85.50 87.40 87.50 86.30
Purity of Last Expressed Juice 76.70 75.80 76.10 74.40 75.00 75.60 74.70 71.50 72.70 74.30 72.30
Purity of Mixed Juice 86.00 85.50 85.10 84.50 85.50 85.60 86.00 83.40 85.30 85.50 84.20
Reducing Sugars/Sucrose Ratio 3.40 3.30 3.70 4.30 3.50 3.30 3.30 5.10 3.40 3.30 3.70
MILLING DATA
Imbibition per cent Fibre 204.00 222.00 224.00 207.00 210.00 238.00 253.00 266.00 258.00 256.00 261.00
Lost Absolute Juice per cent Fibre 45.50 42.10 40.90 42.30 43.00 42.00 39.00 37.40 37.50 37.00 37.60
Imbibition per cent Cane 32.10 35.20 34.50 32.90 34.60 36.20 36.70 41.20 39.80 34.40 40.60
Sucrose Extraction 92.30 92.90 93.40 92.90 92.90 93.40 94.20 94.20 94.10 94.20 94.00
Sucrose per cent Bagasse 2.91 2.60 2.47 2.55 2.66 2.60 2.43 2.24 2.29 2.34 2.20
Moisture per cent Bagasse 53.18 53.12 53.06 53.28 53.26 53.01 52.54 52.17 52.46 52.64 52.98
Lower Calorific Value 3,003 3,014 3,021 3,000 3,000 3,023 3,067 3,105 3,066 3,061 3,033
RECOVERIES
Overall Recovery 83.60 83.40 84.40 83.10 83.00 83.40 84.50 82.70 84.30 84.40 82.40
Boiling House Recovery 90.50 89.80 90.40 89.50 89.40 89.40 89.70 87.80 89.60 89.60 87.70
Boiling House Performance 97.90 97.40 98.50 97.80 97.10 96.90 97.00 96.60 97.20 97.10 95.60
Tons Cane per Ton Sugar 8.53 8.88 8.95 9.09 8.74 8.70 8.54 9.01 8.63 8.38 9.20
FILTER CAKE
Sucrose per cent G.ke 1.18 1.12 1.03 1.30 1.57 1.66 1.63 1.27 1.37 1.30 1.57
Cal

Proceedings of The South African Sugar Technologists' Association—March 1966 27

The South African Sugar Technologists' Association—March 1966
When I listen to Mr. Perk's paper my mind goes
back to when I started the factory data reporting
system in 1925, with the co-operation of twelve
factories out of twenty-five. It was all very secretive,
however, and instead of factory names, numbers
were used. This lasted for a few years until one day
Tongaat proudly painted the figure 5 on the wall of
the factory.
Mr. Boyes: Mr. Perk, do you know if a climbing
film type evaporator causes sucrose decomposition?
I agree with separate steamings of third boilings.
Although Entumeni showed a higher weight for
weighed bagasse compared to the calculated weight,
Tongaat showed just the opposite, no doubt because
there was no loss of cane weight due to a short period
of intermediate storage. Our difference in weight was
thought to be caused by unweighed water going onto
the milling tandem. My point is that the figures
obtained at Entumeni do not necessarily apply
throughout the industry as other factors must be
considered and therefore it is wrong to calculate
sucrose differences for all factories based on results
from this one mill alone.
Mr. Perk: Climbing film evaporators were introduced
in order to reduce the residence time of the
juice at high temperatures and therefore there is no
sucrose decomposition.
Regarding separate steamings, a separate set of
gutters should be fitted under the vacuum pans to
collect the steamings, which are stirred until all the
grain is dissolved and then returned to the weighed
mixed juice tank.
Mr. Lenferna: Are the heat transfer coefficients
determined in sugar house practice?
Mr. Perk: In investigations carried out at Mackay
Research Institute not only were heat transfer coefficients
determined but also the composition of the
scale in the different tubes of the pre-heaters.

Proceedings of The South African Sugar Technologists' Association—March 1966 29
BOILER DESIGN AND SELECTION IN THE CANE SUGAR
INDUSTRY
1.00 Properties and Uses of Steam
1.1 Properties
Steam is a colourless odourless gas produced by
adding heat to water. Its physical properties have been
determined experimentally and published in the form
of steam tables. There are slight differences between
the results obtained by various workers, but these do
not materially affect the design of steam generating
plant.
Calendar's 1939 steam tables (1) shown graphically
in Fig. 1.1 illustrate the following imporant
characteristics:
(a) The saturation temperature of steam rises
continuously with pressure.
(h) The total heat of dry saturated steam reaches a
maximum at about 465 p.s.i.a. The proportion
of latent heat to total heat falling constantly
until at the critical pressure (3,206.2 p.s.i.a.) no
latent heat is added at all.
(c) The specific volume of steam decreases with
pressure until it equals that of water at the
critical pressure.
1.2 Uses
In a sugar factory steam is used primarily for:
(a) Generating power.
(b) Concentrating sugar juices.
Set out in table 1.1 are the pressure and temperature
conditions normally encountered in a modern factory,
power generating conditions are usually determined
by the size and type of machine whilst process conditions
are limited by a combination of the carameli­
Electrical Power
Generation
Mill Drives
Process
By N. MAGASINER
TABLE 1.1
Steam Conditions in a Cane Sugar Factory
sing temperature and pressure vessel design factors,
as well as overall factory thermal balance requirements.
2.00 Fuels
The economic viability of the cane sugar industry
largely depends upon the use of bagasse as a fuel to
generate power and process steam. In a well balanced
raw sugar factory the quantity of bagasse available
should be just sufficient to meet the total energy load.
Unbalanced factories may experience a short fall or
surplus of bagasse in which case either costly auxiliary
fuels have to be used or costly bagasse disposal
techniques employed. Where off crop outside power
and irrigation loads are high, these can be met either
by increasing factory thermal efficiency and stock
piling baled bagasse for use during the off crop or by
burning an auxiliary fuel. This choice is dependent
upon local conditions.
Scheduled in table 2.1 are the chemical and physical
properties of typical fuels used in the industry.
Bagasse and hogged wood have similar characteristics,
but differ radically from the other two. On a dry
basis their chemical characteristics, as with most
fibrous fuels, are almost identical, i.e. they have low
ash and high volatile contents. Physically in the "as
fired" condition they have high moisture contents,
and low calorific values and bulk densities.
The average Natal bituminous coal is free burning,
has a high ash fusion temperature (plus 1,400° C) and
exhibits reasonable swelling characteristics. Sulphur
content is low whilst its calorific value is fairly high
(11,500 — 12,500 B.T.U/lb). Its bulk density is about
eight times that of bagasse, while from 15-20 times
Use Equipment Steam Conditions Remarks
Turbo alternator of:
back pressure,
pass out/condensing or
condensing design
a) Electrical
b) Steam Turbine
c) Reciprocating
Steam Engines
250-900 p.s.i.g.
600-850° F.
300-450 p.s.i.g.
650-750 F.
100-250 p.s.i.g.
sat-550° F.
Evaporators, Juice heaters, up to 40 p.s.i.a. sat.
pans, etc.
Power is generated by expanding H.P. steam
down to process conditions, i.e. 30-40 psia sat.
Where condensing facilities are included, these
cater for balancing electrical and steam loads
and/or meeting off crop power demands.
Power obtained from main turbo-alternator
station.
Small horse powers preclude higher steam
conditions. Due to high efficiencies, pressure
reducing and desuperheating plant required to
balance factory load.
High capital cost of plant and foundations and
oil entrainment in steam have tended to make
this type of prime mover obsolete.
Since the heat transfer co-efficient of saturated
steam is about 10 times higher than superheated
steam, superheated conditions should be avoided.

30
Property Bagasse
Chemical properties
1) Proximate analysis from mill
"as fired"
Carbon % 11.5
Volatiles % 37.0
Water % 50.0
Ash % 1.5
Proceedings of The South African Sugar Technologists' 1 Association—March 1966
TABLE 2.1
Chemical and Physical Properties of Fuels used in Cane Sugar Industry
100.0
2) Ultimate analysis
"as fired"
Carbon % 22.5
Hydrogen % 3.0
Sulphur % trace
Nitrogen % —
Oxygen % 23.0
Phosphorus % —
Moisture % 50.0
Ash % 1.5
100.0
3) Gross Calorific Value G.C.V.
(BTU/lb.) 4,108
4) Net Calorific Value N.C.V.
(BTU/lb.) 3,298
5) Theoretical weight of air required
for complete combustion
per 10,000 BTU on G.C.V. (lbs) 6.37
6) Max. theoretical CO. in flue
gases measured by Orsat % . 20.70
7) Practical excess air requirements
for complete combustion % . 48
8) Corresponding % CO. in flue
gas as measured by Orsat % . 14.0
9) % moisture by weight in flue
gases at above CO;. % . . 15.95
10) Practical boiler efficiencies
(100,000 pph units) with heat
recovery equipment designed to
bring final gas temperatures
down to 450' : F at the above
excess air figures at 80 : F
ambient air temperatures .
Losses
a) Unburnt carbon in ashes, grits
and stack discharge . . . % 3.00
b) Dry gas loss "/ g. 71
c) Wet gas loss °/' (1,265-48)/
G.C.V.
Varies from 0.1 to 0.2 depending upon
humidity.

Proceedings of The South African Sugar Technologists' Association—March 1966 31
Property
Losses—continued.

32
Constituent
C
H,
S
N,
o2
H.O
Ash
Parts by wt.
0.225
0.030
—
—
0.230
0.500
0.015
1.000
Oxygen in fuel . . . .
Oxygen required
Weight of air required
Weight of Nitrogen
Theoretical dry gas %C02
Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE 3.1
Combustion Reactions of Bagasse
Weight of Oxygen
required
0.600
0.240
—
—
—
0.840
0.230
0.610
0.610
0.232
6.74
-• x 100
32.54
large storage and thermal inertia characteristics can
be installed. Fig. 9. 2B illustrates a typical example of
this type of unit. Should the bagasse supply fail,
continuous steaming can be maintained for a period
of some 10 to 20 minutes thus providing a reasonable
time for bagasse to be reclaimed from store to maintain
load, or in the event of a bagasse carrier failure to
enable auxiliary power equipment to be brought on
line.
Whilst most of the ash produced in this type of
unit is disposed of while the boiler is on range through
collectors in the boiler itself, the furnace must be shut
down at weekly intervals to be manually cleaned. The
shutdown can be timed to coincide with the normal
weekly factory shutdown. The furnace is extremely
simple, has no moving parts and possesses self feeding
characteristics which simplifies auto-control.
The state of the fuel bed is quiescent in relation to
suspension firing which reduces grit carryover and
smut emission considerably. Grit collectors can be
dispensed with whereas they are considered essential
with suspension firing.
2.62 lb.
2.01 lb.
Product
co2
H20
so2
N.,
—
H.O
Ash
- 20.7%
Weight of
product
0.825
0.270
2.010
—
0.500
0.015
Orsat vol.
of product
in ft. 3 at 32° F.
6.74
_
—
25.80
—
—
32.54
2) From Avagadro's Hypotheses at a given pressure and
temperature all gases have the same number of molecules
per ft 3
i.e. PV - KT where in the f.p.s. system K = 10.7/mol.wt.
At 32° F, 30" Mercury the volume of 1 lb. of:
COjis 8.157 ft'
SO, is 5.61 ft 3
N.is 12.81 ft 3
Air is 12.385 ft'
4.2 Coal Firing
There are a number of different ways of firing coal.
The choice of equipment being dependent upon the
fuel characteristics and the size of the unit.
Where unit capacities do not exceed 250,000 p.p.h.
South African bituminous coals are best burnt on
carrier bar stokers, of the type shown in Fig. 9.2D,
having ratings of from 33-38 lb. of coal burnt per
hour per ft. 2 of active grate area. Above 250,000 p.p.h.
an economic case can be made out for pulverising the
fuel down to the consistency of a face powder and
firing it in suspension using a portion of the combustion
air as the conveying medium.
If coal and bagasse are to be fired simultaneously,
they must be intimately mixed and the fuel bed must
not be allowed to build up to more than a thickness of
6". This can best be achieved with the equipment
shown in Fig. 9. 2C.
Grit carry-over with spreader firing is high and
reasonably high efficiency collectors should be installed
to avoid a grit carry-over nuisance.

Proceedings oj The South African Sugar Technologists' Association—March 1966
4.3 Oil Firing
Of the fuels used fairly extensively in the sugar
industry, oil is probably the simplest to burn efficiently.
The type of combustion equipment required
depends upon its grading, i.e. whether it is a light or
a heavy oil and the size of the combustion chamber,
but in general consists of a nozzle, through which the
oil can be introduced and atomised, surrounded by a
chamber through which the combustion air is introduced.
Atomising can be achieved by introducing the
oil through tiny orifices at high pressure or by means
of a mechanical rotor or by steam injection. Fig. 9.2E
illustrates a typical oil fired boiler.
ignored. This point is very important in practical
boiler design.
5.2 Convection
Heat transfer by convection is due to fluid motion.
Cold fluid receives heat from a fluid at a higher
temperature by mixing with it. Free or natural
convection occurs when the fluid motion is not implemented
by mechanical agitation, but merely by
differences in fluid densities. When, however, a fluid
is mechanically agitated, heat is transferred by a
process known as forced convection.
Convection heat transfer in boilers is usually forced,
natural convection playing only a very small part
which is usually neglected. The rate of heat transfer
by forced convection from a hot gas flowing at a
constant mass rate to a cold tube of uniform diameter
has been found to be influenced by the velocity v,
density p, specific heat Cp, thermal conductivity k
and viscosity p., of the gas as well as the outside
diameter D of the tube. The velocity, viscosity,
density and tube diameter affect the thickness of the
fluid film at the tube wall through which heat must
first be conducted, as well as the extent of fluid
mixing. The average temperatuie of the fluid is a
function of its thermal conductivity and specific heat.
By means of dimensional analysis the relationship
given in the following equation
hD /DG\a /Cpp\b
33

34
that heat transfer is greater for smaller diameter rubes,
and higher mean film temperatures and mass gas
gas velocities. Average gas side convection heat transfer
co-efficients normally encountered in boiler design
vary from 5 BTU ft. 2 hr. °F in the low temperature
passes to 14 BTU/ft. 2 hr. C F in the high temperature
pases. The average heat transfer co-efficient from a hot
tube to a steamAvater mixture flowing through it is
about 1,000 BTU/ft. 2 hr. °F whilst the average heat
transfer co-efficient froma hot tube to superheatedsteam
flowing through the tube is about 120 BTU/ft. 2 br. °F.
From these figures it will be appreciated that the
overall heat transfer co-efficient from gas to a steam'
water mixture is virtually determined by the gas side
co-efficient whilst the overall heat transfer co-efficient
from gas to steam is determined by both gas side and
steam side co-efficients. This is illustrated in Fig. 5.1.
Note also the rise in Metal temperature in the second
case.
5.3 Radiation
The relative importance of the three modes of heat
transfer differs with the temperature level of the
system. In conduction through solids the mechanism
consists of an energy transfer through a body whose
molecules except for small oscillations around a space
within themselves remain continuously in a fixed
position. In convection, heat is first absorbed by
particles of fluid immediately adjacent to the source
and then transferred to the interior of the fluid by
mixing with it. Both mechanisms require the presence
of a medium to convey heat from the source to the
receiver. Radiant heat transfer does not require an
intervening medium.
If the phenomena of conduction and convection on
the one hand are contrasted with thermal radiation
on the other, it is found that the former are affected
by temperature differences and very little by temperature
level whereas the latter increases rapidly with
increase in temperature level. The rate of heat transfer
by radiation varies in fact as the difference between the
fourth powers of the absolute temperatures of source
and sink. The relationship
q--=CTA(T1 4 —T2 4 ) ... 5.5
expresses mathematically this concept and is known
as the Stefan-Boltzmann law. The constant of proportionality
in this equation is known as the Stefan-
Boltzmann constant. The relationship assumes that
the emissivitv of the source and sink are both unity,
i.e. they are both referred to as black bodies. A black
body is a body which will not reflect any radiant
energy. In the combustion chamber of a boiler the
emissivity of the flame and heating surfaces are not
unity, nor do tie heating surfaces present the equivalent
of an infinite parallel plane to the radiating gases.
The Stefan-Boltzmann relationship must therefore
be corrected accordingly. The corrected relationship
can be written as
where Fa is a dimensionless geometric factor and Fel
is a dimensionless emissivity factor.
Proceedings of The South African Sugar Technologists' Association—March 1966
Bagasse
Wood . . .
Bituminous Coai .
Oil . . . .
0.72
0.72
0.81
0.85
0.85
0.85
0.90
0.90
0.85
0.85
0.90
0.95

Proceedings of The South African Sugar Technologists' Association—March 1966 35
On leaving the furnace the hot gases are nonluminous
and at a temperature of 1,700-2,200 deg. F.
Heteropolar gases however such as water vapour and
carbon dioxide possess radiant emission bands in the
infra-red range of sufficient magnitude to merit
consideration in the convection passes of the boiler.
Heat transfer by non-luminous gas radiation varies
with the partial pressure of the gas and also with a
mean beam length which is defined as the radius of
an equivalent hemispherical gas mass. For gas shapes
of industrial importance it is found that any shape is
approximately representable by an "equivalent"
hemisphere of proper radius. The evaluation of the
equivalent hemisphere involves tedious graphical or
analytical methods and a number of text books deal
with this aspect of the problem. Fig. 5.4 illustrates
typical curves for 2" o.d. tubes pitched at two diameters.
While it is not strictly correct to talk in. terms of a
radiant heat transfer co-efficient in the same sense as
conductive or convective heat transfer co-efficients, it
is nevertheless useful to use such a concept in order
to obtain a quantitive comparison of the three types
of heat transfer. Typical figures are therefore given
in table 5.2. From this table it will be noted that the
luminous radiant heat transfer co-efficient is much
higher than the convective co-efficient, and therefore it
follows that it is more economical to transfer heat
radiantly than convectively. Other considerations,
however, which are discussed later limit the amount
of heat which can be transferred in this manner.
TABLE 5.2
Approximate Overall Heat Transfer Co-efficients encountered
in Boiler Design
Location
Combustion Chamber
Superheater
Convective boiler passes
Economiser—Bare tube
Cast Iron
Air Heater—Tubular
overall heat
Type of heat transfer coeff
transfer (BTU/ft 2 hr°F)
Luminous radiation
Convection
Non-luminous
radiation
Convection
Non-luminous
radiation
Convection
Non-luminous
radiation
Convection
Convection
20—30
10—14
2—3
4—14
0.5—2
6—10
0.4—0.5
3.5-4.5
2—6
6.00 Fluid Friction
As early as 1874 Osborne Reynolds pointed out
that there was a relationship between convective heat
transfer and fluid friction. The analogy between the
two arises from the fact that the transfer of heat and
the transfer of fluid momentum can be related. In
simple terms the relationship is:
Heat actually given, up
Total heat available to be given up ~
Momentum lost by friction
Total momentum available
The Reynolds analogy was extended by Prantdl to
include the laminar layer which exists near a pipe wall
as well as the turbulent layer. The Prandtl modification
is sometimes called the Prantdl analogy. Modern
theory now presumes that the distribution, of velocities
no longer ends abtuptly at the laminar layer but that
there is instead a buffer layer within the laminar layer
in which transition occurs. Other extensions, therefore,
of the analogies have been derived.
In practical boiler terms the Reynolds analogy
implies that in order to obtain a higher convective
heat transfer co-efficient, more work must be done in
overcoming the fluid friction between hot gases and
the heating surfaces. Since this involves using more
fan power the economic advantages of higher heat
transfer rates and hence smaller boilers must be
balanced against the cost of additional fan power
consumption.
7.00 Circulation
The water side heat transfer co-efficient as in the
case of the convective gas side heat transfer co-efficient
is a function of the velocity of the fluid flowing. If the
water side co-efficient becomes small in relation to the
gas side co-efficient, the tube metal temperature will
approach the gas temperature, thus overheating the
metal surfaces and causing blistering and possibly
even rupture. In boilers with good circulation metal
temperatures are only about 10°-20° F higher than
the saturated water temperatures.
Natural circulation in a boiler is caused by the
difference in density between water in the feeding leg
of a circuit and the water/steam mixture in the riser
leg of the circuit. In Fig. 7.1a simple circuit is shown
where two pressure vessels are connected by a bank of
heated downcomers and a bank of heated risers. The
number and diameter of the tubes in the riser bank
is fixed by the heating surface necessary to transfer
the required amount of heat to the water, while the
number of tubes in the downcomer circuit is fixed by
the cross sectional area of tubing required to maintain
an adequate water supply to the risers. The term
"adequate water supply" is defined empirically by a
circulating factor which is the ratio of the number of
pounds of water/steam mixture at any point in a
circuit to the number of pounds of steam at that point.
The circulating factor is introduced to fix the quantity
of water to be supplied to a circuit and is a function
of the heating intensity and an empirical overheating
parameter. The ratio of steam to water obviously
affects the density of the mixture and hence the circulating
factor must affect the static head which
produces circulation, the static head being the difference
in weight of fluid per unit area in the downcomer
and riser tubes. It should be remembered that density
is affected by the difference in velocity between steam
and water flowing in a tube. This difference has the

36 Proceedings of The South African Sugar Technologists' Association—March 1966
apparent effect of increasing the circulation factor
and hence decreasing the available static head.
The static head in the circuit is used to overcome
the head drop across the riser tubes and downcomer
tubes due to friction, entrance and exit losses, acceleration
losses and bend losses all of which are functions
of the quantity and density of the water/steam mixtures
flowing.
Since the static head is a function of the difference
in density between the fluid flowing in the riser and
downcomer tubes and the vertical height of these
tubes it follows that tubes where possible should
always be placed vertically in order to obtain maximum
circulation. Fig. 7.2 illustrates a typical boiler
circuit.
As pressure conditions rise the difference in density
between water and steam approaches zero (see Fig.
1.1) and the available static head decreases. High
pressure steam generators therefore have to be built
as high head units or alternatively some means of
forced or assisted circulation must be employed. A
typical example of a forced circulation steam generator
is the Lamont unit. Steam generators operating at
super critical pressures are generally of the "once
through" type.
8.00 Boiler Design
8.1 Boiler Heat Balance
The previous sections have dealt with the properties
and uses of steam, types of fuel available, their
methods of combustion and the relationships between
the rate of heat transfer, friction loss and circulation.
The performance of a boiler can be measured
without any reference to its physical characteristics.
Only the initial conditions of the feedwater and fuels
and the final conditions of the steam, exhaust gases,
ash and grits and casing temperatures need be considered.
Fig. 8.1 shows diagrammatically the factors
which go to make up a boiler heat balance; i.e.
Heat in = Heat transferred to steam
+ Heat lost in dry exit gases
+ Heat lost in evaporating moisture
produced when hydrogen in fuel is
oxidised to water.
+ Heat lost in evaporating free and
inherent moisture in fuel
+ Heat lost in evaporating moisture in
combustion air.
+ Heat lost due to unburnt carbon in
ashes.
+ Heat lost due to unburnt carbon being
carried over with grits to grit hoppers
and stack.
+ Heat lost by radiation and convection
from the boiler casing.
+ unaccounted losses.
Heat-In
The "Heat-In" is the gross calorific value of the
fuel determined by the bomb calorimeter, all products
of combustion having been cooled down to 60°F
and the moisture condensed.
Heat-In Steam
The "Heat-In Steam" is the amount of heat which
is transferred to the water to produce steam.
Dry Gas Loss
The "Dry Gas Loss 1 ' is the heat lost in the exhaust
gases by virtue of their temperature above ambient.
Wet Gas Loss
The "Wet Gas Loss" is the heat lost in the exhaust
gases in evaporating and superheating the moisture
contained in the fuel and that moisture which is
formed by the oxidation of hydrogen during
combustion.
Moisture in A ir Loss
Air used for combustion contains a small amount
of moisture which must be evaporated and superheated
only to be rejected in the exhaust gases.
Ash Loss
A small percentage of carbon is physically entrained
in the ash and its heating value is therefore lost.
Grit Carry-Over Loss
Lighter particles of fuel are sometimes carried
over in suspension with the gases and their heating
value is lost. If grit collectors are installed, a portion
of the grits can be collected and retired.
Radiation Loss
A boiler is generally insulated to bring the casing
temperature down to between 120° and 170° F.
Insulating to this temperature introduces a small
radiation loss from the surface of the boiler which
amounts to about 0.5 %-l .5 % of the total heat input.
Unaccounted Loss
To make up the final boiler heat balance, a figure
of between 0.5 % to 1.0 % is usually included to cover
errors in measurement and small losses such as the
sensible heat in ash, etc. which are not measured.
Laid out in Table 2.1 are the approximate heat
balances of four 100,000 p.p.h. boilers. Each boiler
is designed to burn a specific fuel, i.e. bagasse, wood,
coal and oil and has sufficient heat recovery equipment
fitted to bring the final gas temperature at the designed
COa down to 450° F.
As a rought guide a reduction of approximately
30° F in final gas temperature will be proportional to
an increase of 1 % in boiler efficiency. Further a
reduction of approximately 1 % in CO;, will be
proportional to a reduction of 1 % in boiler efficiency.
Boiler Blowdown
Modern boilerplant requires careful water treatment
control. In addition impurities which collect in a
boiler must be blown down. From Fig. 8.2 the
percentage blowdown can be calculated for specific
conditions. In general blowdown should not exceed
5% of the steam generated. This represents a heat
loss which is not normally included in the boiler heat
balance. However, since water and not steam is
blown down only sensible heat is lost and in most
installations in the sugar industry a 5% blowdown
rate is proportional to a drop in efficiency of only
approximately 0.5 %.

Characteristic Dump Grate
a) Relative cost for 100,000 pph unit
based on dump grate furnace .
b) Floor area occupied by boiler
(assumes ID fan and grit collector
outside building).
c) Height of building required to
d) Relative building cost (assuming
proportional to building volume)
e) Approx. installed HP
/) Approx. power consumption at
M.C.R. on major fuel
/) Superheater
00
(iii)
(iv)
j) Convection passes . . . .
k) Heat recovery equipment required
for normal industrial
/) Optimum final gas temperature
m) Other factors limiting lower final
gas temperatures
n) Grit Collector
o) Ash Handling
p) Automatic Controls . . . .
q) Ratio weight of water to M.C.R.
evaporation
1
TABLE 9.1
BOILER CHARACTERISTICS COMPARISON SCHEDULE
2,400 sq. ft.
48 ft.
1
480
270 KW
Oil
No thermal storage
Refractory sidewalls
require annual
maintenance
Turbulent thermal
conditions. Heavy
carry over
H.P. secondary air
required
Pendant. Not
drainable
Vertical tubes cross
baffled
Airheater
0.70
Bagasse Self
Feeding
0.91
2,500 sq. ft.
40 ft.
0.87
390
220 KW
Oil
Large thermalstorage
Large refractory
areas require annual
maintenance
Quiescent furnace
conditions limited
carry over
Medium pressure
secondary air
required
Cross flow.
Drainable
Vertical tubes cross
baffled
Airheater
420/480" F.
420/480" F.
420/480' F.
320/380" F.
Expensive economiser Expensive economisei • Expensive economiser Low temperature
required for small required for small required for small deposits. Cleaning
gain in efficiency gain in efficiency gain in efficiency severe problem
Required
Not required Required
Not required
Intermittant
mechanical ashing
Furnace must be
cleaned manually
once a week
Continuous
mechanical ashing
Continuous
mechanical ashing
Fuel feed, air flow
and furnace pressure
regulation required
Due to self feeding Fuel feed, air flow
properties of furnace and furnace pressure
auto controls simpli­ regulation required
fied only air flow and
furnace pressure
regulation required
Fuel feed, air flow
and furnace pressure
regulation required
0.78
Moving Grate
1.14
2,420 sq. ft.
60 ft.
1.27
495
280 KW
Coal
No thermal storage
on bagasse reasonable
on coal
Limited refractory
maintenance
Turbulent furnace
conditions on coal
and bagasse. Heavy
carry over
H.P. secondary air
required
Pendant. Not
drainable
Vertical tubes cross
baffled
Airheater
0.65
Coal
1.05
2,600 sq. ft.
54 ft.
1.23
300
170 KW
—
Reasonable thermal
storage
Limited refractory
maintenance
Quiescent fnurnace
conditions. Limited
carry over
H.P. secondary air
required
Cross flow.
Drainable
Vertical tubes cross
baffled
Economiser
0.58
Oil
0.73
1,680 sq. ft.
40 ft.
0.58
95
53 KW
—
No thermal storage
Limited refractory
maintenance
No carry over
Medium pressure
secondary air
required
Pendant or Cross
flow. Drainable
Vertical tubes cross
baffled
Economiser
350/400" F.
Low temperature
deposits. Cleaning
severe problem
Not required
Fuel feed, air flow
and furnace pressure
regulation required
0.63
Remarks
Allows 10 ft. and 5 ft.
clearance in front and at
rear of boiler respectively.
Gas can be burnt on unit
fitted with oil burners.

38 Proceedings of The South African Sugar Technologists , Association—March 1966
9.00 Boiler Layout
Boiler design is largely dictated by:
1) Fuel:
Physical and chemical properties;
Cost.
2) Final Steam Conditions:
Pressure;
Temperature.
3) Feedwater Conditions:
Temperature;
Impurities.
4) Application:
Industrial;
Power Generation.
Fig. 9.2 illustrates a few of the basic boiler designs
used in the Cane Sugar Industry. Their characteristics
are compared in Table 9.1 while an indication of the
effectiveness of the different types of heating surface
used in the design illustrated in Fig. 9.2B is shown in
Fig. 9.1A. Fig. 9. IB shows the energy flow diagram
for this unit.
Industrial design considerations are as follows:
9.1 Combustion Chamber
Furnace geometry is governed by the fuel and type
of combustion equipment used. High moisture content
fibrous fuels require fairly large furnaces with a
refractory belt in the combustion zone to ensure
stable conditions. With fuels having higher combustion
temperatures, however, such as coal and oil, furnace
volumes can be reduced and fully water cooled walls
included.
To minimise screen and superheater tube fouling,
sufficient heating surface should always be incorporated
to ensure that the gas leaving temperature is at
least 150' to 200° F below the ash deformation
temperature. Reducing the gas leaving temperature
much below this point reduces the temperature
potential in the superheater zone thus making the
superheater itself much larger.
9.2 Superheater
Superheaters can be broadly classified into two
groups:
fa) Drainable
(b) Non-Drainable.
Drainable superheaters are less prone to overheating
on start-up. The pendant non-drainable type
however are readily accommodated in the tall combustion
chambers essential with suspension firing.
Tube configuration can be critical. A minimum
pressure drop varying from 5 to 25 p.s.i. at M.C.R.,
is usually necessary to ensure an even steam flow
distribution through each tube. Further, by varying
tube pitching superheater characteristics in relation
to load can be appreciably altered. Increasing the
tube pitching increases radiation heat transfer at the
expense of convection heat transfer. Conversely,
decreasing tube pitching increases convection heat
transfer at the expense of radiation heat transfer. By
carefully adjusting tube geometry therefore an almost
flat characteristic can be obtained over a wide range
of boiler loading. Fig. 9.3 illustrates the convection
and radiation transfer characteristics in relation to
load.
9.3 Convection Surfaces
The boiler exit gas temperature can be economically
reduced to within 250° and 300'' F of the saturation
temperature by means of convection heating surface.
As boiler pressures rise this surface is less effective
and heat recovery surface, in turn becomes more
important. Up to about 650 p.s.i.g. convection
heating surface plays an important role in the design
while at pressures exceeding 1,000 p.s.i.g. it is virtually
redundant.
9 • 4 Reco very Equipment
Heat recovery equipment is designed to increase
boiler efficiency by still further reducing the boiler
exit gas temperature. This is accomplished by either
an economiser which transfers heat in the flue gases
to the feed water, or an air heater which transfers
heat in the flue gases to the combustion air.
In industrial plant the quantity of heat which can
be transferred in each case is limited by:
(a) The gas temperature can only be reduced
economically to within 150' to 200 u F of the
feedwater or ambient air temperatures.
(b) The quantity of heat transferred to an economiser
operating under industrial water treatment
conditions should be limited to avoid depositing
feedwater solids in the tubes which occurs if
boiling takes place,
(c) The combustion air temperature should not
exceed a figure which would induce high stoker
and furnace maintenance costs due to excessive
combustion chamber temperatures. This figure
is usually limited to between 350° and 400° F.
Economisers and air heaters are made of either cast
iron or mild steel, the choice being dependent upon
factors such as the sulphur, phosphorous and moisture
content of the fuel and operating metal temperatures.
Wherever possible designs should ensure that operating
metal temperatures are higher than the dew
point of the gases so as to minimise corrosion. There
are a number of techniques used to accomplish this,
such as:
(a) Recirculating hot air through the cold passes of
an air heater.
(b) Recirculating hot water from the boiler drum
through the economiser to elevate the feed water
temperature.
(c) Arranging heating surfaces as parallel rather
than counter flow exchangers.
(d) Adjusting the heat transfer co-efficient by varying
mass flow rates so that metal temperatures
are closer to the gas temperature rather than
the air temperature fsee Fig. 5.1).
All these techniques unfortunately reduce the
effective temperature difference between the hot gases
and the water or air. Larger heating surfaces are
therefore required but maintenance and replacements
problems are minimised. Where metal temperatures

Proceedings of The South AJrican Sugar Technologists'' Association—, •March 1966 39
cannot be effectively controlled, e.g. where feed water
temperatures to economisers are below 180° F or
where high efficiencies are required heat recovery
equipment should be made of either cast iron which
enables a heavier metal section to be used economically
or a specially treated steel.
The choice and extent of heat recovery plant
required largely depends on fuel costs and characteristics.
As final gas temperatures are reduced to
increase efficiency, more and more heat recovery
equipment is required until the increase in cost of
plant and additional fan power outweighs the saving
in fuel. Further, lower final gas temperatures introduce
corrosion and fouling problems. The economic limit
has therefore to be carefully assessed.
The combustion stability of high moisture content
fuels is improved if the combustion reaction temperature
is increased. Air heaters are therefore far more
advantageous than economisers when, burning these
fuels.
9.5 Draught Plant
In general modern bagasse fired boiler plant operates
under balanced draught conditions, i.e. a forced
draught fan provides the combustion air and an
induced draught fan draws the products of combustion
over the boiler heating surfaces and exhausts them to
atmosphere.
A high pressure (10" - 30" water gauge) secondary
air fan is often used to inject high velocity air into the
furnace to increase turbulence and hence combustion
efficiency.
9.51 Forced Draught Fans
These are normally high volume, low pressure fans
fitted with horsepower limiting backward bladed
impellers. To ensure flexible boiler operating
characteristics they should be designed to supply at
least 15 % more air than that required at the design
C02 at M.C.R. against a pressure 32% in excess of
the M.C.R. draught loss. The fans should be capable
of providing 100 % of the combustion air requirements.
9.52 Secondary Air Fans
These are normally low volume, high pressure fans.
Their blade shape depends on the duty which they
have to perform and hence can vary over a wide
range. Since they normally generate fairly high
pressure their operating noise levels are high and care
should be taken in the design stages to limit this to
95 decibels.
9.53 Induced Draught Fans
These are high volume, medium pressure, high
temperature fans fitted with forward curved radially
tipped blades having anti-fouling and anti-erosion
properties. The exhaust gases from bagasse fired
boilers contain a particularly abrasive dust and
special precautions must be taken in impeller design
to minimise erosion, if a life of more than one crop
is to be obtained. Renewable anti-erosion stellite
ridges placed normal to the direction of the gas flow
have proved satisfactory. Peripheral velocities however,
should not exceed 15,000 ft. per minute.
They should be designed to handle at least 20 % by
volume more than that required at the design C02 at
M.C.R. against a pressure of at least 44% in excess of
the M.C.R. draught loss.
9.54 General
Boiler fans are driven by means of either steam
turbines or electric motors, the latter becoming
increasingly more popular.
Electric motors where installed should be of T.E.F.C.
weatherproof design which although more expensive
initially are the most suitable for operation in the
dusty conditions prevalent in a boilerhouse. Fan
drives should be carefully chosen in relation to the
high inertia masses which they have to accelerate
when starting.
A large number of different types of fan controls
are available, varying from simple discharge damper
controls to complex variable speed drives. In a well
balanced factory where a surplus of bagasse can be
produced if necessary, simple damper controls are
the most effective.
9.6 Grit Collectors
In general grit collectors need only be fitted to units
employing suspension firing.
Smut particles are highly inflammable and care
should be exercised while operating the plant to
ensure that hoppers are always free flowing. Unless
high efficiencies are required grit refiring is not
recommended due to the abrasive nature of the
retired particles which increases maintenance costs
and also aggravates the fire hazard problem in
hoppers from which refiring is taking place.
9.7 Boiler Valves and Mountings
A well balanced complement of boiler valves and
mountings fitted to a bagasse fired boiler of 100,000
p.p.h. capacity is illustrated in Fig. 9.4. Provision is
made for the future installation of an economiser as
well as for water sampling, continuous blowdown and
high pressure chemical injection. Two absolute water
gauges are fitted as well as a high and low level water
alarm and remote water level indicator. The feed
valve arrangement ensures maximum flexibility.
9.8 Feed Water System
To protect boilerplant operating at over 350 p.s.i.g.
against oxygen and C02 corrosion deaerating plant
should be installed in the feed circuit. A simple but
effective arrangement which provides sufficient storage
to cover most emergency conditions is illustrated in
Fig. 9.5. A pressure operated auto cut-in gear should
be fitted to the turbo feed pump to ensure that an
adequate supply of feed water is always available.
This gear should of course be checked once per shift
for satisfactory operation. Feed pumps should be
sized to deliver at least 15% more than the M.C.R.
rating of the plant to cater for load surges.
9.9 Instruments and Controls
Boiler instruments and controls, which should
preferably be centralised, should be kept as simple
as possible. Only those instruments which materially

40
affect boiler operation should be fitted. Log sheets
should be designed to ensure that vital operating
factors, such as bearings, etc are checked and their
condition logged at least once every hour.
A well engineered complement of instruments and
controls should not exceed the following:
Instruments:
(a) Pressure Indicating:
1) Superheater outlet steam pressure;
2) Boiler drum pressure;
3) Feed water mains pressure.
(b) Temperature Indicating:
1) Final steam temperature;
2) Feed water temperature;
3) Economiser water outlet temperature (if
economiser fitted);
4) Boiler outlet gas temperature;
5) Aii heater outlet gas temperature (if air
heater fitted);
6) Final exhaust gas temperature;
7) Ambient air temperature;
8) Air heater air outlet temperature (if air
heater fitted).
(c) Level Indicating:
1) Local boiler drum level indicator;
2) Remote boiler drum level indicator;
3) Feed tank and deaerator level indicators.
(d) Power Indicating:
1) I.D. fan amps;
2) F.D. fan amps;
3) Other fan amps;
4) Feed booster pump amps;
5) Electro feed pump amps;
6) Fuel feeder amps.
Recorders:
1) Combined steam flow steam pressure recorder.
Audible Alarms:
1) Boiler low,high level water alarm.
Visual Alarms:
1) Feedwater tank low/high level alarm;
2) Deaerator low/high level alarm;
3) Bagasse supply failure alarm.
Analysers:
1) Portable hand-operated C02 analyser for spot
checks.
Auto controls:
1) Boiler water level control;
2) Feed tank level control;
3) Deaerator level control;
4) Steam pressure control based on simple air/fuel
ratio controllers;
5) Furnace pressure controller.
Proceedings of The South African Sugar Technologists' Association—March 1966
Conclusions
A wide range of boiler plant is available to suit
most factory requirements; the choice being dependent
upon local conditions. Where positive economic
thinking dictates that only bagasse shall be burnt as a
fuel, the boiler fitted with a self-feeding furnace very
adequately meets these requirements. Further in
areas where skilled operators are at a premium, this
unit provides broad operating characteristics without
the need for sophisticated control systems. In areas
where skilled labour is readily available and a maximum
degree of mechanisation is required, the air
glide furnace with dump grate presents a reasonable
solution to the problem. If, however, the burning of
auxiliary fuels is to be avoided with this unit an
assured bagasse supply must be available. Fig. 10.1
illustrates a possible solution to this problem.
Oil burners for use during the Off-Crop, can be
fitted to either of these two units at a very small
increase in capital cost.
Auxiliary coal firing can only be accommodated
readily in a unit of the type illustrated in Fig. 9.2C.
The unit is considerably more expensive and is
essentially a compromise design.
Since auxiliary fuels in a well balanced factory
should only be burnt during the off crop, the off crop
load should be carefully assessed, especially where
coal is the auxiliary fuel, in order to establish whether
it would not be more economical to install a separate
coal fired boiler to cater for this load at maximum
efficiency and minimum capital outlay.
Modern bagasse fired water tube boilerplant can be
constructed with unit capacities in the range 20,000
p.p.h. to 300,000 p.p.h. This means that factories milling
up to 300 tons of cane per hour can be served, if
necessary by one boiler. Reference to Fig. 10.2 will
indicate that this is in fact the cheapest installation,
but it also necessitates putting "all one's eggs in one
basket". The final choice must rest with the factory
management. However, since reliability of plant is
increasing continuously, large unit sizes should be
carefully considered.
Summary
The paper is intended to provide a useful reference
work on modern steam generating plant used in the
Cane Sugar Industry.
The paper is divided into three sections:
The first section deals with the
1.00 Properties and Uses of Steam.
2.00 Fuels.
3.00 Combustion.
4.00 Combustion Equipment.
5.00 Fundamental Heat Transfer Relationships.
6.00 Friction Loss; and
7.00 Circulation Relationships
used in boiler design.
The second section deals firstly with
8.00 Overall boiler heat balance, and secondly,
with

Proceedings of The South African Sugar Technologists' 1 Association—March 1966 41
9.00 Boiler heating surface disposition and Ta Ambient air temperature
structural design.
la. the last section a number of designs are analysed
and possible future trends discussed.
10.00 Conclusions.
Acknowledgments
I would like to thank John Thompson Africa (Pty)
Ltd. for allowing me to publish this paper and my
colleagues D. J. van Blerk and D. A. Jones for their
assistance in its preparation.
List of Symbols
A Area through which heat is
transferred normal to the direction
of heat transfer, absorbed or
reflected normal to the radiating
body.
Cp Specific heat of fluid at constant
pressure.
D Outside diameter of tubes over
which fluid flows.
F„ Radiation geometric factor introduced
in order to extend the
Stefan-Boltzmann relationship
to cases other than infinite
parallel planes.
Fe Emissivity factor introduced in
1
order to extend the Stefan-
Boltzmann relationship to bodies
other than black bodies.
Ft, Corrected emissivity factor for
use in equation 5.8
G Mass gas flow
Hj Total heat of dry saturated steam
h Total conductive heat transfer
co-efficient
he Convective heat transfer coefficient
hg
hr
Gas side heat transfer co-efficient
Radiant heat transfer co-efficient
hj Sensible heat in steam
hL Conductive heat transfer coefficient
through body 1
h2 Conductive heat transfer coefficient
through body 2
k Thermal conductivity
L Effective vertical length of downcomers
P Pressure
Q Heat transferred
q Heat transferred per unit time
T Temperature
ft 2
BTU/lb.
ft.
dimensionless
dimensionless
dimensionless
lb/ft. 2 hr.
BTU/lb.
BTU/ft. 20 F
hr.
BTU/ft. 20 F
hr.
BTU'ft. 2 °F
hr.
BTU/ft. 20 F
hr.
BTU/lb.
BTU/ft. 20 F
hr.
BTU/ft. 20 F
hr.
BTU/ft.hr.
L F
ft.
PS.I.A.
BTU
BTU/hr.
°F
absolute
Tfe Temperature of gases leaving
furnace
T/j, Temperature of furnace gases
Tw
Temperature of tube walls absorbing
heat
Tl5 T2 Temperatures of source and sink
respectively
At Temperature difference between
hot and cold faces of a body
through which heat is flowing
t. Saturation temperature of steam
at pressure P
V Volume of gas
Vj Specific volume of dry saturated
steam
v Velocity of fluid flowing
u Specific volume of water
u„, Specific volume of water/steam
mixture
Wy Weight of fuel burnt
Wj Weight of gas per lb. of fuel in
combustion chamber
x Distance between hot and cold
faces
t Unit of time
fj, Viscosity of fluid
P Density of fluid
p,„ Density of steam/water mixtuie
a . 1723 Stefan-Boltzmann constant
8 Unit of time
Abbreviations:
p.p.h.
M.C.R.
N.C.V.
G.C.V.
°F.
absolute
°F.
absolute
°F.
absolute
°F.
absolute
°F.
absolute
°F.
°F.
ft 3
ft 3 /lb.
ft/sec.
ft s /lb
ft 3 /lb
lb/hr.
lb.
ft.
sees.
lb/ft. sec.
lb/ft 3
lb/ft 3
hr.
pounds per hour
Maximum continuous rating
Nett calorific value
Gross calorific value
References
Callendar, G. S. and Egerton, A. C. The 1939 Callendar
steam tables. Second Edition 1952. Published by Edward
Arnold.
Fishenden, M. and Saunders, O. A. An introduction to
heat transfer. First edition 1961. Published by Oxford
University Press.
Mullikan, H. F. Evaluation of effective radiant heating
surface and application of the Stefan-Boltzmann Law to
heat absorption in boiler furnaces. Trans, of the American
Society of Mechanical Engineering.
Gaffert, G. A. Steam Power Stations. Fourth Edition 1952.
Published by McGraw-Hill.
Kern, D. Q. Process Heat Transfer. 1950. Published by
McGraw-Hill.
6. King, J. G. and Brame, J. S. S. Fuels—Solid, Liquid and
Gaseous. 1961. Published by Edward Arnold.
Lyle, O. The Efficient use of Steam. 1947. Published by Her
Majesty's Stationery Office.
McAdams, W. H. Heat Transmission. Third Edition 1954.
Published by McGraw-Hill.

42
Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' 1 Association—March 1966 43

44 Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 45
O IOOO 2000 3000 4000 SOOO 6000
MASS GAS FL0W(LBS/FT. 2 HR)
FIGURE 5.2: Heat Transfer Coefficient for Flue Gas Flowing at Different Mass Velocities across a Bank
of Tubes (Mean Film Temp. 200° F.) at least Six Rows Deep Square pitched at 2D

46 Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 47

48 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 7.1: Simple Boiler Circuit.

Proceedings of The South African Sugar Technologists' Association—March 1966 49
COLLECTOR
TUBES
The system comprises a number of parallel heated circuits which generate a potential in the form of a static head.
The flow through each circuit is controlled by careful sizing of the downcomer tubes to provide the necessary
system resistance thus ensuring adequate water supply to each circuit and preventing overheating of the tubes.
FIGURE 7.2—Typical Boiler Circuit

Proceedings of The South African Sugar Technologists' Association—March 1966 51
FEEDWATER CONCENTRATION P.P.M.
FIGURE 8.2: Blowdown Rate as a function of Feed Water Concentration.

52 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 9.1 A: Chart showing the effectiveness of Heat Transfer Surfaces.

Proceedings of The South African Sugar Technologists' Association—March 1966 53
FIGURE 9.1 B: Energy Flow Diagram.

54. Proceedings of The South African Sugar Technologists' Association March 1966
FIGURE 9.2 A: Bagasse Fired Boiler with Dump Grate and Auxiliary Oil Burners

56 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 9.2 C: Bagasse Fired Boiler with continuous Ash Discharge Stoker and Auxiliary Coal Firing Equipment.
f

Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 9.2 D: Coal Fired Boiler.
57

c
m
•o
to
a.
o
ECONOMISER
GAS FLOW
— AIR FLOW
•~\*
PENDANT SUPERHEATER -CROSSFLOW DRAINABLE
SUPERHEATER CAN ALSO BE FITTED
WATERCOOLED COMBUSTION CHAMBER
OIL BURNERS
fci
^
c
I

Proceedings of The South African Sugar Teclmologisis" Association—March 1966 59

V V

62 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 10.2: Curve showing Relative Cost per pound of Steam for Various Sizes of Boilers.

Proceedings of The South African Sugar Technologists' Association—March 1966 63
Mr. Ashe: It is indicated in figure 10.2 that the
cost per pound of steam is much higher for smaller
boilers than for larger boilers but this has not been
our experience with a 50,000 lb. per hour boiler.
Mr. Magasiner: With small boilers it is possible to
simplify the design and thus effect economies.
Mr. Hurter: How does the availability of boiler
plant compare to the availability of turbo-alternators
and turbines?
Mr. Magasiner: The availability of boiler plant is
improving rapidly. Provided the boiler water treatment
is adequate and attention is paid to the plant
while it is operating there is no reason why a boiler
should not steam throughout the crop without a loss
of production. It should be possible to operate on one
boiler only.
Mr. Hurter: It is not feasible to operate on one boiler
if both bagasse and coal are being burned.
Mr. Magasiner: If it is running either on coal or on
bagasse there is no reason why a factory should not
operate on one boiler.

64 Proceedings of The Soutli African Sugar Technologists Association—March 1966
Scope
The subject matter of this paper, if treated in the
general sense, covers an extremely wide field to which
it is impossible to do justice in the time allocated. It
is even difficult to cover the subject adequately if we
confine ourselves only to the furnace design of bagasse
fired sugar mill boilers.
I therefore intend to give a short summary of past
and present bagasse furnace designs and cover the
subject of auxiliary fuels and how the latter have
affected basic boiler design considerations.
For the purposes of this paper I have assumed that
the majority of you have been concerned, one way or
the other, with the practical aspects of steam raising
equipment in sugar mills.
Historical Background
Bagasse has been used as a fuel for the boiling
processes of sugar production since the 18th century.
With the introduction of steam boiler plant in the
early 19th century and the growing mechanization of
sugar mills, it became necessary to overcome the
problems of burning this wet and rather unmanageable
fuei in large quantities.
About 100 years ago the first Dutch ovens appeared,
typically the Cook's Hearth furnace, which was the
forerunner of many variations which are still in use
up to the present day. However these variations took
different forms in different parts of the world.
The original Cook's Hearth, which consisted of only
one cell, with an approximately square grate on to
which the fuel was fed in a conical pile, is still the
basic form used in many parts of the world.
In this country over the years we have seen the
development of this basic design with the introduction
of forced draught, preheated air, flat suspended roofs
and arches and the final refinements of secondary overfire
air and tertiary air into the secondary combustion
chamber. With the rise in capacity of boiler plant, the
number of cells was often increased to three, each cell
maintaining the approximately square shape of the
grate.
A parallel development, mainly in the Western
hemisphere, was the Horseshoe Furnace, in which each
cell was built up with refractories with a horseshoe
shape in the plan view. The main difference from the
Cook's Hearth was the elimination of the grate, the
fuel burning in a conical pile on the Moor. Primary air
was introduced through cast iron tuyeres built into
the first course of brickwork, with secondary air
admitted higher up and all the way round the periphery
of the horseshoe. These furnaces attained very
high ratings indeed. With hot air, ratings of up to 1.5
X 10 6 BTU/hr. per sq. ft. of floor area were fairly
common. This compares with the Cook's Hearth,
where grate ratings seldom exceed 1.0 X 10 6 BTU/sq.
ft./hr. with hot air.
FUELS AND FURNACES
By P. R. A. GLENNIE
The modern counterpart of the Horseshoe Furnace
was developed in the 1930's in the West Indies and
was known as the Ward Furnace. These furnaces
eliminated the Dutch oven construction in front of the
boiler and were placed directly below the main boiler
furnace with an intervening refractory throat. The
basic horseshoe shape was originally retained with
later variations to oval and then rectangular cells.
Although these furnaces have been introduced to
many parts of the world, South Africa has been a
notable exception.
Another parallel development was the inclined step
grate originally introduced at the end of the 19th
century. In its original Dutch oven form it still exists
commonly in the East and West Indies. Particularly
in more primitive areas these furnaces operate on
natural draught without forced draught or airheaters.
In some cases the step grate has been brought in
under the main furnace in a similar way to the Ward
Furnace. In other cases preheated air and forced
draught have been introduced.
A post war development has been the Eisner
Furnace which started as a variation of the step grate
Dutch oven. In its final form the furnace is also placed
beneath the main combustion chamber, eliminating
the typical Dutch oven contours, but the method of
combustion and shape of fuel bed are similar to that
of the step grate.
With the exception of the Eisner Furnace all the
foregoing designs with their many variations were in
use throughout the cane sugar milling industry until
the late 1940's.
In the years immediately prior to the last war coal
spreader fired boilers had become extremely popular
in the United States both for power station and
industrial use. It was found that this method of firing
was extremely flexible and could be used for all
varieties and sizings of bituminous coal. It could also
be used for the auxiliary, or even main, firing of refuse
fuels.
In the vast timber growing area of North America
considerable quantities of electricity are generated
from sawmill wood waste and it became common
practice to fire this waste by spreader methods with
coal or oil as the standby fuel.
It was therefore only logical that the American
boiler and stoker companies should turn their minds
to the possibility of burning bagasse on spreader
stokers and developments were started in this direction
very shortly after the second world war. The advantages
of spreader firing were almost immediately
apparent and the new system caught on very rapidly
indeed.
In Southern Africa the first spreader fired units
were installed in 1953/4 and by 1960 probably 90%
of all new installations were of the spreader type.
Many Dutch oven fired boilers were converted to the
new system.

Proceedings of The South African Sugar Technologists' Association—March 1966 65
In this country it is indeed remarkable that, in the
short space of less than 15 years, spreader firing has
taken over almost entirely from the older methods.
Fuels for Sugar Mills
Bagasse always has been, and for the forseeable
future must be, the primary fuel of a sugar mill.
In the average modern mill, having a cane input of
over 200 tons per hour, the weight of bagasse produced
per day is approximately 1,800 tons and, with its low
bulk density, the volume is prodigious.
Although there have been many schemes to make
better use of this fibrous substance, particularly in
paper and cardboard manufacture, these have not been
of any particular economic significance. It must
therefore be stressed that bagasse must remain the
primary fuel of all cane sugar mills.
With the increase of refining facilities in larger
mills and the necessity to generate electricity for
irrigation, the problem often arises of a certain shortage
of bagasse and the engineering designer must
either economise in his use of steam or increase his
boiler efficiency to overcome the shortfall of this fuel.
Only too frequently, and usually to minimise capital
expenditure, recourse is made to the use of auxiliary
fuels such as wood, coal and, in countries where coal
is not as cheap as it is in South Africa, oil, rather than
increase the efficiency of the bagasse firing equipment.
Auxiliary fuels, of course, have their advantages
from the point or view of shutting down and starting
up procedures and in the event of temporary mill
outages. They obviate, to a large extent, the necessity
to store very large quantities of bagasse, which is so
awkward to handle mechanically.
Of course, with the generation of electricity for
irrigation purposes, it is necessary to fire boiler plant
with such auxiliary fuels during mill shutdown and
off-crop periods.
A nice solution from the boiler designers point of
view would be to install the highest possible efficiency
boiler plant so that an excess of bagasse is produced
and the fuel stored for periods of mill shutdown.
Unfortunately nobody has yet come forward with a
really practical solution to the problems of bagasse
storage.
In this country we have the factor of probably the
cheapest coal supplies in the world and this is the
main reason for the very rapid development of spreader
fired boilers in Southern Africa with their adaptability
to coal/bagasse dual firing.
In most other sugar producing countries, particularly
the West Indies, oil is cheap and hearth type
furnaces, such as the Ward, with oil burners in the
secondary furnace are able, by dint of lower capital
costs, to hold their own against the spreader stoker.
Due to clinkering difficulties with high furnace
temperatures it is, of course, not practical to fire high
ash coal in Dutch oven furnaces.
On the other hand wood, with very similar burning
qualities to bagasse, can be burnt either in hearth
furnaces, or for that matter, on spreader stokers.
I have often wondered why so much labour is
expended on bringing large quantities of wood logs
to the boilerhouse and manhandling them into the
furnaces for lighting up purposes. If this fuel were
put through a simple hogging machine it could so
easily be fed to the boilers through the normal bagasse
carrier system.
In the following table I give the approximate
furnace gas quantities for a boiler steaming at 100,000
lb/hr. with the four different fuels that we have
discussed above:
Bagasse Coal Wood Oil
% moisture in
fuel . . .
G.C.V. BTU/lb.
Furnace gas
50
4,150
5
12,000
50
4,300
—
18,500
weight lb/hr. . 205,000 150,000 200,000 120,000
Assuming that we have a hypothetical boiler that
can fire all four of these fuels, obviously the fuel with
the lowest gas weight will have the most heat removed
from these gases in the course of their passage through
the boiler. If we have designed the boiler for operating
with an economiser, or airheater, outlet temperature
of 400" F. on bagasse, the equivalent temperature on
wood will probably be 390° F, on coal 350° F. and
on oil 320° F.
Consequently, with dual firing and where we desire
to install boilers of high efficiency, we came up against
the problem of corrosion with the higher grade fuel.
Particularly with auxiliary oil firing in boilers
specifically designed for bagasse, it is essential to bypass
heating surfaces and/or provide parallel flow
airheaters in order to prevent dew point corrosion in
the airheaters and draught plant. This is only one of
the many criteria for the design of modern high
efficiency boilers for sugar mills.
Design Considerations—Hearth Type Furnaces
With the majority of hearth furnaces, such as the
Horseshoe, Ward or designs best known in this
country, the Vincent & Pullar and Murray type
furnaces, the fuel bed is ostensibly in the form of a
cone with the fuel fed from over-head on to either a
grate or the floor of the cell.
With very high temperatures maintained in the
hearth cell the moisture in the fuel is driven off and
most of the fuel ignites from the bottom periphery
of the cone upwards. With the fuel falling through the
flames a limited proportion representing the smaller
particles of fuel, are flash dried and readily ignite;
sometimes before they reach the cone.
The whole operation necessitates furnace temperatures
of a very high order and this brings with it the
following disadvantages:
Firstly, we have the difficulty of slag formation
in the primary and secondary combustion
chambers, although this will vary considerably with
the type of cane and the area in which it is grown.
Secondly, due to the high temperatures in the fuel

66
bed there is a tendency towards the production of
gases which may pass through the boiler to the
chimney in the unburnt state thus constituting a
loss in boiler efficiency. For example, carbon dioxide
will re-associate with carbon at high temperature
levels to form carbon monoxide which may not all
be burnt in the secondary combustion chamber.
Admittedly, with modern furnace designs, secondary
air blows directly on to the cone of fuel and tertiary
air is introduced at the entrance to the secondary
combustion chamber and the prevalence of unburnt
gases in the flues is not very apparent.
Thirdly, the temperatures associated with hearth
designs, with their large areas of brickwork, produce
high radiation losses.
Fourthly, brickwork maintenance at these high
temperatures is obviously a serious consideration.
There are, however, certain advantages in favour
of the hearth type furnace and these can be summarised
as follows:
The first and principal consideration is simplicity
of operation. With comparatively primitive labour
the fuel feed can be controlled easily as the level of
the top of the cone is not critical. As long as this
level is kept approximately constant, boiler output
can be controlled by the simple manipulation of
dampers.
Secondly there is the advantage of a fairly
considerable reserve of mill in the furnaces in the
event of fuel stoppage and cessation of supplies of
bagasse.
It is when one comes to the question of the advance
of boiler design over the last 20 years that one comes
across the most serious objections to hearth type
furnaces.
Up to comparatively recent times sugar mill boilers
were very lowly rated, the evaporation seldom exceeding
5 ib.hr. per sq. ft. of boiler heating surface.
In most parts of the world 3- and 5-drum bent
tube boilers were used although, in South Africa,
preference was shown for the sectional header type
longitudinal drum boiler.
The design of these two types of boilers was such
that a maximum of approximately 3,000 lb. hr. of
steam could be generated for each foot in width
between boiler side walls. Thus a 50,000 lb. hr. boiler
would be about 18 ft. wide.
The Horseshoe and Ward Furnaces with their
comparatively high furnace rating, fitted in almost
exactly with this criterion of boiler design. The more
lowly rated furnaces such as the Cook's Hearth and
Step Grate usually required a greater width than the
boiler and the side walls of the latter were stepped out
to accommodate the Dutch oven width.
With the advent of more highly rated boilers in
recent years the permissible evaporation per unit
width has increased rapidly until today the figure
given above has very nearly doubled.
Obviously the more highly rated boilers of modern
design have the advantage of reduced initial costs and
furnace designs must fall into line. In the case of the
Proceedings of The South African Sugar Technologists' Associalion—March 1966
Cook's Hearth, the Horseshoe Furnace and the Step
Grate Dutch oven we theoretically arrive at a furnace
width nearly twice that of the boiler and this is
obviously ridiculous.
At Mercedita Sugar Mill in the West Indies an
attempt was made to overcome this problem with a
Ward Furnace design having an individual cell width
of 6 ft. and a length of II ft. Instead of having a
cone of bagasse on the floor of the hearth a fuel bed
shaped like the top of a grave was produced by means
of an air spreader using a varying air pressure.
At Medine in Mauritius a similar idea was introduced
only this was not a true Ward. Furnace in that
dumping grates were used and there were only two of
them for a 35 tonnes, hour boiler. This design, was, in
effect, a compromise between a hearth furnace and a
full spreader design and can be classed under either
category. Grate ratings for this design are midway
between those employed for conventional spreader
firing and those for typical hearth designs.
Design Considerations—Spreader Firing
With the spreader firing of bagasse, or any other
wet vegetable fuel, intimate contact is made in
suspension between the hot incoming air, flames and
furnace gases on the one hand and the fuel on the
other.
This results in flash drying of the fuel which loses
a large proportion of its moisture before it lands on
the grate. Bagacillo and smaller fibres actually burn in
suspension. The fuel bed on the grate is relatively
thin and most of the rest of the moisture is driven off
by radiation from the furnace flames themselves rather
than radiation from the refractories.
The grate itself is either of the stationary, dumping
or travelling grate type. Due to difficulties of cleaning
fires and ashing, the stationary grate can only be used
for very small boilers and we will not consider it in
this context.
The dumping grate can be used for boilers up to
about 60,000 Ib.hr., a steam dumping mechanism
usually being employed. This type of grate is perfectly
suitable where bagasse is the main fuel and coal is
only used for supplementary purposes or for limited
periods on its own. Obviously dumping grates are
also suitable where the auxiliary fuel is wood or oil.
The travelling grate design with the grate moving
from rear to front, is essential where it is required to
burn coal over extended periods. For coal firing alone
boiler sizing may go as high as 300,000 Ib./hr. although
I believe the largest bagasse fired unit is only 150,000
lb.hr. This type of grate obviously has the advantage
of bringing all ashes to the front of the boiler automatically
whence they can be removed by a variety
of mechanical means.
Regarding the methods of spreading the fuel over
the grate, there are two principal means of accomplishing
this.
The first method consists of a rotor, usually with
four blades, on to which the fuel is fed and thrown into
the furnace. Very much higher rotor speeds are

Proceedings of The South African Sugar Technologists' Association—March 1966 67
required to obtain a satisfactory spread with bagasse
when, compared with coal. It is therefore desirable to
have means of varying over a wide range the speed of
the rotors according to the fuel being burnt.
The second and rather simpler method of distributing
bagasse over the length of the grate is by means
of pneumatic air spreaders. These obviously have
no moving parts and operate on. an air supply from
the boiler secondary air system. However they are
unfortunately unsuitable for coal except in a special
applications.
Consequently, and particularly in this country where
the need is to burn coal and bagasse, the tendency
has been towards the rotor type spreader which can
handle either fuel.
Personally I am very much in favour of air spreading
as it must increase the Hash drying of bagasse especially
when, hot air is used in. the spreaders themselves.
The main disadvantages of spreader firing can be
summarised as follows:
Firstly, with spreader firing it is essential to meter
the fuel in. order to obtain a completely steady
rate of feed. This means that we must provide
some form of feeder with a variable speed drive
and it is by no means easy to design a feeder which
will handle such an awkward fuel as bagasse.
However this problem, has been reasonably well
overcome with the drum type and the slat type
conveyor feeders. The latter has been adapted to
meter both the bagasse and coal in the same unit.
Secondly, and partly as a result of the first disadvantage,
it is necessary to have more complicated
controls and more skilled operators than is obtained
with hearth furnace operation.
Thirdly, due to the relatively thin fuel bed, there
is little reserve of fuel on the grate in the event of
mill stoppage and it is necessary to change over to
coal within a fairly short space of time. Operators
have to be trained to carry out this operation
without loss of steam pressure.
Fourthly, especially where travelling grate stokers
are used, the cost of firing equipment is much
higher with spreader stokers than with the older
hearth designs. This is to a certain extent offset by
the narrower and more highly rated boilers that
can be used with spreader firing.
Fifthly, the very nature of spreader firing means
that higher excess air is used in the furnace than
is obtained with, hearth furnaces. The basic figures
are probably 40 % and 30% respectively, in other
words one third more excess air is used in the case
of spreader firing and this increases the chimney
losses. Again this feature is offset by the fact that
furnace temperatures are very much lower with
reduced radiation losses.
The advantages of spreader firing far outweigh the
disadvantages and they can be summarised as follows:
Firstly, there is the intimate mixing of fuel and
air in suspension with consequent flash drying.
This brings with it the almost complete elimination
of unburnt gases.
Secondly, we have comparatively low furnace
temperatures with a consequent reduction in brickwork
maintenance. Also under this heading should
be mentioned the ability to use partially water
cooled walls which again have the effect of reducing
furnace maintenance.
Thirdly, the use of auxiliary fuels, such as coal or
wood waste, on the stoker is facilitated. Also in the
case of combined bagasse and oil firing, the heavy
slagging which is experienced with Ward Furnaces
is not here apparent due to lower furnace
temperatures.
Fourthly, it is possible to mechanise ashing
completely.
Fifthly, and as mentioned above, it is possible
to obtain a considerable reduction in boiler width.
Sixthly, the spreader fired boiler is highly flexible
and has "quick steaming" characteristics.
One can only come to the conclusion that, especially
where larger boilers over 50,000 Ib./hr. are concerned
and where higher steam pressures and temperatures
are to be used, the only possible economic solution
lies with the spreader stoker. However it is my
opinion that, where coal is only used to a limited
extent or where the auxiliary fuel is oil, a compromise
such as the Medine design described previously may
be a better proposition.
Yet another variation, which has arisen in my
experience, is the case of a boiler which is primarily
coal fired for long offcrop periods for irrigation
purposes and here we used a conventional backward
moving chain grate stoker with air spreader firing of
bagasse. In the actual case this method was adopted
because the boiler was secondhand and had an existing
chain grate stoker. For new plant the backward
moving travelling grate stoker is slightly more
expensive than the conventional spreader travelling
stoker which operates in the opposite direction. It
may therefore not be normally a very economic
solution, but the point is that the combustion of
coal on a conventional travelling grate or chain grate
stoker is more efficient than spreader firing. The main
reasons for this are lower excess air and lower grit
emission, both giving reduced chimney losses.
Future Developments
The cane sugar mill of the future may well develop
a steam demand in excess of 1,000,000 lb./hr. It will
then be necessary to consider boiler unit sizes of the
order of 300,000 to 400,000 lb./hr. of steam.
To the boiler designer this size of unit has no
problems whatsoever as units of ten times this size
are currently being installed in power stations all
over the world.
However the real problem is how we are going to
fire these larger units with bagasse and in my opinion,
the only possible solution can lie with pulverizing of
bagasse in milling equipment.
These mills would be swept with very high temperature
air and the powdered fuel fed into a bin system.

68 Proceedings of The South African Sugar Technologists' Association—March 1966
The dry pulverized fuel would be led to special
feeders which would distribute the fuel to burners set
in completely water cooled boiler furnaces.
We should not get too enthusiastic about this system
until such time as there is a call for much larger units.
I say this because it is unlikely that such a system
with its complicated controls and the necessary firing
skills, will become economic until a boiler size
of at least 250,000 Ib./hr. is reached. I envisage here
a minimum of three such boilers which would mean a
sugar mill handling something over 600 tons of cane
per hour.
A second development which I have in mind is
the possibility of somebody solving the problem of
bagasse storage. This would go a long way towards
increasing the efficiency of the use of bagasse and
effect a saving in the use of auxiliary fuels.
Thirdly and finally I would like to comment on the
possibilities of higher efficiency boiler plant being
employed. Quite a few mills are now using economises
as well as airheaters, particularly where higher boiler
pressures have been employed, but in sugar mill
practice we are still a long way off the order of
efficiencies employed in power station practice.
Especially where large quantities of coal are used,
consideration should be given to back end temperatures
of the order of 280° F. on coal, equivalent to
say, 320° F. on bagasse.
With this order of low exit gas temperature it is
probably better to place the Economiser after the
main airheater with possibly a small primary cast iron
airheater after the Economiser.
With these low outlet gas temperatures it will be
necessary to go to concrete, brick or gunite lined
mild steel chimneys.
Conclusion
I have by no means covered the whole field of this
most absorbing subject and have probably done less
than justice to some furnace designs. For any
omissions in this direction I can only apologise.
Mr. Magasiner: What order of boiler capacity
do you recommend for pulverised bagasse units?
Mr. Glennie: Definitely not below 250,000 lb./hr.
Mr. Hulett: It is mentioned in the paper that the
new large boilers with which the industry may be
equipped will require the bagasse to be further
pulverised. Final bagasse is already being pulverised
by about 5,000 h.p. before it reaches the boilers.
How much more horse power does the author suggest
would be required?
Mr. Glennie: It is extremely difficult to assess horse
power requirements for pulverising bagasse at this
stage. However, it should be borne in mind that the
use of dry fuel in firing boiler plant will increase the
efficiency to a very considerable extent and that this
will more than compensate for the extra horse power
requirements.
Mr. Hulett: The spreader stoker is recommended
but it is very often the cause of breakdowns. With
regard to furnace designs, I would like to mention that
the Eisner furnace has a simple brick construction
and requires very little maintenance.
Mr. Glennie: Returning to the question of future
higher capacity boilers, it should be pointed out that
if, for example, two 250,000 lb./hr. boilers are installed
instead of ten 50,000 lb./hr. units, or five
100,000 lb./hr. units, the capital and amortisation
charges will be considerably less. In the assumption
I have made for mill steam requirements, I have not
taken diffusion into account. However, I understand
that there would be certain savings of steam requirements.
Mr. Hulett: High velocity secondary air in spreaderfired
furnaces boosts the output of a boiler tremendously.
A twenty per cent increase was obtained in a
boiler at Darnall.
Mr. Glennie: When the bagasse spreader-fired units
were first introduced in the early 1950's, it was the
opinion of a number of people in the boiler industry
that these units required adequate provision for
secondary air. Subsequently, high pressure secondary
air was introduced by most manufacturers.
Mr. Magasiner: There is a lot more horse power
consumption if you use this high pressure, high velocity
air.

Proceedings of The South African Sugar Technologists'' Association — March 1966 69
STEAM AND VAPOUR DISTRIBUTION
by R. E. MARSH
General
Whole books, chapters of books, papers, correspondence
courses etc. etc. ad infinitum are available
on the subject of steam, its uses, generation, conservation
and distribution. It is not the purpose of
this paper to serve as a substitute for this literature
or even to quote from it more than is necessary. The
purpose of this paper is to consider the problems of
steam and vapour distribution with regard to the
Cane Sugar Industry in South Africa, with some
reference to economy.
From this point of view it is considered best to
follow the process of design and to consider the problems
in detail as they arise. This gives rise to the first
problem which is, where should the design start?
In some industries, textiles for instance, steam is
generated, distributed to whichever machines or
processes require it, possibly with some pressure
reduction where required, and what condensate can
be recovered is returned to the Hot Well lor boiler
feed. In such cases, all that is needed is a plan of the
building, showing the positions of the vessels and
boiler, and upon this with little trouble can be marked
the runs of the steam distribution. In a Sugar Mill this
is far from being the best point at which to start.
First of all, where turbine mill drives are used, the
steam not only is distributed but must also be recollected.
Secondly, due to multiple effect evaporation,
more than one type and "quality" of steam is available.
Thus the correct point at which to start the Steam
distribution design is that which is used in the Chemical
and Oil industries, namely the Flow Sheet.
Flow Sheets
Flow sheets have many advantages for initial
design. They are completely diagrammatic and, therefore,
a pipe which will wind tortuously through the
length of the Mill may be shown as a straight line.
Vessels and equipment may be shown in positions
which will make the reading and understanding of
the flow and process as simple as possible. In the early
stages of design, alterations and redirection of flow
can easily be made and understood. Pipes may be
sized in most cases before the general arrangement
drawings showing their position in the Mill are
commenced. This avoids the necessity for repositioning
piping on the building plan due to it being larger
than at first anticipated. And so on. and so forth.
It is in the Flow Sheet stage, and at no other stage
in the design procedure, that the amount of fuel
economy should be decided. Whilst, of course, it is
possible to make revisions to design at any time, even
after the completion of construction, it is at no time
cheaper or more easily accomplished than when the
flow sheets are being prepared.
The extent of the Steam and Vapour distribution
must now be considered. In this connection, the flow
of steam and. vapour must be followed through to the
bitter end, even in the condensed state in order to give
the complete picture. If maximum economy is to be
attained, every pipe which discharges to atmosphere
must be treated with suspicion and examined to
ascertain whether useful heat is being thrown away.
In the Cane Sugar Industry where the bagasse obtained
from the cane is used as fuel, it is generally considered
that fuel economy can be overdone and result in
enormous quantities of surplus bagasse for which
there is no market or use. This is a dangerous thought.
The actual balancing of fuel against demand should
be an operational problem and not one of design.
Design should be based on maximum steam economy.
It is very easy to waste steam even with a high efficiency
system, but extremely difficult and often expensive
to economise once construction is complete.
Further, it should be remembered that markets
seldom, if ever, exist for produce which is not available
except in theory. Therefore, it is preferable for
the flow sheets to be laid out on the basis of maximum
efficiency. It must be borne in mind, whilst doing so,
however, that less efficiency may be necessary under
certain circumstances. Unless there is a shortage of
water, reduction in efficiency is quite easily obtained
by blowing off steam to atmosphere, etc.
The main points which need serious consideration
in a Sugar Mill with regard to economy are:
(a) Exhaust steam should only be used for heating
where it is not possible to use a lower "grade"
heating media. This means in effect that it
should only be used for the first effect of the
multiple effect evaporator and for juice heaters
where the leaving juice temperature is such that
the use of other media would give too low a
temperature difference.
(b) Condensate from many of the vessels is at a
pressure and temperature above atmospheric
boiling point. In these cases the condensate
should be allowed to "flash" as soon as
possible after leaving the heating surface and
this flash steam collected and usefully employed.
(c) As all condensate is not required for boiler
feed and much in fact is not suitable for this
purpose due to contamination in multiple
effect evaporation, the surplus condensate heat
content should be used as far as possible.
Where, as is the case in some Mills, the condensate
is used for Maceration on the Mills,
it must be cooled before its re-use. In this case
the best method of cooling is by passing it
through a heat exchanger for primary juice
heating.
(d) Providing that the minimum process steam
requirement is sufficiently greater than the
maximum steam requirement for power generation,
the possibility of vapour compression on
the first effect of the evaporator station should
be examined.

70 Proceedings of The South African Sugar Technologists'' Association March 1966
When the flow sheets have been completed, consideration
can then be given to the preparation of
general arrangement and detail drawings and the
physical side of the pipework system. Before considering
this latter, however, some comment regarding
existing installations is considered necessary.
Where flow sheets exist from the original construction,
alterations and additions can be quite easily
planned, except for the final survey, in the comfort
of an office. Even where no flow sheets exist, it is
preferable to prepare at least a partial one of the part
of the piping which is affected. Upon this may be tried
the various alternatives until a satisfactory solution
to the problem, whatever it may be, is obtained and
only then is it necessary to embark on site measurements
and the preparation of working drawings.
The alternative of spending lengthy periods climbing
around steam pipes and steelwork either when the
Mill is working or when the weather is hot is not one
which is to be lightly undertaken. The Engineer of an
existing Mill can be sure that flow sheets will, in the
long run, prove to be worth a thousand times their
weight in perspiration.
We now pass to the design of the physical side of
the piping, that is to say, the determination of size,
route, wall thickness, location and size of valves etc.,
the order of procedure is generally as follows:
Pipe Sizing
This is a procedure which is quite often assumed by
the uninitiated to be difficult and highly technical.
For the Engineer who is faced with the problems of
the natural circulation of hot water, the gravity flow
of water supplies etc., this can often be the case. Indeed,
some steam pipe sizing problems can be tricky,
but for the Sugar Mill Engineer, who is concerned with
the sizing of additions to existing process mains in
the great majority of cases, no great difficulty exists.
A typical formula for the flow of steam through
pipes is:
Z,-Z„ --0.00010123 . L
(J5.027

Proceedings of The South African Sugar Technologists- Association - March 1966 71
This is, of course, enough to put anyone off for life.
fortunately however, pipe-sizing tables and charts
are available in profusion and the solving of such
formulae as that given above is very seldom necessary.
The available charts, it is necessary to point out,
seldom cover piping over 12 in. nominal bore and at
first glance this is unsatisfactory for a Sugar Mill
where a great proportion of the low pressure process
steam piping is much larger than this and may even
be as large as 48 in. diameter.
The steam pressures with which one has to deal in
the Sugar Industry in general cover a wide range.
On the high pressure side (steam for power) pressures
may vary from 100 p.s.i.g. to 450 p.s.i.g. depending
to a certain extent on the age of the Mill as the more
modern trend is towards higher pressures with the
trend towards higher efficiencies. On the low pressure
side (exhaust steam etc. for process) pressures above
15 p.s.i.g. are not normally encountered and the final
vapour from the multiple effect evaporator may he
at a pressure of 2 p.s.i.a. (26 in. Hg. Vacuum), in
the case of the high pressure side, piping is seldom
greater than. 12 in. diameter and recourse is best made
to tables or charts for this piping. Two things must be
checked. Firstly the pressure drop, which should not
exceed 1 p.s.i.g. per 100 ft. run at JOOp.s.i.g. At higher
pressures, a greater pressure drop may be tolerated,
but care should be taken to check that this will not
be so great as to effect the performance of the prime
movers. Secondly, the velocity should be kept below
100 to 130 ft. sec. for saturated steam and 150 to
200 ft. sec. for superheated steam, regardless of the
consequent pressure drop. In other words, if pipe sizes
are selected to satisfy both the pressure drop and
velocity conditions, the larger size should be selected.
There are two reasons for limiting the steam velocity,
namely to prevent excessive erosion of bends and to
prevent excessive noise due to Row in the piping.
For the low pressure side, a different approach can
be adopted. A few examples will show the reason
why:
(i) For the same initial pressure and velocity, the
pressure drop decreases with increase in pice
size:

72
Thus: For 15 p.s.i.g. and 100 ft/sec.
2 in. pipe pressure loss = 1.07 p.s.i.g./100 ft.
4 in. pipe pressure loss = 0.48 p.s.i.g.,/100 ft.
6 in. pipe pressure loss = 0.30 p.s.i.g.,/100 ft.
(ii) For the same initial pressure and pressure drop,
the velocity increases with increase in pipe
size:
Thus: For 15 p.s.i.g. and 1 p.s.i.g. 100 ft.
2 in. pipe velocity — 94 ft/sec.
4 in. pipe velocity = 147 ft sec.
6 in. pipe velocity •-•-•- 194 ft sec.
As few low pressure pipe runs exceed more than
200 ft. in the average Sugar Mill, and it can be assumed
that a total pressure drop of 1 p.s.i.g. can be
tolerated on an average, it will be seen from examination
of the above figures, that even for a pipe as small
as 4 in. diameter, the pressure drop will be acceptable
if a velocity of 100 ft sec. is used for sizing a 15
p.s.i.g. main. One other aspect needs consideration
and that is the effect of initial pressure:
For equal velocity and pipe size, pressure drop
decreases with increase in initial pressure:
Thus: For 4 in. dia. pipe and 100 ft/sec.
15 p.s.i.g. pressure drop - 0.48 p.s.i.g. 100 ft.
10 p.s.i.g. pressure drop =0.40 p.s.i.g. 100 ft.
5 p.s.i.g. pressure drop - 0.35 p.s.i.g. 100 ft.
Thus, it will be seen that the tendency towards the
velocity being the controlling factor in this class of
pipe sizing is accentuated by a lower initial pressure.
Proceedings of The South African Sugar Technologists Association — March 1966
FIGURE 3: Control for Heating Application.
However, due to the lower density at lower pressure,
higher velocities can be tolerated and this tends to
offset this last characteristic.
The above examples have been given merely to
prove the general trend. Without going into further
details, it can be accepted that the low pressure side
of the steam piping in a Sugar Mill may be safely
sized purely on the basis of velocity for piping over
6 in. diameter. Below this size, a check should be made
on the pressure drop. In addition, where exceptionally
long runs are encountered, it may be necessary to
carry out a check on larger piping possibly even up
to 12 in. diameter. This will rarely be justified.
Opinions vary considerably on what velocities should
be used as a design basis. Probably the most representative
figures are those given by Oliver Lylc in
"Efficient Use of Steam". On condition that the
exhaust steam pressure is 15 p.s.i.g. and is dry saturated,
then the sort of velocities which can be used, as a
design basis are as follows:
15 p.s.i.g. (exhaust steam) 100 to 130 ft/sec.
5 p.s.i.g. (1st vapour) 120 to 160 ft/sec.
0 p.s.i.g. (2nd vapour) 150 to 180 ft/sec.
12 in. Hg. Vac. (3rd vapour) 160 to 220 ft/sec.
26 in. Hg. Vac. (4th vapour) 220 to 350 ft/sec.
Obviously these figures must be adjusted for particular
cases but serve as an indication of normal sizing
figures for this sort of application.

Proceedings of The South African Sugar Technologists'' Association — March 1966 73
Once the velocities are decided, it is only necessary
to apply the formula Q =-•-. AV, where Q is the quantity
in cusecs, A is the cross-sectional area of the
pipe in sq. ft. and V is the velocity in ft/sec. This may
be converted to a more convenient form for steam,
pipe sizing thus:
The final selection of pipe sizes depends, of course,
on the standard sizes available. It will be noted from
references to B.S. 10:1962 which gives details of
standard pipe flanges up to 72 in. diameter, that an
enormous range of sizes are standard. However, up
to 12 in. diameter, 7>{ in., 7 in. and. 9 in. should be
ignored as these are not normally manufactured.
The latter two sizes sometimes occur on pumps but
as a reducer is invariably fitted directly to the pump,
they can be ignored from the point of view of actual
piping. Above 12 in. it is normal to use outside
diameter sizes, and although a considerable range of
these is given, it is as well to standardise somewhat
further, and a suggestion in this respect is that 16 in.,
20 in., 24 in., 30 in., 36 in., 42 in. and 48 in. should
suffice. The use of sizes at closer intervals has really
only an academic advantage and only complicates
maintenance etc.
Having established the pipe sizes, we must now
consider:
Pipe Wall Thickness
This is adequately covered by British Standards
1387 and 806. These standards should be rigidly
adhered to as far as all high pressure piping is concerned.
On the low pressure piping however, more thought
must be given to the matter. Piping over 12 in. diameter
must be imported and is often expensive.
Trouble may be encountered for piping between 8in.
and 12 in. diameter. On low pressure piping, there is
no reason why the larger piping should not be fabricated
locally, and this course is normally adopted.
The only question which may arise is in respect to
the wall thickness. What should this be? If the formula
given in B.S. 806 is applied to 15 p.s.i.g. working
pressure for a 24 in. diameter pipe we arrive at a
wall thickness of 0.0146 ins. which is approximately
28 s.w.g. Obviously no one in their right senses would
contemplate either manufacturing or installing such a
pipe on steam service. Other factors than pressure
therefore become critical. Firstly, it is necessary that
the piping shall be self-supporting over a reasonable
length, secondly it must take stress set up by expansion
and thirdly, corrosion must be taken into account.
Generally for the low pressures and comparatively
moderate temperatures involved, a wall thickness of
i inch will be adequate. This will also meet normal
corrosion encountered on steam services. Under some
circumstances, such as where there is carry over of
sugar in vapour on the latter effects of the evaporator
station, a greater wall thickness may be considered
necessary but this is a matter for individual consideration.
Pipe Bends
On larger size piping, it is usual to employ fabricated
"lobster back" bends as illustrated in Figure
1. From the economic point of view, it is obviously
preferable to have as few "cuts" or "pieces" as possible.
However, most engineers have been thoroughly
frightened at an early and impressionable age by the
bogey "extra resistance". It has been instilled into
them that the biggest crime that they can commit on a
piping job is to unnecessarily increase the resistance
and pressure drop. On systems where limited gravity
flow is available, this may possibly be true in extreme
cases. On natural circulation for hot water heating or
hot water supply, this can very often be the case,
especially as the head available to produce circulation
may be as low as 1.0 in. w.g. However, the case of
steam distribution is somewhat different. Figures and
formulae for the pressure loss in bends are available
but not for the various types of bends shown in Figure
1 for steam service. This type of bend is, however,
used quite frequently for air ducting (which is one
case where resistance can be very important) and
figures for this service are available. As the density
of air at sea level is approximately 13.5 ft 3 /lb. and
the density of steam at 15 p.s.i.g. is 13.73 ft 3 /lb. it is
reasonable to assume that the figures for air may
be applied to steam in order to ascertain how much
difference is caused by the number of "pieces" used
for a bend. The figures thus obtained may not be
100 per cent accurate but will prove sufficiently so
for our purposes. The figures given by the I.H.V.E.
"Guide to Current Practice" for the bends illustrated
are as follows:

74 Proceedings of The Soutli African Sugar Technologists' Association —March 1966
As 1 p.s.i.g. = 27.6 in. w.g. it is obvious that unless
there are an enormous number of bends in a run
of piping the difference is not worth worrying about
and certainly not worth the additional expense involved
in fabricating radius (or near to radius) bends.
It should also be borne in mind that the insulation
of bends is also more expensive the nearer they are
to radius type, especially if sheet metal cladding is
used.
Pipe Expansion
The subject of pipe stressing is lengthy and involved.
A discussion of this subject would very probably
double the length of this paper and would not really
be of interest to the Sugar Engineer. Suffice to say
that the stressing and layout of high pressure piping
should be left to the specialist. Low pressure piping
is not subject to the large expansion of high pressure
piping and here, if normal good practice is followed,
difficulties are seldom experienced. The piping should
be left as free as possible to move with expansion,
with plenty of length and bends on branches connecting
to equipment. Anchors should, be kept to a minimum
and expansion devices fitted wherever it is
necessary to confine a straight. length of piping.
To be more specific, it is the following points
which should be given particular attention:
(i) Exhaust pipe connections to turbines must impose
as little stress as possible on the turbine.
Not only must the expansion of the piping
be considered but also the movement of the
turbine casing. The use of expansion bellows
is indicated in this case. However, it should be
noted that the internal pressure in bellows acts
on the internal faces of the corrugations and
results in a longitudinal force which tends to

Proceedings of The South African Sugar Technologists' Association — March 1966
lengthen them and which must be overcome in
order to compress them to absorb expansion.
This force may be transmitted onto the turbine
unless a very complicated method of
containing the expansion is adopted. For this
reason it is better to use the lateral (or sideways)
movement type of bellows on this application
as the forces needed to move the bellows
are very much less.
(ii) Connections to vessels and equipment should be
designed so as to keep the stress acting on the
vessel body to a minimum. Long branches
from the main with one or two bends will
usually suffice but each case should be considered
separately. On one hand it must be
realised that the force exerted by a 24 in. pipe
when heating up from 60° F to 250° F is 377
tons, unless it is allowed to expand freely in
which case the increase in length will be 1.6 in.
per 100 ft. Both these figures should be halved
by the use of "cold draw". This means that
when the piping is installed, each run should be
left with a gap between flanges equal to half
of the calculated expansion. The flanges are
then drawn together cold and the piping thus
stressed in the opposite direction to the expansion
movement, so halving the hot stresses.
Even under these circumstances, there are some
pretty frightening forces still at work and these
should be as far as possible allowed to dissipate
themselves on the piping. Where this is
not possible and considerable forces act on the
vessels, consideration should be given to the
use of expansion devices to alleviate this. On
the other hand,a visit to a few of the Sugar Mills
in Natal will show that considerable stresses
are exerted on many existing vessels without
any detrimental effect. However, it is not advisable
to assume that this is not therefore worth
worrying about. When in doubt it is as well to
check with similar cases in existing installations
and if not 100 per cent satisfied, to play safe
and fit expansion devices.
Pipe Supports
These should allow the free movement of the piping
for expansion except where anchors are particularly
required. Detailed design of supports is not a subject
for a paper such as this as supports may vary considerably
in detail. However, basically they fall into
three main types. These are:
(i) Hanger Supports
Providing the hanger is of sufficient length,
these allow free expansion of the piping for
both longitudinal and transverse movement.
Where vertical piping occurs the connected
and adjacent horizontal piping will also be
subject to vertical movement. In this case,
spring type hangers should be used in order
to accommodate this movement.
(ii) Rollers
Horizontal runs may be supported on rollers
which will allow free movement of the piping.
(iii) Sliding supports
Where these are employed, the pipe should be
fitted with a shoe, which should be of sufficient
depth to allow the insulation to clear the supporting
bracket and should also present as
little area of contact to the support as possible.
The pipe supports should be arranged to give a fall
on the pipe in direction of flow. Condensate is always
present in the bottom of the pipe due to condensation
to meet heat losses from the pipe wall. This must be
drained. On the smaller piping, failure to provide
adequate fall and drainage will result in the build up
of condensate which may result in the formation of a
"slug" which, driven along by the steam at high
velocity, will cause hammer and possibly damage to
the piping. On larger piping, this is not so possible
and the fall provided may be less. In fact on the largest
piping, a level run may be quite satisfactory although
condensate left behind when the plant is shut down
may give rise to corrosion problems.
Steam Valves
In the past, valves for larger piping have presented
difficulties. Valves for 24 in. piping are extremely
bulky and weighty. Further, in the past, gate or
sluice valves have been the only type available although
some large globe or angle valves have been
manufactured where tight shut off is essential. Butterfly
valves have now become available and solve some
of the problems previously encountered. They are
much less bulky, much lighter and are more easily
operated. Further, for low steam pressures they can be
arranged for tight shut off. Butterfly type non-return
valves are also available. These enable automatic
isolation where required (such as on turbine exhausts)
and where automation is required, perform this task
more economically. With this wider range of valves
available, the Engineer may find easier solutions to
many of his previous problems especially for automatic
operation.
Whatever type of valve is used, the sizing of these
must be carried out entirely separately from the piping.
This does not of course apply to valves which are
intended purely for isolation or change-over purposes.
These are normally of the same size as the piping.
Control valves, however, are an entirely different
matter. Globe or butterfly type valves should be used
for control, as the control characteristics of gate
type valves are very poor. There exist a number of
pressure reducing valves which have never worked
properly. The manufacturers usually are blessed with
the blame for this. In the great majority of cases, they
do not work due to the fact that they are oversized,
being the same size as the piping, or have been
incorrectly installed. An oversized control valve will
spend most of its life open only a fraction of its full
travel and vainly trying to achieve control under these
circumstances whilst its seat is destroyed by wiredrawing.
A few examples of control problems are
given to illustrate why the valve sizing should be
considered separately:
Reference is made to Figure 2. This is a water
control problem (purely hypothetical). The control
75

76 Proceedings of The South African Sugar Technologists' Association — March 1966
valve is assumed to be of the butterfly type, as figures
for various degrees of opening are available for this
type which enable the full picture to be given. If the
resistance of the piping is 8.5 p.s.i.g. at full flow of
150 g.p.m., then the following figures may be tabulated:
Full flow
4 in. pipe loss 8.5 p.s.i.
4 in. valve loss 0.44 p.s.i. at 60° open (full open)
4 in. valve loss 1.50 p.s.i. at 46° open
3 in. valve loss 1.40 p.s.i. at 60° open
From this it can be seen that the 4 in. valves, sized
to suit the pipe will be a quarter shut even at full
flow. The 3 in. valve will be virtually fully open at
full flow.
Half flow
4 in. pipe loss
4 in. valve loss
3 in. valve loss
2.10 p.s.i.g.
7.90 p.s.i.g. at 19° open
7.90 p.s.i.g. at 27° open
As the head available is constant at 10 p.s.i.g. this
must be the resistance of the system at all flow conditions,
and as the piping resistance drops with reduction
of flow, the valve must compensate for this.
From the half flow figures it will be seen that the smaller
valve achieves this with less movement than the
larger, and thus has a greater movement available for
the control of very low flows where accuracy is often
important.
What the above examples illustrate is the better
quality control that can be achieved by the use of the
right size of valve which is almost always smaller than
the piping. Not only is the smaller valve better, it is
also cheaper! This is one of the few examples in this
modern age where the better job costs less.
It should be mentioned that were the level of water
in the top tank shown in Figure 2 to rise, the difference
in control quality given by the different valves would
be accentuated. On pumped applications this is particularly
important as the pump head tends to increase
with reduced flow. For good quality control, the valve
resistance should be between 10 and 20 per cent of the
total system resistance.
With steam (and gases) a slightly different approach
is necessary. Control valves on these services have
critical sizes. A particular size of valve has a maximum
flow capacity; for a particular inlet pressure, regardless
of the pressure differential across the valve, this
maximum cannot be exceeded. If a valve is selected
as nearly as possible to this critical capacity, control
will be of the best quality possible. Figure 3 illustrates
an example of this.
The following figures show the action of firstly
a valve sized to suit the piping, and secondly, a valve
selected for critical flow. Steam temperatures equivalent
to the outlet steam pressure are quoted as it is
assumed that the heat transfer, and thus the steam
flow will be proportional to the temperature difference.
The problem shown is one of temperature control
in a tank.
Full Flow
6 in. valve 60° open (full) 0.33 p.s.i. drop 250° F
3 in. valve 60° open 7.10 p.s.i. drop 235° F
Half Flow
6 in. valve 25° open 16.00 p.s.i. drop 210° F
3 in valve 43° open 18.00 p.s.i. drop 203° F
Quarter Flow
6 in. valve 17° open 20.5 p.s.i. drop 190° F
3 in. valve 31° open 21.5 p.s.i. drop 186° F
It will be noticed that the differences in valve opening
are very pronounced in this case. However, it
will also be noticed that even at half flow, the outlet
pressure of the control valves must be below atmospheric
pressure. Unless a pump is installed on the
condensate or sufficient height is available for
natural gravity extraction, what will actually happen
is that the steam coil will flood, thus reducing the
effective heating surface. Another point is that although
the 3 in. valve may be cheaper, the pressure
loss through this even when full open will necessitate
an increase in heating surface which may result in an
overall increase in cost of the plant.
On high or medium pressure heating systems, this is
not so much the case and the critical flow valve is the
one to use. With the low pressure met in Sugar Mill
process heating, it is preferable to fit the control
valve on the condensate outlet. This gives flooding
control, thus varying the effective heating surface to
suit requirements. It also ensures that the full steam
pressure is always available both for heating and for
condensate disposal. As this valve is to handle condensate,
with a certain amount of flash steam, it will
be considerably smaller and consequently cheaper than
a steam control valve on the inlet to the coil.
Figure 4 illustrates the control of the steam supply
to an evaporator. Flooding control cannot be used on
this application as it has been found in practice to be
impractical. Neither can the valve be selected for
critical flow. Such a valve for this application would
be 14 in. and would have a full open (60°) pressure
drop of 7.2 p.s.i. As in this example the total pressure
difference available is only 9 p.s.i. this would result in
such an increase in calandria heating surface as to be
completely uneconomical. However, on such an application,
variation in evaporation rate does not cover
the full range from maximum to nil. Under operating
conditions, a reduction from maximum to half
should prove more than sufficient. Therefore, the
valve may be selected more from an economic point
of view with due regard to the available pressure
difference. For instance a 30 in. valve would have a
full open pressure drop of 0.19 p.s.i., a 24 in. valve
would have a pressure drop of 0.50 p.s.i. The latter
should be quite suitable and under some circumstances,
an even smaller valve may be practical.
The moral of the foregoing is quite clear. Never,
never size a control valve to suit the piping. Consider
it's location carefullly. If information is not available
to the Engineer to enable him to size the valve himself,
then he should give the manufacturers full details
of application, steam quantities, allowable

Proceedings of The South African Sugar Technologists' Association —
March 1966 77
pressure drop etc. Never, never consider the pipe
size. Reducers are relatively cheap. Oversized valves
are expensive, especially if they have to be thrown
away and replaced.
Thermal Insulation
Thermal insulation or lagging is required to satisfy
three basic conditions:
(i) to keep heat losses to a minimum to prevent
wastage of fuel;
(ii) to prevent excessive heat loss from hot surfaces
creating uncomfortable temperatures in
the building;
(iii) to prevent operating personnel suffering injury
from burning by hot surfaces.
Where fuel has to be purchased, (i) is of prime
importance and insulation design practice is based on
this. Any insulation designed on this basis will automatically
take care of (ii) and (iii). Where, in a Sugar
Mill, there is a surplus of bagasse, the importance of
(i) tends to fall away unless a ready market is available
for the surplus. However, in order to satisfy condition
(ii) it will be found in practice, that very little
reduction in insulation will be possible. Possibly some
of the condensate piping may be left uninsulated but
only 7 ft. above the floor as otherwise condition (iii)
will not be met. Even with an average insulation
efficiency of 85 per cent throughout (including vessels)
the heat losses from hot surfaces in a 250 ton/hr.
Sugar Mill will be equivalent to about 1,500 lbs. steam
per hour. This represents a lot of heat. Therefore it
is suggested that normal insulation design be used
even when there is a bagasse surplus, although in such
a case, efficiencies may be lowered and cold face
temperatures raised in certain cases. Care should be
taken when designing insulation to consider both the
efficiency and cold face temperature. Surface coefficients
play a large part in heat transfer, and it
should be noted that whilst the surface temperature
of sheet metal covered insulation is higher than that
for a cement finished insulation of the same thickness,
the efficiency is higher. This is due to the fact
that the proportion of radiated losses and convection
losses is different.
The use of sheet metal cladding as insulation on
piping is now almost an accepted standard. This
finish is not expensive and has the virtue of being far
more durable than other finishes.
Whilst this is outside the scope of this paper, it is
felt necessary to mention that condition (iii) should
be considered with relation to the clarified juice
piping. This has a surface temperature of up to at least
220° F in parts and this temperature is dangerous to
personnel. When one considers that all but the most
hardened humans cannot tolerate more than a temperature
of 110° F when washing their hands, the
danger is brought home to one more effectively. Any
surface at a temperature greater than 180° F should be
at least provided with a token insulation to protect
personnel unless it is at least 7 ft. above the floor.
Condensate
As mentioned earlier in relation to flow sheets,
condensate is essentiaLly a part of the steam and vapour
distribution. Detailed consideration of the condensate
system would, however, it is felt, be out of place
in a paper directed mainly at the actual steam side of
the installation. Only a few points are therefore dealt
with here.
Condensate is corrosive. Any idea that the Engineer
may have of it being pure distilled water, may be
safely dismissed from his mind. Condensate from the
latter effects of the evaporator station may be contaminated
with sucrose and therefore acidic. Even
with a "clean" system, the condensate often picks
up carbon dioxide or other gases and forms an acidic
solution. In hospitals and industries where the condensate
piping is fairly small in size, although extensive
in run, it has been found time and time again
that copper piping pays in the long run due to lower
maintenance and replacement costs. The absolute
minimum specification for condensate piping has been
found to be galvanised mild steel. With the larger
sizes of piping encountered in a Sugar Mill, copper
piping is not available. Therefore galvanised mild steel
piping should be used. A lighter tube may be used
(medium weight) than if black piping were employed
and the cost is therefore much the same. The life of
the piping will be extended considerably and in the
long run financial savings will accrue.
In a Sugar Mill, condensate must, sooner or later
reach a stage at which it must be pumped in order to
return it to the boiler feed tank or to dispose of the
surplus in other ways. It must be noted that the
condensate may be near to atmospheric boiling
temperature and if the pressure in the pump suction
pipe or in the pump casing should drop below the
pressure at which the condensate temperature is equal
to boiling point, cavitation will occur in the pump.
This is not only noisy, it is disastrous for the pump
itself and can result in the necessity for the complete
replacement of the unit. Wherever condensate, or
any hot liquid for that matter, is to be pumped,
sufficient positive pressure must be provided at the
suction of the pump to prevent cavitation. This will
mean generally fitting the pump well below the collection
tank from which it is drawing. The pump
manufacturers should be consulted on this matter as
different pumps have different requirements in this
respect.
The bane of the life of an Engineer responsible for
maintenance, is the task of maintaining vast numbers
of steam traps. Fortunately, from this point of view
vessels working under vacuum do not need traps as an
atmospheric leg and seal will suffice, unless, of course,
the height is not available for this in which case an
extraction pump must be fitted. The draining of steam
mains is of course one case where nothing can be done
to reduce the number of traps. The case of the larger
vessels, however, deserves more consideration. It is
the practice in some cases to avoid the use of traps
altogether by using U-tubes. These devices have it to
be said in their favour that they are simple. That they
are not the ideal solution is also quite obvious. Apart

78 Proceedings of The South African Sugar Technologists' Association — March 1966
from their being unsightly, untidy and cumbersome,
they will not pass air. This is an extremely important
point. It is assumed that in this enlightened day and
age, every engineer dealing with steam is aware of the
evils of air being present in any steam heating system.
Steam traps are, of course, available with thermostatic
air releases. Normally, group trapping, that is to say,
one trap serving two or more vessels is completely
unsatisfactory, due to the fact that any difference in
working pressure between the vessels will result in the
condensate being unable to flow from that having the
lower pressure. There is one case where this does not
apply and that is when the steam trap is located at a
lower level and the condensate piped separately from
each vessel to the trap. The height of the vessel above
the trap must be such that any differences in pressure
can be balanced by a water column in the respective
vertical condensate pipes. If sufficient height is available
and this system can be adopted, separate traps
need only be provided where it is essential to segregate
the various "qualities" of condensate. Thus a trap
would be necessary for exhaust steam condensate.
A separate trap for first effect vapour condensate
would be needed as this may be clean most of the time
but is subject to occasional contamination. Also, in
some cases, this may be capable of producing flash
steam in which case, it must be separated from cooler
condensates. Finally, a trap is necessary for the other
grades of condensate. Thus with this system, a minimum
of three traps are required. In practice it may
prove necessary to provide four or five but even this
number is reasonable especially if they can be located
in a group. This system does away with U-tubes and
their disadvantages, and at the same time centralises
the main traps for maintenance. Unfortunately,
standard steam traps cannot be used as they are not
available for the required capacities. However, there
is no reason why purpose-made traps cannot be used
for these large sizes. The separate condensate pipes
should be brought to and connected into a closed tank
or cylinder. Each pipe should be fitted with a check
valve to prevent back flow into idle vessels. The pipes
should be carried down in the tank to below the
design water level to ensure a seal. All that is needed
to make this into a steam trap is an automatic valve
on the outlet controlled by level in the tank so that
if the water level falls, the outlet valve shuts. In
addition, the trap should be able to pass air. An air
vent on the top of the tank takes care of this and to
meet all possibilities this should consist of a thermostatic
device in series with a float device, thus preventing
the egress of both steam and water. It should be
noted that these external devices are easily accessible
for maintenance unlike the standard steam trap
which must be opened up for servicing. Whilst this
system may not be the ideal answer to the problem,
it is felt that it does tend to alleviate the maintenance
problem whilst retaining the advantages of proper
steam trapping.
Summary
Whilst, needless to say, piping design should generally
follow the procedure followed on all engineering
design of:
Preliminary examination
General Arrangement drawings
Detail drawings and
Specification of Materials,
it is suggested that in the case of piping, the general
arrangement drawings should be preceded by the
preparation of flow sheets. Further, particular attention
should be paid to the sizing of low pressure piping
(which may need a different approach than that given
to high pressure piping), the economics of the bends
to be used, the sizes of control valves, expansion,
supports and insulation. The condensate system also
needs particular attention in respect of pipe material,
pumping arrangements and steam trapping.
It is further suggested that, the more attention that
is given to the above in the design stage, the less will
be the amount of attention required during the construction
and the less the likelihood of alterations to
the actual installation.
References
Institution of Heating and Ventilating Engineers, (1959).
"A Guide to Current Practice".
Lyle, Oliver, (1947). "Efficient Use of Steam".
British Standards Institution (1954) "B.S. 806".
British Standards Institution (1962) "B.S. 10".

Proceedings of The South African Sugar Technologists' Association—March 1966 79
CONDITIONING BOILER FEEDWATER
Introduction
In the opening address of the annual congress of
your association last year, Dr. Aston. R. Williams
quoted Sir Richard Livingstone as having said in
1954:
"In all subjects, not only in technology, there
tends to be too much detail put in. If you are going
to study a subject you must know the general
principles behind it. You must know the way in
which to learn all about it."
Dr. Williams concluded his address with:
"In an age of specialisation it is all the more, not
all the less, important to understand basic principles."
You are sugar technologists and your members
comprise people engaged in numerous specialised
occupations all linked with the production of sugar.
This paper deals with a specialised activity within
your industry, a vital process to you all, because the
whole mill, the refinery, the personell engaged, all
rely on the boilers providing steam.
The term "water conditioning" is being used more
and more now in preference to the term "water
treatment". "Water conditioning" is the application
of proved scientific methods to render a water suitable
for a specialised industrial usage.
Water and the Sugar Mill
Water is as much a raw material in the manufacture
of sugar as sugar cane itself. Both of these raw
materials are of variable compositions and require
regular control testing. Mill laboratories are naturally
occupied in testing all stages of the manufacture of the
final product. Efficient modern practice calls for
regular control testing of water quality, particularly
the quality of the boiler feedwater and the control of
feedwater conditioning.
Water conditioning has developed practical techniques
that permit of "Tailor made" water, a feature
quite unheard of thirty years ago. What we are required
to do is to integrate the new techniques into
the industry, to obtain maximum use of the equipment
our predecessors have left us. This equipment is still
of immense value in practice in the mills processing
cane today, no matter what value the accountants
have recorded in their balance sheets. Many mills
are installing new boilers and the fact that you are
holding this symposium is evidence of the importance
you are placing on the efficiency of steam generation
in your industry.
FOR THE SUGAR MILL
by G. E. ANGUS
The extraction of sugar from sugar cane depends
on steam being available at the mill. The site chosen
for a mill depends on a number of factors. Among the
principal of these is the positioning on the sugar estate
in relation to transport of cane to the mill and also
the availability of a source of water of sufficient
quantity, and. of suitable quality for the general water
requirements of the mill and residential areas adjacent
to it.
The class of water utilised is in general classified as
"lowland surface water".
This class of water depends on catchment areas.
By the very nature of its origin, and the course of the
river or canal, it takes up a number of substances in
solution or suspension.
Lowland surface water may contain excessive
quantities of mud and silt in suspension, and variable
quantities of mineral salts and vegetable matter. The
mineral salts are derived from the rocks and soils
and farmlands through which the water passes.
In lowland areas also, very frequently swampy lands
contribute organic contamination, and excess of
fertilisers from farmlands sometimes complicates
matters.
Components of water
Completely pure water is non-existent.
The impurities in water may be roughly classified
as:
Dissolved solids.
Dissolved gases.
Suspended matter.
Dissolved Solids
The minerals which water picks up from rocks and
soil consist chiefly of:
Calcium carbonate (limestone), CaC03.
Magnesium carbonate (magnesite) MgC03.
Calcium sulphate (gypsum) CaS04.
Magnesium sulphate (Epsom salts) MgS04.
Silica (silicates), Si02.
Sodium carbonate (soda ash) Na2C03.
Sodium chloride (common salt) NaCl.
Sodium sulphate (Glauber's salt) Na2S04.
These are soluble in water under various conditions
and constitute the inorganic dissolved solids, i.e. the
inorganic residue that is left when a filtered water is
evaporated to dryness. To the above must be added
any dissolved vegetable matter, and also material
derived from trade effluents.
Dissolved Gases
All natural water contains dissolved gases.
The principal gases are;

80
Oxygen.
Carbon dioxide.
Oxygen is soluble at atmospheric pressure and
ambient temperature to the extent of approximately
9 parts per million by weight (p.p.m.). Carbon dioxide
can be absorbed from the air to the extent of approximately
10 p.p.m.
The solubility of gases in water decreases as the
temperature is raised.
Carbon dioxide dissolved in water reacts with chalk
or limestone to form soluble bicarbonate of lime,
and it is this form that we find in water supplies.
Suspended Solids
These are materials in suspension, such as finely
divided vegetable matter, and silt.
Units and Terminology
In view of the fact, that in the past and also in
various countries at the present time, a number of
different methods are used for describing the quantity
of substances found in water, it is only natural that
an attempt at standardisation has been made.
We in the Republic of South Africa in common with
a number of overseas countries, including the United
Kingdom and the United States of America, have
adopted as our unit, "parts per million by weight"
(p.p.m.) and this unit will be used in this paper.
Each profession has its own technical terminology.
The following terms commonly used in analytical
reports and discussions dealing with the quality of
water for industry are set out to simplify matters.
They will be used in this paper.
The scale forming substances, with the exception of
iron, alumina and silica, contribute what is called the
total hardness of water or "H" figure. The terms
"temporary" and "permanent hardness" that one
sees in many of the older textbooks, are being replaced
today by the terms, "carbonate" and "non-carbonate
hardness".
Total hardness = carbonate hardness + non-carbonate
hardness.
Total Hardness is also sub-divided into: calcium
hardness and magnesium hardness or CaH and MgH
figures.
The terms "hardness" is a relic from early days
when it was used in relation to soap consuming power
in laundry and domestic usage.
Other commonly used terms and symbols used in
industrial water analyses are as follows:
T.D.S. Total Dissolved Solid matter.
"P" Alkalinity to Phenolphthalein.
"M" Methyl orange alkalinity or
Total Alkalinity.
P(BaCl2)
or "O" Hydroxides (2 P-M).
"S" Soda Alkalinity (M-H).
P04 Phosphates.
NaCl Chlorides.
pH Hydrogen ion concentration
(degree of acidity or alkalinity)
Proceedings of The South African Sugar Technologists' Association—March 1966
For convenience in calculations, H, CaH, MgH, P,
M, O and S are expressed in terms of CaC03.
pH values range from 0 to 14.
Values from 0 to 7 indicate decreasing acidity.
Values from 7 to 14 indicate increasing alkalinity.
pH 7.0 indicates neutrality.
Boilers in Sugar Mills
Boilers in Sugar mills are what are known as water
tube boilers.
In a water tube boiler, the air and fuel (bagasse
and/or coal) are burned in a furnace. The hot products
of combustion are guided from the combustion
zone over the exterior surfaces of the water-tubes. As
the gases flow through the boiler they are cooled by
transferring their heat to the water contained in the
tubes. The tubes are in nests inter-connecting cylindrical
pressure vessels called "drums". A mixture of
steam and water is formed in the tubes and considerable
velocity of flow is attained as the water and steam
move to the drum where steam is separated for use.
The water which has become concentrated by formation
of the steam is diluted by incoming feedwater
and returns to the steam generating tubes through
special drums. Modern water-tube boilers often are
provided with water walls. These are essentially cages
of vertical tubes that surround the furnace areas and
are supplied by down-comers and headers with water.
These discharge their mixtures of steam and water into
the steam drums as described above.
It is natural that in the older mills there are numerous
boilers which by modern standards are oldfashioned.
These have been in use for many years
and are still good. They form the back-bone of the
power stations. Modern boilers are often added to
these power stations but do not necessarily replace
the old boilers.
The operating pressures of these older boilers vary
in general from 160 to 200 p.s.i.g.
Modern boilers operating at 400 p.s.i.g. are in use
at several mills, and being erected at others.
It is noted that at the factory of Triangle Limited
in Rhodesia the boilers operate at 475 p.s.i.g.
There is a considerable difference between the
conditioning required for the feedwater to the old
boilers operating at approximately 200 p.s.i.g. and
that required for the boilers operating at 400 p.s.i.g.
and higher.
Feedwater
In a sugar mill the feedwater consists largely of
condensate, i.e. distilled water. In a mill in full production
more than enough condensate is provided
from various sources to provide 100 per cent of the
feedwater and have some over. It is claimed that
some mills are near to attaining this in practice. There
are inevitable steam losses as we all know. The extra
condensate is derived from the cane in the form of
water in the cane and in make-up water used in mill
tandems for leaching the cane to form the mixed
juice. This water is recovered by evaporation and
condensation. The magnitude of the percentage

Proceedings of The South African Sugar Technologists' Association—March 1966
which can be made available to the boiler is determined
by the purity of the condensate. The main
contaminants are sugar, lubricating oil and cooling
water.
Mr. J. Bruijn 2 of the Sugar Milling Research
Institute reports: "In the alkalised feedwater, sugar
tends to decompose into sugar acids and other products
which lower the pH and in this way increase
the rate of corrosion of boiler tubes. In addition the
boiler water has a tendency to foam which results in
carry-over of salts in the steam".
Mr. Bruijn records a fact that should be appreciated
widely, namely that pure sucrose is not readily detected
by conductivity readings. The impurities that
accompany sugar in early stages of manufacture
increase conductivity and thus contamination in the
early stages can be detected by this means.
The Sugar Milling Research Institute has reported
on the use of conductivity apparatus in detecting
contamination, from sugar. The general question of
conservation of condensate has received much
attention.
Modern, practice in a mill appears to be keeping
various condensates separate and testing each flow.
If the quality is satisfactory it is sent direct to the
feedwater tanks. If it contains sugar it is not wasted
but diverted to where it can best be used.
Every endeavour should continue to be made to
develop practical automatic equipment to take care
of this matter.
In older mills using steam operated reciprocating
drives the condensate from these machines becomes
contaminated with lubricating oil. The steam and
condensate should receive special attention and never
be released into the feedwater unless the oil content is
consistently only in small trace quantities.
Leakage of cooling water into turbine condensers
can be detected in several ways of which conductivity
is probably the best. This can cause a serious increase
in costs of internal chemicals if not checked in time.
In. the majority of mills it is necessary to use makeup
of river water to add to the condensate available,
in order to provide the quantity of feedwater required
by the power station.
You may well ask if it is water impurities which
cause boiler problems, why not remove all the impurities
before the water is used? Completely purifying
water is costly and providing purer water than is
necessary is economically unsound.
Preparing water for boilers requires considerable
study and planning. It is not just a case of finding
a magic "muti" or magic gadget and using either of
these and relaxing.
Careful analyses of raw water supplies must be
conducted and these should be done on a number of
samples over a reasonable period of years to gain some
idea of seasonal variations. The older mills have this
information, and also much experience. New mills
have not as a rule got this information. It is frankly
amazing how frequently water conditioning specia-
lists are asked to advise on schemes of which only one
"grab" sample has been submitted for analysis.
Quite frequently an analysis is set before one that gives
detailed information as to the suitability of the water
for general domestic purposes. In many cases essential
details required for calculations for water conditioning
are not available.
As the title of this paper indicates, the feedwater
and its composition are decisive factors in deciding
the conditioning required. Feedwater is the material
that provides steam and its composition must be such
that its impurities can be concentrated inside the
boiler without exceeding the tolerance limits of the
particular boiler design.
The water supply to sugar mills is normally clarified
(coagulated), filtered and chlorinated. These
processes which are part of what is termed external
treatment, render the water suitable for domestic
purposes. The result is a water, crystal clear, but
containing dissolved salts and the gases, oxygen and
carbon dioxide.
It is the presence of these substances that give rise
to the major problems of water conditioning for
boilers. Admixture with good quality condensate
merely dilutes these impurities. If the condensate
carries impurities such as oil or sugar these have
to be taken into consideration also. The first step to
be taken then is to calculate the composition of the
feedwater without any special conditioning having
modified its composition. This is the starting point
for developing a conditioning scheme.
In order to be able to appreciate the necessity for
the various steps in a conditioning programme let us
examine what might occur if we did nothing but mix
average domestic water with condensate and proceeded
to generate steam.
Table I gives the solubility of various chemical
compounds that are common in waters:
Table I
Solubility of Chemical Compounds
ppm. as CaC03
Calcium 32° F 212° F
Bicarbonate 1,620 decomposes
Carbonate 15 13
Sulphate 1,290 1,250
Magnesium
Bicarbonate 37,100 decomposes
Carbonate 101 75
Sulphate 170,000 356,000
Sodium
Bicarbonate 38,700 decomposes
Carbonate 61,400 290,000
Chloride 225,000 243,000
Hydroxide 370,000 970,000
Sulphate 33,600 210,000

82 Proceedings of The South African Sugar Technologists' Association—March 1966
Scale Formation in the Boiler
Calcium and Magnesium compounds in the water
are precipitated by the heat and pressure and form
scale and sludge.
Magnesium sulphate is very soluble, as will be seen,
but it usually reacts in boiler water to form less soluble
magnesium salts such as the hydroxide. Calcium
sulphate tends to remain in solution in the colder
areas of the boiler but becomes a deposit, crystal by
crystal, inside steam bubbles that are forming in the
hotter parts of the boiler, where scale can cause the
most damage. The presence of dissolved silica, iron
and alumina complicate matters and these deposit
with the calcium and magnesium scales.
The scales derived from the carbonates or bicarbonates
of calcium and magnesium usually form suspended
solids in the boiler water, and if permitted to
concentrate too much, settle out and become "baked
on" to heat-transfer surfaces to form scale. Oil if
present in condensate, adsorbs on to sludge (or suspended
matter), and if this becomes baked on forms
a very dense insulating scale.
The danger of scale formation in tubes deserves a
short discussion in order to examine the principles
involved.
"As water circulates through boiler tubes it absorbs
heat and cools the metal. Scale forms a barrier
between the circulating water and the tube, decreasing
the efficiency of heat transfer. As the result, the
metal of a scaled tube has to be hotter to transfer the
same amount of heat as a clean tube. When boilertube
steel is heated to about 900° F it starts to weaken."
When it is appreciated that furnace temperatures are
normally in excess of 2000° F it can be realised how
important it is to maintain relatively scale-free
conditions at points of h'gh heat transfer.
Foaming in the Boiler
It will be noted that sodium salts are very soluble
and this solubility increases as water is heated. Sodium
salts do not normally cause boiler deposits but
contribute to causing foaming which will be discussed
later. Soluble organic matter such as sugar often
assists in giving the foam "body" or strength. Some
forms of suspended material have the same effect.
The result of course is entrainment of boiler water in
the steam which is very undesirable.
Corrosion of Steel in the Boiler
There are two main agents concerned with corrosion
of boiler steel: oxygen and salinity.
Oxygen: At atmospheric pressure and ambient
temperature about 9 parts per million of oxygen
can be dissolved in water. Water, as delivered to the
mill after only external conditioning for domestic
purposes, is saturated. As water is heated the oxygen
becomes less soluble and some escapes. As feedwater
enters a boiler the oxygen remaining either
begins corroding (oxidising steel) or goes off into
the steam lines.
Salinity: Sodium chloride particularly, and
sodium sulphate to a lesser extent, assist oxygen to
penetrate through deposits to steel surfaces, and
the result is corrosion unless inhibitors are present.
Damage to Superheaters
Superheaters consist of tubes carrying steam that
are exposed to hot products of combustion with the
result that the steam is heated well above the temperature
for saturated steam at the pressure of generation.
Maintaining the steel of the tube within safe temperatures
depends on. the cooling effect of the flowing
steam which is being heated. Foaming or "carry over"
of boiler saline in water causes re-boiling in the superheaters
and formation of deposits. These deposits
interfere with heat transfer and also if excessive, cause
restrictions and diminution of flow of steam. This
can cause either softening followed by bursting of the
superheater tube (or element as it is commonly called)
or corrosion by direct action of hot steam on hot
steel forming oxide of iron and hydrogen gas.
Damage to Turbines
Sodium salts carried over with the steam can form
incrustations on turbines. These if detected in time
can be removed by washing with water.
In boilers operating above 400 p.s.i.g. selective
silica carryover has to be considered. Scientific study
by G. C. Kennedy and other 3 4 7 has shown that:
(a) Steam is a solvent for silica.
(b) The ratio of silica in the steam to silica in the
boiler water increases rapidly as boiler pressures
increase.
(c) At 400 p.s.i.g. silica in the boiler water should
not be permitted to rise above 100 p.p.m.
(as SiO2) in order to keep silica in the steam at
or below 0.02 p.p.m.
(d) With silica in steam at or below 0.02 p.p.m.
(as Si02) appreciable turbine deposits would
not normally occur.
Corrosion of Non-Ferrous materials in the boiler
If concentration of alkaline water is carried too far
and very high pH values result, bronze fittings can
be attacked and fail.
Corrosion (Erosion) of gauge glasses
Slight leaks even on boilers operating as low as
160° F causing concentration of alkali at the point
of leakage can cause solution of glass and failure.
In high pressure boilers gauge glasses are protected
by mica sheet from direct alkaline attack.
External Conditioning
By this term we mean conditioning that is completed
external to the boiler itself such as:
(a) Clarification, (Coagulation), Filtration.
(b) Precipitation Softening.
(c) Ion Exchange techniques.
(d) Evaporation.
(e) De-aeration.

Proceedings of The South African Sugar Technologists' Association March 1966 83
Clarification (Coagulation), Filtration
These are normally arranged for the main water
supply to the mill and for domestic purposes.
Only a small proportion of this supply is used as
make-up to the boilers.
It is important nevertheless, that the fact that it is
used for boiler feed should be realised. Sometimes
large quantities of aluminium sulphate are used for
coagulation. To render the water alkaline with this
large quantity a correspondingly large quantity of
lime is required. The result is a considerable increase
in calcium sulphate content and this has to be removed
by special external conditioning or dealt with
in the boiler itself. A thing like this literally causes a
chain reaction and can cause costs of conditioning to
soar. Sometimes, particularly with water coloured with
organic material it is necessary to coagulate with
aluminium sulphate at pH values nearer 5.0.
This type of water needs the addition of alkaliusually
lime, after filtration, to minimise corrosion of
steel piping, tanks, etc.
For boiler water conditioning the use of caustic
soda as the alkali would be preferable, but this is
normally impossible to justify for the whole supply
from the cost aspect. This also applies often to the
use of soda ash which is the second choice.
Normal chemical reactions in this type of conditioning,
with the primary coagulant alum, are:
The use of very small dosages (up to 4 p.p.m.) of an
organic coagulant-aid can, in many cases, enable
satisfactory coagulation to be achieved with a minimum
amount of primary coagulant. This has considerable
merit from all points of view including that of
conditioning of the boiler make-up water. Organic
coagulant aids add virtually no dissolved solids to the
water.
Precipitation Softening
The principle of operation may be summarised as
follows:
A predetermined quantity of selected chemicals is
added to a quantity of water and rapid mixing ensures
the dispersion of the chemicals. The treated water is
subjected to a slow mixing to enable chemical reactions
to proceed, crystal growth to occur, and separation
of the precipitates to commence. This period,
known as the flocculation period, is followed by a
period of sedimentation either in a quiescent state or
under continuous flow conditions, but sufficiently
slow to permit of the separation of the precipitates
to the bottom of the tank or basin from which point
they can. be removed as a sludge.
The chemicals used are lime, soda ash and. sodium
aluminale. These are normally mixed into a thin
slurry with water and fed by proportionating dosage
equipment. Such equipment must be regularly cleaned
of all deposits liable to affect accurate dosing.
This type of softening has not found favour with
sugar mills because of control difficulties associated
with seasonal quality variation of lowland surface
supplies. Without adequate chemical control this
type of softener can cause considerable difficulty in
the power station.
Ion Exchange techniques
We now meet the terms ion exchange materials
and cations and anions. The origin and exact meaning
are matters for physical chemists, but let us regard
them as terms for use in our processes.
Let us take common salt, sodium chloride: This is
written as NaCl. In aqueous solution, it splits up
into the Sodium cation Na + and the chloride anion
CI . Similarly calcium sulphate, CaS04, splits up into
cation Ca ++ and anion S04 -- . The + and -- signs
indicate the nature of the electrical charges carried
by the ions.
The common feature with all ion exchange processes
is the use of ion exchange resins (complex
synthetic organic materials) in granular or bead form.
These materials, which are insoluble, are rather like
storage batteries. When exhausted, they can be
regenerated and are thus rendered ready for a further
cycle of operation. The normal usage of these resins
is to allow the water to be treated to flow through a
bed of the material not less than 30 inches deep. The
ion-exchange reaction is, under these conditions, very
rapid.
Cation Exchange—Sodium cycle
Water is passed through a bed of ion exchange
material. During its passage through the bed, hardness
cations, calcium and magnesium, are taken up
by the exchange material and are replaced in the
water by sodium. In this reaction the ion exchange
material eventually becomes saturated with calcium
and magnesium and further exchange action ceases.
To restore this capacity a solution of common salt
(sodium chloride) is passed slowly through the exchanger.
The effect of an excess of this is to drive out
the calcium and magnesium and to replenish the
sodium in the material, which is then available for a
further cycle of softening.
The reactions in this process merely change the
bases. Calcium and magnesium ions go into the ionexchange
material and are replaced in the water by
sodium ions. This process does not alter the quantity
of dissolved solids in water appreciably, nor does it
alter the anions (bicarbonate, sulphate, chloride).
This process does not remove silica from solution.
Most of the modern base exchange plants offered
utilise polystyrene synthetic resins in bead form.
These resins are available from a number of manufacturers.
They have a high exchange capacity and
high resistance to deterioration under service conditions.
The exchange capacity of an ion exchange
material is usually expressed in terms of the grains of
hardness (expressed as CaCO3) removed by a cubic

84
foot of the material. Water, having a total hardness of
100 p.p.m. (as CaCO3) has a hardness of 7 grains per
Imperial gallon. A gramme (metric) is equivalent to
15.43 grains.
The exchange capacity varies with the salt usage
and a general average for a polystyrene resin would be
the following:
Salt lb/ft 3 Exchange capacity, grains/ft 3
as CaCO3
6 20 000
10 25 000
15 30 000
A 30 inch depth of bed is recommended as a minimum.
The operation of a base exchange softener
consists of:
(a) Backwashing to clean the resin and loosen it up.
(b) Brining to regenerate the resin (i.e. the use of
a salt solution).
(c) Rinsing to wash all excess salt away.
(d) Softening.
Reactions concerned in this type of softening can be
expressed in a simplified form as follows:
Softening cycle
Ca(HC03)2 + Na2 ® -> 2NaHC03 + Ca ®
Calcium carbonate + Base exchange material in
sodium form—> Sodium bicarbonate + Base exchange
material in calcium form.
MgS04 + Na2 ® ---> Na2S04 + Mg ®
Magnesium sulphate + Base exchange material in
sodium form ---> Sodium bicarbonate + Base exchange
material in magnesium form.
Regeneration using common salt (sodium chloride)
2NaCl (in excess) — Ca ® —> Na2 ®
+ CaCl2 + excess NaCl
Sodium chloride + Base exchange in calcium form—>
Base exchange in sodium form+Calcium chloride.
In the above ® is the Resin material
The plant required consists essentially of a cylinder
containing the active ion exchange material supported
on gravel beds, beneath which is a specially designed
collecting system, and a salt saturator, suitable piping
arrangement and valves. Salt solution is injected into
the system at a 10 per cent strength by use of a hydraulic
ejector diluting a 30 per cent brine solution
(saturated), three times. This brine solution requires
to be removed thoroughly by rinsing and then the
softener is ready for operation on the softening cycle.
When the softening action is exhausted, backwashing
is done to cleanse and loosen the ion exchange
resin bed and regeneration using salt is done as
described above.
The above process can be arranged for manual
operation using standard valves.
By eliminating the standard valves and using in
their place a single control valve whereby all operations
are carried out either by the turning of one handle or
by the moving of a lever to fixed settings, the sequence
Proceedings of The South African Sugar Technologists' Association March
of operations is simplified. Backwash and rinse flows
should be controlled by butterfly valves operating
against orifice plates to enable correct rates of flow
from each operation to be attained.
With softeners using a multiport valve with a rotating
motion it is a relatively straightforward matter
to render the complete operation fully automatic.
The multiport valve is motorised and is connected
to a make and break contactor head which rotates
the valve to the correct positions for softening, backwashing,
brining and rinsing. The period between
regenerations for all types should be altered to suit
seasonal variations in the hardness of the raw watersupply.
This is a vital control procedure.
Cation exchange—Weakly Acidic
This unit is similar to the base exchange unit
described above, except that the cylinder and piping
are rubber lined, the ion exchange material is a special
weakly acidic cation exchange material and. the regenerant
is dilute sulphuric acid.
Weakly acidic cation exchange material only reacts
with cations combined with the bicarbonate anion.
The cations are absorbed and replaced by the hydrogen
ion. The result is the formation of carbonic acid
(H2CO3). Sulphates and all other salts pass through
the plant unchanged. Carbon dioxide is removed, from
the water by a scrubbing tower in which the carbonic
acid is broken up by an up-flow air stream. The water
which has been treated in this way is then passed
through a standard base exchange softener to remove
the non-carbonate hardness. The water is now of
low alkalinity, of zero hardness and with sodium
chloride and sodium sulphate. The pH of the water
after degassing is 6.5. Caustic Soda solution is added
to raise the pH to 8.2 to prevent corrosion of the base
exchange unit, the pipelines and the pre-boiler system.
This type of treatment plant enables control of
alkalinity to be achieved, and if carefully supervised,
zero hardness can be assured.
It has to be appreciated that this process does not
remove silica.
As is well known fully demineralised water is obtainable
with special ion-exchange plant. With boilers
operating under 500 p.s.i.g. the expense of obtaining
this is seldom justified.
At sugar mills as yet, as far as known, none have
been installed.
A feature of the ion exchange processes described,
is the fact that the operation is not complicated,
control is relatively simple and the reactions arc automatic.
Deaerating equipment
In low pressure equipment with cast iron economisers
it is not vital to provide deaerating plant for
the feedwater, to protect the boiler from corrosion.
In the case of boilers operating at 400 p.s.i.g. and
above, with steel tube economises, it is necessary to
deaerate efficiently.

Proceedings of The South African Sugar Technologists' Association—March 1966 85
The method most commonly used is to provide a
pressure type deaerator heater. In this type of equipment
the feedwater is broken up into a fine spray
and scrubbed with low pressure steam. Part of the
steam is vented carrying with it the bulk of the dissolved
oxygen from the water. This type of deaerator
heater reduces the dissolved oxygen content of the
feedwater to 0.007 parts per million. Final oxygen
scavenging is done chemically as described later.
Hot well arrangements
The hot well or surge tanks should be protected
from ingress of air, if possible utilising steam blanketing.
Hot wells should be painted with special protective
paints to prevent corrosion and they require
regular maintenance. Feedwater in hot wells should
be maintained at a pH value in excess of 8.2.
Internal Conditioning
Since even minute amounts of impurities can cause
trouble, and these can escape from external conditioning,
a follow-up treatment is needed, regardless of
how the make-up water is prepared. In a sugar mill
the make-up water is only a small percentage of the
feedwater, and hence the feedwater contains only
minute quantities of impurities per 1,000 gallons.
Nevertheless these impurities cannot be ignored, and
it is general practice to use a form of internal treatment
as a final conditioning.
A complete final conditioning for sugar mill boilers
includes:
(a) Softening chemicals to react with feedwater
hardness.
(b) Sludge conditioners to disperse sludge and keep
it from sticking to metal surfaces.
(c) Oxygen scavengers and alkali to prevent corrosion.
(d) Chemical antifoams to prevent carryover.
Softening Chemicals
Softening chemicals used are sodium phosphates
and caustic soda. The type of phosphate selected
depends on boiler conditions and the method of
feeding used.
Sodium phosphates used are:
Disodium phosphate Na2HP04, Trisodium phosphate
Na3P04 and sodium polyphosphates such as
hexameta-phosphate Na6P6018 and septaphosphate
Na9P7022 and similar complex phosphates.
Where a number of boilers are fed from one ring
main and there is no provision for phosphate addition
to individual boilers, formulae containing selected
organic matter and scdium polyphosphates are used.
These do not react and form precipitates in the mains,
and economisers. It is necessary with these to ensure
that there is alkali (caustic soda) in excess. This
ensures that when the polyphosphates reach the boiler
they form trisodium ortho phosphate. This trisodium
phosphate reacts with the small amounts of calcium
hardness and forms hydroxy apatite Ca]0 (P04)6
(OH)2. In this form it is non-adherent, and is removed
via the blowdown cock.
Under these conditions also magnesium hardness
is precipitated as either magnesium silicate or magnesium
hydroxide.
All of these are non-adherent particularly if dispersing
type organic materials are used in the treatment.
If insufficient caustic soda (or sodium carbonate)
is used to satisfy the requirements to form trisodium
ortho phosphate then adherent calcium compounds
can be formed, and calcium may combine with silica
to form very hard scale.
It is, for the above reasons, vital to ensure that
sufficient residual trisodium phosphate and caustic
soda are carried in the boiler water. Caustic soda
excess assists in maintaining silica in solution as
sodium silicate.
It must be appreciated that internal conditioning
for control of scale formation is done in the boiler
itself despite the fact that the chemicals are usually
added to the feedwater at the feed pump suction.
It is when the feedwater meets the boiler water under
full pressure, and consequent temperature, that
softening reactions proceed speedily to completion.
The softening chemicals brought in with the feedwater
merely replenish the reserve of softening chemicals
in circulation in the boiler itself.
Sludge conditioning chemicals
Specially processed temperature stable organics are
used. Under conditions in sugar mills where feedwater
is very soft due to a high percentage of condensate
return and the volume of sludge formed from
each 1,000 gallons of feedwater is relatively low, it is
considered that the sludge conditioning preferred is
more a dispersion that a coagulation. It is for this
reason that specially processed lignin derivatives
combined with starch type organics are being utilised.
Tannins are particularly effective in conditioning
calcium carbonate and magnesium hydroxide precipitates
where the total sludge content is high. The
action of tannins is more coagulation that dispersion.
The starch type organics are particularly effective in
waters having high silica content, preventing adherence
of silica scale. When moderate oil contamination is
present the use of starch type organics is beneficial.
The action here appears to be keeping the calcium
phosphate sludge bulky and in a condition suitable for
maximum adsorption of oil into the sludge.
In this form oil is removed with the blowdown, and
does not adhere to metal surfaces. Boilers with a fair
quantity of precipitated sludge are more tolerant of
oil than boilers which have little or no sludge.
Oxygen scavengers and alakali
Even with the best deaerating equipment some
oxygen gets through. Chemical oxygen scavengers
complete the work in the case of boilers in the 300
p.s.i.g. (and higher) class. Sodium sulphite is normally
used with these boilers, and this is preferably catalysed
with a special chemical to provide very speedy reactions.
Hydrazine is used normally in boilers operating
at higher pressures where the fact that it does not
contribute to boiler water dissolved solids is important.

86
Oxygen, scavenging chemicals are dosed to the feedwater
after it has passed through the deaerator, if one
has been installed.
As explained earlier in many of the older installations
operating at 200 p.s.i.g. or lower, deaeratio.n as
such is not practised. In these the use of the lignin
based organic, and alkali in sufficient quantity, protect
boiler metal by protective film formation, and retardation
of corrosion by relatively high pH values.
Chemical antifoams to prevent carry-over
The causes of foaming or "carry-over" are quite
complex but in general this distressing phenomenon
is related to the concentration of dissolved solids in
the boiler water. High alkalinities generally assist in
the formation of foam, and organic matter such as
sugar can give it strength. Most modern antifoams
utilise complex synthetic materials to combat this.
These specially developed chemicals have two basic
properties; they are insoluble in water, and are surface
active. When properly dispersed in boiler water
exceedingly small quantities affect the skin of the
steam bubble and weaken it. There are distinct advantages
in combining antifoams with sludge conditioners
which act as carriers of the antifoam to where
it is needed.
Internal Conditioning—Summary
Internal conditioning for sugar mills with a large
percentage of condensate in the feedwater requires:
Phosphates.
Alkali.
Dispersing organic material preferably lignins
and processed starches.
Antifoam chemicals.
Oxygen scavenging materials.
Chemical feeding
The chemicals described in this paper and most
offered today are in briquette or pulverised form, and
require to be dissolved in water and fed as a solution.
The major companies specialising in this field of
chemistry have spent a great deal of time and money
on research and scientific experimentation. This
accounts for the fact that often the exact composition
of the chemical formulations is not fully revealed to
the customer. Nevertheless, the general principles are
usually revealed, and analytical control methods used,
follow accepted chemical techniques.
It is best to feed the alkalis and sludge conditioning
chemicals at the feed pump suction in the hot well.
If a deaerator is used sodium sulphite should be fed
into the feedline just after it.
As explained earlier, it is preferable to feed the
sodium phosphate direct to the boiler drum, but where
this is not possible sodium polyphosphate is fed with
the sludge conditioning chemicals at the hot well.
The use of positive dosing equipment operated by
electric power or coupled to the reciprocating steam
supply at the feed pumps merits serious consideration.
Proceedings of The South African Sugar Technologists' Association—March 1966
Table II
Boiler Water Analytical Control Limits (p.p.m.)
Applicable to Water Tube Boilers in the Sugar Industry
Work ing Pressure
Total Dissolved Solids .
Suspended Solids
Hydroxide Alkalinity
(as CaC03)
Soluble Silica (as SiO2)
Phosphate (as PO4) .
Sulphite (as Na.,SO3) .
Hardness (as CaCO3) .
Oil
PHOSPHATE RESIDUAL CONDITIONING
Up to 200
p.s.i.g.
3,000 max.
800 max.
mm. 100
preferable
10% of
Dissolved
solids
max. 400
150 max.
40 to 80
zero
7 max.
200 to 350
p.s.i.g.
2,500 max.
300 max.
min. 100
preferable
10% of
Dissolved
solids
max. 350
100 max.
30 to 60
30 to 100
zero
5 max.
350 to 450
p.s.i.g.
2,000 max.
150 max.
min. 100
preferable
10% of
Dissolved
solids
max. 300
90 max.
20 to 40
20 to 70
zero
zero
It will be seen that residuals of alkalinity, phosphate
and sulphite are set out. Individual water treatment
companies often add to these limits by requiring
additional analytical data to suit their particular
formulae.
Control of the dissolved solids figure is by blowing
down. As the blowdown. water is replaced with lower
solids feedwater the boiler water is essentially being
diluted. Other things being equal more economical
usage of chemicals is obtained by endeavouring to
operate near the limit of dissolved solids set down.
The pH of the boiler water should always be above
10.0. If the above limits are used the pH always will
be above this figure. Sugar can depress this and remedial
action by addition of alkali should immediately
follow detection.
It will be noted that there is a considerable difference
between the analytical control limits applicable
to water in boilers operating up to 200 p.s.i.g. and
those applicable to water in boilers operating at
400 p.s.i.g. and above.
Table II gives the control limits for "phosphate
residual" internal conditioning, which system has been
in use for many years. Relatively recent refinements
in this technique have been the development of special
temperature stable organic materials as sludge conditioners,
and antifoams, capable of use in boilers
operating up to 2000 p.s.i.g.
A New Technique—Chelating
A relatively new technique which differs considerably
from the above is now being tried out in South
Africa as internal conditioning, after intensive research
and practical tests overseas. It is called "Chelating".
Specialised synthetic organic materials are used.
These have the ability to combine with, or "chelate"
calcium and magnesium ions, which are present in
boiler waters as carbonates, sulphates, etc. A soluble
complex compound is formed. The chelated calcium
and magnesium cannot be precipitated by carbonates,
phosphate or hydroxide or heat and remain in solution.

Proceedings of The South African Sugar Technologists' Association—March 1966 87
At this stage of our experience with this new technique,
considerable care must be taken to ensure that
the conditions under which it is used are correct.
If boilers are being kept satisfactorily clean with a
conventional organic phosphate technique, the use of
"chelating" agents will be of little or no additional
benefit. This technique requires much more accurate
control than the organic phosphate technique.
There are potential corrosion, hazards that must be
taken, into consideration. A. very high degree of
deaeration of feedwater is essential at all times.
"Chelating" has a very marked cleaning effect on
calcium and magnesium scale. It is considered that with
old boilers of riveted construction it would be exceedingly
unwise to try this technique. Leakage could
become a serious problem and cleaning out seams
could create conditions favourable to caustic embrittlement
of steel.
With new boilers of all-welded construction, up-todate
pretreatment of make-up, excellent quality
condensate, and precision equipment for feeding
chemicals coupled with close analytical control, this
new technique is worthy of examination.
The only difference to be noted in the control limits
set out in Table II if a "chelating" agent is used, are:
Phosphate and hardness are no longer determined.
Residual chelating agent is required to be present
at all times, and special practical analytical techniques
have been developed to enable this residual
to be controlled. This residual should not normally
exceed 100 p.p.m. of chelating agent in the boiler,
otherwise corrosion may be encountered.
Dissolved solids, alkalinity and sulphite limits
remain unchanged.
We will be hearing a lot more about this technique
in the future, as experience is gained in its use.
Caustic Cracking
(Embrittlenient of Boiler Steel)
It would not be right to ignore this matter in a
paper dealing with feedwater conditioning. It does
not occur very often, but when it does it is a serious
matter. The cracking which develops is continuous,
predominantly intergranular, and originates at the
surface of the metal.
To get embrittlement of boiler steel, there arc four
conditions which must be present at the same time.
1. The metal must be under stress near its yield
point.
2. The boiler water must contain caustic soda.
(The presence of some silica along with caustic
soda accelerates attack).
3. Inhibiting chemicals must be either absent or in
too low a quantity to stifle attack.
4. There must be some mechanism (a crevice, seam,
leak, etc.) permitting the boiler water solids to
concentrate on the stressed metal.
The advent of the all welded boiler appears to have
caused a slackening of research on this matter. The
modern method of ensuring that a water is nonembrittling
relies on. either tannins or lignins by themselves
or sodium nitrate either by itself or with the
tannins or lignins. Caustic alkalinities should be kept
within the limits laid down and preferably should not
exceed 250 p.p.m. If caustic alkalinities are higher
than this, consideration should be given to definite
control measures, particularly where riveted boilers
are in use. It is worthy of note that the water in steam
accumulators of riveted construction should conform
to anti-embrittling standards. There is no simple
test to determine if a boiler water will contribute to
embrittlement.
Oil Contamination of Boiler Feedwater
Detecting and eliminating oil contamination is an
important part of a feedwater treatment programme
in a sugar mill using steam driven reciprocating drives.
Many of these are still in use in the older mills.
Lubricating oil is injected into the steam used in
these driving engines, and when the steam has done its
work and been condensed, it contains oil.
This, if it gains access to the boilers in more than
trace quantities, causes trouble in a number of ways
such as increasing insulating power of scale, causing
sludge to become sticky and "baked on", and aggravating
foaming.
Oil contamination becomes more critical as boiler
pressures and heat transfer rates increase.
Removing oil contamination from steam is done by
using baffle separators or by centrifugal separators.
Removing oil contamination from condensate is
done by filters often using chemical coagulation prior
to filtration.
In the case of boilers operating on condensate
slightly contaminated with oil, it would be a safe
precaution to boil out during the off season with a
special chemical formula combining detergent ability
and oil emulsifying action, to ensure oil free metal
surfaces.
Conclusion
Adequate water conditioning arrangements and
control are vital to the successful generation and usage
of steam in the sugar mill. It is now a highly developed
and specialised branch of chemistry and engineering.
The exact mechanism and chemistry of many of the
reactions described in this paper are beyond the
author's comprehension. It is with the authority of the
research chemists, and with records of successful
practical experience that they are put forward. Sugar
technologists can rest assured that what has been
stated can be substantiated, and the conditioning
methods described are practical and scientific. The
time for guess work and magic in water quality work is
past.
Mr. Oliver Lyle in the introduction to his excellent
book "The Efficient Use of Steam" states:

88 Proceedings of The South African Sugar Technologists' Association—March 1966
"Some of the explanations may not be scientifically
correct, but the author believes that it is
more important to be nearly right and understandable,
than to be academically accurate and
incomprehensible. (The author is insufficiently
equipped to be academically accurate)".
The author of this paper is in full agreement with
this, including the portion in parentheses.
Acknowledgments
Permission to present this paper has been granted
by The Alexander Martin Corporation of Johannesburg
and Durban, distributors for the Nalco Chemical
Company of Chicago, from whose technical publications
and bulletins much has been taken, in many
cases in the original wording used. The author has
extracted liberally from the paper he presented to the
Institution of Certified Engineers South Africa (Natal
Branch) at Durban in 1958-9. Extracts have also been
taken from a paper, prepared by Mr. W. A. Low,
Mr. W. L. P. Davies and the author, presented to the
Institution of Certificated Mechanical and Electrical
Engineers, South Africa, in 1965 in Johannesburg.
Detailed references to these sources of information
would only complicate matters in this paper.
Summary
This paper surveys modern water conditioning
practice in respect of feedwater for boilers in sugar
mills in Southern Africa. The water conditioning
techniques described are applicable to water tube
boilers operating with condensate return being a high
percentage of total evaporation. Emphasis in this
paper is on basic principles and practical application.
References
1 Angus, G. E. (1959): Water Conditioning with special reference
to generation of steam, heat exchangers and cooling
systems. J. Instn. Cert. Engrs. S.A. 32, 133.
2 Bruijn, J. (1962): Automatic Sugar Detection in Boiler Feedwater.
Proc. S.A. Sugar Tech. Assoc.
3 Jacklin, C. and Brower, S. R.: Correlation of Silica carry-over
and solubility studies. A.S.M.E. Paper No. 51-A-91.
4 Kennedy, G. C. (1950): A Portion of the system Silica-Water,
Economic Geology 45, 629.
5 Low, W. A., Davies, W. L. P., Angus, G. E. (1965): The
selection and operation of water conditioning plant for steam
boiler installations. J. Instn. Cert. Mech. & Elect. Engrs.
S.A. 38, 253.
6 Lyle, O. (1947): The Efficient use of Steam. His Majesty's
Stationery Office, London.
7 Nalco Chemical Company, Power Industry Dept. (1961 to
1965): Technifax Bulletins.
8 Straub, F. G. and Grabowski, H. A. (1945): Silica Deposition
in Steam Turbines. Trans. A.S.M.E. 67, 309.
Mr. Cargill: What p.p.m. of sucrose is acceptable
in boiler feed water?
Mr. Angus: At a guess I think 100 to 150 p.p.m.
in the boiler itself on the basis of organic content
would be the upper limit.
Mr. Robinson: In Table II the figure is given of
3,000 p.p.m. for total dissolved solids. Is this for
treated or untreated water?
Mr. Angus: You can run on 10,000 p.p.m. in some
cases but the figure of 3,000 was set by the American
Boiler and Affiliated Industries Standards Committee
(1956) who laid down the figures for manufacturers
to consider in respect of dissolved and suspended
solids. It is certainly possible to operate at very much
higher figures, particularly with shell type boilers but
it is advisable to stick to the limits set out in the
paper.
Mr. Phipson: We used to rely on magnesia to remove
silica in the water by forming magnesium silicate.
Mr. Angus talks of removing magnesia but in that
case what happens to the silica?
Mr. Angus: Magnesia can be added to the water,
and magnesium silicate will settle in flocculant form
and be removable by blow-down. Some raw waters
have sufficient magnesia but softened waters have
none. In the absence of sufficient magnesia enough
caustic soda must be present to maintan the silica
in the form of sodium silicate and the amount required
at 200 p.s.i. would be 1.5 times the silica
content. Up to 400 p.s.i. it would be 2.5 times. Above
400 p.s.i., particularly where turbines are concerned,
other factors must be taken into account.
The maximum use of condensate is required, as the
more condensate you have, the less silica will go into
the boilers.

Proceedings of The South African Sugar Technologists' Association — March 1966 89
ECONOMIC DESIGN AND OPERATION OF PROCESS
HEAT EXCHANGE EQUIPMENT
Introduction
A thorough investigation of all equipment in a
sugar factory covered by this title would result in a
monumental publication. This relatively modest paper
purports to draw attention to a few specific aspects of
economy in juice heating. Considering the abundance
of comprehensive articles on liquid-vapour and liquidliquid
heat exchangers appearing in chemical engineering
journals it is surprising that the superficial
treatment of this subject as shown by articles in sugar
journals indicates that little application is made of this
valuable fund of knowledge in the sugar industry.
Although the empirical approach may be adequate for
routine specification and control, the application of
general chemical engineering techniques developed
By E. J. BUCHANAN
in this field would facilitate the attainment of a
maximum operating economy and the optimum
design of new equipment. It is hoped that this article
will produce a stimulus to the application of established
heat engineering techniques to the economic design
and operation of juice heaters and heat exchange
equipment in general in the sugar industry.
Derivation of Heat Transfer Coefficients
In the general case of heating juice flowing inside a
tube we are concerned with convectional heat transfer
to a liquid under turbulent flow. Consider for example
a juice velocity of only 3 ft. per sec. through a 1.5 in.
i.d. tube. If the density is 65 lb per cu ft and the
viscosity 0.5 cP, i.e. 0.5x6.72x 10 -4 =3.36x 10 -4

90
lb/(ft)(sec), (taking conservative figures), then the
value of the Reynolds number is:
NRe = Duρ/μ = (1.5x3x65)/(12x3.36xl0 -4
)
= 7.25 x10 4
which is well over the laminar flow region (below a
value of 2,100).
The overall heat transfer coefficient may be predicted
from a knowledge of the physical conditions existing
on either side of the tube wall but since it is
dependent on a number of variables it is usual to
estimate individual coefficients for the inner and outer
pipe surfaces and to summate these, as discussed later.
Liquid Film Coefficient Inside Tubes
The fluid adjacent to the pipe wall is in laminar
flow, hence heat transfer through this film is by
conduction and the liquid film resistance will be
dependent on the Reynolds number, as well as the
thermal conductivity and specific heat of the fluid, i.e.
h i = f(D,G,μ,k,c)
By dimensional analysis it may be shown that
h i D i k=f(DG/μ,cμk)
the three dimensionless groups being known respectively
as, the Nusselt number (N Nu ), the Reynolds
number (N Re ) and the Prandtl number (N Pr ), i.e.
N = f(N ,N )
Nu
Re Pr
A considerable amount of research has resulted in
the correlation
h D.k = 0.023(DG/μ) i 0.80 (cμ/k) 1/3
which holds for Reynolds numbers between 10,000
and 400,000 and Prandtl numbers between 0.7 and
120 18 . For liquids, this equation may be condensed to
h = 0.023G i 0.8 k 2/3 c 1/3 ,D 0.2 μ 0.47 ... (1)
Equation (1) may be solved approximately by the use
of the nomograph 3 in fig. 1.
Since μ decreases rapidly with an increase in
temperature, the film coefficient increases and it is
usual to calculate a mean coefficient for conditions
prevailing at the mean temperature of the liquid in
the exchanger. This is satisfactory for the case of low
viscosity liquids where a small temperature difference
prevails across the tube. However, in general it is
necessary to estimate the actual wall temperature in
contact with the heated fluid as discussed later.
Equation (1) shows that the liquid velocity is the
most important factor determining the film coefficient
inside the tubes. For example, if the liquid velocity
was increased from 3 to 6 ft per sec (all other variables
being constant) the film coefficient would increase,
according to equation (1), by a factor of 1 .74. Consequently,
the liquid velocity should be as high as
possible, the upper limit being economically dependent
on the incremental cost of the exchanger and the
pumping charges. 17
Outside Film coefficients
In the sugar industry we are concerned with the
condensation of the low pressure steam in the case of
Proceedings of The South African Sugar Technologists' Association — March 1966
juice heaters and also in the transfer of heat through
liquid films outside tubes in liquid-liquid heat exchangers.
In the latter case, equation (1) may be used
by substituting D e for D;
D e = 4% free area of cross section/perimeter
which applies when the flow is parallel to the tubes and
fully turbulent, i.e. NRe > 10.000 21
.
For steam-heated tubes, the installation may be
either horizontal or vertical. For film condensation,
Nusselt has developed the following equations 19
which apply for N Re in the film of less than 2,100. In
practice these equations are conservative by about 20
per cent due to the effect of ripple on the film.
Dropwise condensation would give higher values,
but in general it is safest to assume film-type condensation
for design purposes. When clean steam
condenses on clean surfaces film-type condensation
is always obtained. 15 The investigations of Osment
et al. 20 have shown that overall heat transfer coefficients
in surface condensers may be doubled by the
injection of filming amines into the steam space to
promote drop-type condensation.
It is interesting to note from equations (2) and (3)
that the relative effectiveness of steam condensation
rates for horizontal and vertical tubes is
Assuming that the tubes are 1.6 in outside diameter,
12 ft. long and 8 tubes are arranged in the average
vertical stack, equation (4) indicates that the horizontal
heater will have a 70 per cent greater condensing
film coefficient than the vertical heater.
The factor N in equation (2) accounts for the effect
of the accumulating condensate film around a vertical
stack of horizontal tubes, the film coefficient diminishing
for lower tubes. For this reason a staggered
arrangement of the tubes would promote a higher
film coefficient.' 1
Film and Wall Temperatures
In the case of turbulent liquid flow through tubes,
the difference in temperature between the bulk of the
liquid and the film in contact with the tube wall is
often neglected particularly if the temperature
difference across the wall is small. However, if
correction is necessary then equation (1) is multiplied
by
viscosity being the only variable which is significantly
effected by temperature.

Proceedings of The South African Sugar Technologists' Association — March 1966
Estimation of the wall temperature may be achieved
by a trial-and-error method using the equation 18
in which, h,- is estimated from equation (1) and h0 from
equation (2) or (3). For the preliminary estimation
of h„ the outer wall temperature is chosen midway
between the bulk temperatures on either side of the
wall.
The wall temperature is then obtained from
In the case of condensing a vapour outside a tube,
the condensate is normally under viscous flow and
the temperature drop across the film is more significant.
The mean film temperature is evaluated from 19
The wall temperature is assumed initially and the
value of the film coefficient, calculated by equation
(2) or (3), is checked using equation (6).
For approximate working figures equation (1) is
used without correction and the steam film coefficient
may be determined from nomographs such as fig.
2 25
Overall Heat Transfer Coefficients
The overall heat transfer coefficient is compounded
from the individual resistances due to the inside
scale, the inside film, tube wall, outside film and
outside scale as shown by equations (8) and (9). The
overall coefficient may be based arbitrarily on either
the inside or outside tube area but the chosen area
should be stated. The outside area is the most usual
choice.
In the above equations the diameter ratios correct
the values of the individual coefficients to the selected
area. In some cases one film coefficient may be
considerably greater than any of the others so that the
diameter conection has a small effect. In this case it
is convenient to abbreviate the equation eliminating
the diameters and to express the overall coefficient
in terms of the tube-side area in contact with the
highest resistance, i.e. lowest film coefficient. 18
The coefficients hdi and hdo represent the fouling
factors for the inner and outer tube surfaces, respectively.
Their combined values may be determined by
comparing the overall coefficients of the clean and
scaled heaters. If however the outer wall is clean, the
inside fouling factor may be calculated 14 fiom
Another method of determining the fouling factor
is by means of a Wilson plot, 1, 15 in which the
reciprocal of U is plotted as a function of u 0.8 for both
clean and fouled surfaces.
Estimation of Coefficients in Practice
There is little information available on heat transfer
coefficients of juice heaters under factory conditions
in South Africa. For this reason, even fundamental
questions such as the choice between vertical and
horizontal heaters or the optimum juice velocity are
often still a matter of controversy even after several
decades of experience. In spite of this lack of practical
information, many of the problems may be clarified
by applying the standard chemical engineering
techniques outlined in the previous section.
The author has determined Uo on several local
heaters and found rather low values of not more than
180 after being cleaned inside the tubes. One of these
heaters will be used as an example of the application
of the methods developed previously.
Example
The heater chosen for analysis is a horizontal
tubular type with tubes arranged in a series of vertical
91

92
stacks, 18 per stack on the average. The following data
apply:
Brix of juice = 14.7°
Juice rate =112 ton,hr
Heating range = 96° F to 192° F
Vapour satn. temp. = 218° F
Effective tube length = 11 ft 10? in
Inside tube diameter = 1.495 in
Outside tube diameter = 1.625 in
Total heating surface = 2,010 sq ft (based on o.d.)
Tube arrangement -- square pitch, average 16 per
stack
Tubes per pass --- 18
Heat Transfer Coefficient—Clean Tubes
Inner Film Coefficient:
It may be assumed that flow is turbulent (as calculated
earlier) hence the inner film coefficient may be
estimated from equation (1). The mean juice temperature
is
(96 + 192)/2 = 144° F or 62° C
Using the physical data in the appendix as an
approximation:
μ = 0.65 x 2.42 lb (ft)(hr)
k = 0.346 Btu (ft)(hr)(°F)
c = 0.92 Btu/(lb)(°F)
D i = 1.495/12 = 0.125 ft
Inside section = 3.1416 x (0.125)-4 - 0.0123
sq ft tube
G = 112 x 2,000/(18 x 0.0123)
= 1.012 10 6
From equation (1)
Tube wall transfer rate:
The tube wall coefficient may be determined as
inferred from equation (8). Assuming 70-30 brass
tubes:
k m = 60 Btu(hr)(sqft)(°F/ft)
x w
lb(sqft)(hr)
= (1.625-1.495) 12 = 0.108 sq ft
k m /x w . = 5,556 Btu (hr)(sq ft)( o F)
Steam side film coefficient:
Generally it is preferable to use nomographs based
on practical figures rather than the method discussed
previously, the limitations of which have been pointed
out. The nomograph by Stoever 25 fig. 2 may be
applied for an initial estimate.
Proceedings of The South African Sugar Technologists" Association —• March 1966
The condensing temperature is 218° F and
λ = 966 Btu/lb
q = 112 x 2,000 v 0.92(192-96)
= 1.978 x 10 7
Btu/hr
Go = 1.978 10 7
/(966 x 2,010)
=: 10.19 lb/(sq ft)(hr)
ND' o = 16 x 1.625 = 26
From fig. 2 (see dotted line example), the correction
factor is 0.34. Assuming a wall temperature of (218 +
144)/2 = 181 o
F the mean condensate film temperature
is, from equation (7)
t f = 218-3(218-181 )/4 = 190 o
and the corresponding base factor may be calculated
from
F b = 11.2t f + 1,320 (see fig. 2)
= 11.2 190 + 1,320
= 3,460
The film coefficient is calculated from
K = F b x F c (see fig. 2)
= 3,460 x 0.34
= 1,180 Btu/(sq ft)(hr)( o
F)
Check on Wall Temperature:
From equation (6)
which is sufficiently close to the value assumed above.
Corrected inside coefficient:
Using equation (5) h, may be corrected to the wall
temperature at which μ w = 0.45 cP and hence
h i = 860 (0.65/0.45) 0.14
= 906
This value is obtained approximately, following the
example (dotted line) in fig. 1.
Checking again with equation (6)
F

Proceedings of The South African Sugar Technologists'' Association — March 1966 93
Measurement of U„ in practice for this particular
heater under the given operating conditions provided
the value of Uod = 157.
Little information is available on fouling factors in
cane juice heaters. The Sugar Research Institute,
Mackay, 2 have conducted investigations on a pilot
scale heater which, upon analysis, provided results
in close agreement with U0 = 450 for a clean heater
under the present conditions. The thermal conductivity
of the scale was calculated as about 0.3 and
after 100 hours operation the thickness of scale was
about 0.006 inches.
Applying this information, it is possible to estimate
approximately the effects of fouling. Assuming for
example that the average scale thickness between
cleanings was 0.005 inches, then the fouling factor
would be
hdi = k/x (0.3/0.005)12 = 720
and from equation (10)
This assumes no outside fouling. The Sugar Research
Institute, Mackay, 2 observed on their pilot
heater that the overall heat transfer coefficient
decreased by as much as 30 per cent during a season
due to fouling outside the tubes. The pilot heater was
operated on factory exhaust steam. The heater examined
in the present paper had been operating for a
complete season, hence a similar degree of fouling
could be expected. In the absence of any confirmatory
data, if this is applied to the present case
This value for hdo is quite feasible even for a very
thin film. Oil, for example, has a thermal conductivity
as low as 0.07 so that a fouling factor of 671 could be
accounted for by an oil film of thickness
In this connection it should be pointed out that
the dropwise condensation promotion due to common
oils is relatively inefficient (c.f. filming amines) and of
short duration, particularly when other fouling compounds
are present. 20
The various heat transfer coefficients and fouling
factors for the heater in question are summarised in
Table 1.
Vertical vs. Horizontal Heaters
Equation (4) indicates that, all other conditions
being equal, the film coefficient for condensation in a
horizontal heater will be greater than for a vertical
heater provided that
Most tubes in cane juice heaters have Do =
1.625/ 12 ft and L = 12 ft so that equation (11) would
read

94
Hence the condensing film coefficient for a horizontal
heater is always greater than for a vertical heater. For
the heater discussed above for example, N = 16 and
from equation (4)
hoh/hov = 1 .486 or hov/hoh = 0.672
The condensing film, coefficient (Table I) for a
similar vertical heater would have been
hov = 1,180 X 0.672 = 793
and the overall coefficient would have been (Table I)
which would require an increase of 9 per cent in
heating surface. This is a very conservative example
since the exchanger was poorly designed (square pitch)
and heavily scaled. Had the tubes been staggered, 16
the number in a vertical row might have been reduced
to eight. Using fig. 2, N = 8 hence NDo = 13,
Fc = 0.42 and Fb = 3,460. Hence hoh = 0.42 x 3,
460 = 1453. The overall film coefficient would have
been
and 13 per cent more heating surface would be required
for a vertical heater. If, in addition, the
heating surfaces were clean then
Thus 28 per cent more heating surface would be
required for a clean vertical heater than for a clean
horizontal heater with N = 8. The working value
may range between 13 and 28 per cent, averaging
about 20 per cent.
The existence of this difference between horizontal
and vertical heaters cannot be disputed since it is
based on calculations which have been substantiated
by a large number of practical results from a wide
field of application. Considering that overall coefficients
are dependent on so many variables such as
steam and juice properties, juice velocities, degree of
inside and outside fouling, etc., it is not difficult to
imagine why some sugar factory designers are unwilling
to accept that this difference exists in practice.
Proceedings of The South African Sugar Technologists' Association — March 1966
A survey of costs per sq ft of heating surface for
juice heaters from local suppliers has indicated that
vertical heaters are normally about 5 per cent higher
than horizontal heaters. This means that the total
initial cost is 20 + 5 = 25 per cent higher for vertical
heaters.
The average price of heaters is R6 per sq ft and for
a 250 tch factory, with heating surface at 45 sq ft
per tch, 11 the additional initial cost for correctly
specified vertical heaters would be
To this must be added an additional 20 per cent on
Tunning costs.
This cost difference should be viewed in the light
of convenience of installation and operation for the
particular factory design. The choice of a vertical
heater on either of these grounds is not necessarily
based on economy and consequently falls beyond the
scope of this paper.
Economical Waste Heat Recovery
A typical example of the recovery of waste heat in
a sugar factory is the preheating of cane juice by
means of evaporator vapours and condensates. A
number of useful calculations has been presented by
FIGURE 3. Nomograph for the evaluation of equations (13)
and (14) in the economic optimization cf waste heat recovery
exchangers. 26 Copyright 1944 by the American Chemical Society
and reprinted by permission of the copyright owner.

Proceedings of The South African Sugar Technologists' Association - March 1966
Happel 7 for determining the economic optimum
heat recovery and these are discussed below.
Recovery of Heat from Vapour or Exhaust Steam
For the recovery of heat from steam or vapour at
a constant temperature, th to a liquid which is heated
without vaporization from temperature tc1 to tc2
(th - tc2) opt. = H/H, . . . (12)
where H = 114rE/UY
This calculation assumes a knowledge of the total
exchanger costs and marginal cost of steam production.
The optimum value of tc2 may then be determined.
This equation applies also to the case of waste heat
recovery from flue gases, as in a waste heat boiler,
where the heated liquid temperature remains constant
and the recovered heat is utilised in steam production.
Recovery in Countercurrent Liquid-Liquid Exchangers
For the recovery of heat from a hot liquid at
temperature th1 to a cooler liquid at temperature tc1
in a counlereurrenl (1-1) exchanger, the optimum
final temperature th2 of the hot liquid may be determined
from
. . . (13)
For any given problem the right hand side of the
equation will be a constant and R will be fixed.
Recovery in Multipass Exchangers
Pre-heating of juice by condensates is commonly
carried out in multipass exchangers of the 1-2 or 2-4
type as described by Webre. 27 For the case of a 1-2
type exchanger, the following equation applies in a
similar manner to the previous expression
. . (14)
For exchangers of the 2-4 type graphical differentiation
is most convenient for the evaluation of P.
Ten Broeck 26 has presented a convenient nomograph
for the evaluation of P for all three types of
liquid-liquid exchangers. This nomograph is reproduced
in fig. 3. The evaluation, of Ht, the incremental
cost of supplying heat, may present complications
it is composed, of several elements. First
there will be a saving resulting from decreased fuel
consumption. The value of the heat saved may be
determined from the price of fuel, its heating value
and the expected furnace efficiency. The cost of
supplying heal, by a furnace will include the fixed
charges on incremental cost of furnace as well as the
fuel cost. Also the recovery of waste heat may reduce
condensing and cooling costs.
Economy by Control
Since convection currents cause entrainment in
clarifiers it is essential that the temperature of entering
juice be stable. In the absence of proper control this
is often achieved by superheating and flashing to
constant temperature. The heat from flashed vapour
is rarely recovered in spite of the fact that (e.g.) a 250
tch factory by maintaining 10° F of superheat in the
juice would (if coal was being burnt) lose R7,200 per
year in heat.*
Although the maintenance of 10° F is only necessary
under conditions of very poor control there are cases
where, due to excessive fluctuation in juice velocities
and steam pressures even 10° F flash is insufficient
to maintain a safe margin for occasional peak flow
rates and resulting temperature drops below boiling.
In such extreme cases automatic temperature control
is not only a labour saving device but could be viewed
as an economic advantage.
It should be mentioned that the maintenance of a
small amount of flash is usually regarded as essential
for the release of air from the juice and the acceleration
of otherwise slow reactions but this discussion refers
to excessive flash.
Modes of Control
Conventional Control: The normal method of control
is to measure the outlet juice temperature and adjust
the steam control valve to maintain the desired
temperature. This usually requires a wide proportional
band setting to maintain stability and hence reset
response to correct the resulting offset due to load
changes. When rapid changes in throughput occur
the resulting short-term error can be corrected in part
by the addition of derivative response.
Condensate Throttling: By throttling the condensate,
a less responsive control action will be achieved but
this system has the advantage of reduced initial cost.
The behaviour of this type of system is difficult to
predict. 22 It also assumes an oversized heating surface
and is prone to the danger of excessive fouling on the
steam side of the tubes if condensates are contaminated
with oil, etc.
Pressure-Cascade Control: The most rapid recovery
to load disturbances may be attained by cascading
the output of a standard three-mode temperature
controller into the set point of a proportional plus
reset pressure controller. Changes in steam pressure
are corrected directly by the pressure controller. Load
changes are sensed rapidly by a change in shell
pressure which is compensated by the pressure
controller. The temperature control system senses the
residual error and resets the pressure controller set
point.
Minimum Temperature Control: In cases where more
elaborate control is excluded due to cost, sharp
downward peaks in the flashed juice temperature
recording chart may be eliminated by the injection of
* The above amount was calculated assuming 4,600 hr/yr
12000 Btu/lb coal, a boiler efficiency of 70%, 20% recycle of
filtrate on juice and 0.242 c /lb coal.
95

96
higher pressure steam through a small Sarco type
temperature regulator. As in the case of condensate
throttling this system has some obvious disadvantages
which may outweigh the low initial cost.
Whatever system is adopted the sizing of control
valves and the design of thermometer probes and
pockets should receive careful attention.
Recent Trends in Heating Economy
Recent efforts to increase heater economy have
been directed toward (a) more accurate optimization
by the application of computers to relieve the tedium
of design calculations (b) attempts to increase both
inner and outer film coefficients and (c) the complete
elimination of scaling.
Optimum Design by Computer
The design of a heater for optimum heat transfer,
pressure drop and cost, entails accounting for so many
variables simultaneously that the solution would
generally require the comparison of costs for a
considerable number of preliminary designs. For
example, the heat transfer coefficient is a function of
the liquid velocity which in turn influences the
pressure drop. By the application of computer
techniques both thermal and mechanical aspects may
be considered simultaneously and by initially applying
relatively empirical criteria, uneconomical designs
may be eliminated at an early stage. Another advantage
of computer methods is that it is possible to
design an exchanger considerably more accurately
than there is time to do by hand.
I.C.I. were recently faced with the design of a train
of exchangers for the recovery of waste heat from
gas to feedwater. A programme was developed capable
of designing and costing exchangers for the full
range of operating conditions. A typical design
print-out is reproduced in fig. 4. A complete design
providing all the data necessary for manufacture
takes between five and ten seconds of the machine
time. 12
Increasing Condensing Film Coefficients
Considerable research has been conducted into the
investigation of possible methods for the attainment
of dropwise condensation. Osment et al. 20 conducted
extensive tests on treated copper and brass tubes
using various types of steam. Field tests using industrial
steam showed the main cause of breakdown
to be corrosion and oxidation of the metal surface
rather than breakdown of the promoter film applied
to promote dropwise condensation. Thiosilanes and
xanthate compounds were most successful. After
cleaning the tubes of a condenser by injection of 50
per cent hydrochloric acid followed by 50 per cent
Teepol into the steam, 20 ml of a 1 per cent solution
of thiosilane: Si(SC12H25)4 was injected to promote
dropwise condensation. The test was continued with
weekly injections of promoter and good dropwise
condensation was achieved for one year. The amount
of promoter used was 0.01 ppm on steam. The overall
heat transfer coefficient ranged from 1,750 to 1,300
Btu/(hr)(sqft)(°F).
Proceedings of The South African Sugar Technologists' Association - March 1966
FIGURE 4. Computer print-out for a typical heat exchanger
design 12 .
Complete dropwise condensation for a period of
2,000 hours has been attained by applying submicron
films of paraxylene to chromium plated copper-ruckle
tubes. The use of submicron films of noble metals
also shows promise. 5
Increasing Liquid Film Coefficients
The stagnant liquid film inside heater tubes may
be eliminated by the use of rotary scrapers. This is
usually applied to the heating of viscous liquids in
double walled exchangers although fluids of 0.5
to 25,000 cP are recorded. The rotating scraper
maintains a thin highly agitated film in contact with,
the wall. Heat transfer coefficients are increased and
over-heating eliminated. Hence this type of heater is
suitable for thermally unstable liquids.

Proceedings of The South African Sugar Technologists' Association—March 1966 97
Direct contact liquid-liquid heat exchangers are
receiving increasing interest. The Bradford Institute of
Technology have surveyed recent developments. 6
The unit comprises basically a mixer-settler. If it was
applied to juice heating, a suitable heating medium
(e.g. a light oil) would be heated by steam and mixed
directly with the juice. After decantation, the oil
would be recirculated through the steam heater. The
method is particularly suited to the heating of corrosive
or highly fouling liquids.
Discussion
Considering that most of the heat consumed in a
sugar factory is absorbed by process demands, the
optimization of heat exchanger design and performance
is important. An essential basis for the attainment
of optimum conditions is a thorough knowledge
of heat transfer coefficients and fouling factors under
varying conditions. In spite of the fact that in other
industries the optimization of heat exchangers has
advanced to the stage of design selection by means of
computers, in the sugar industry there are few factories
where even records are kept of heat exchanger
performance. This paucity of information precludes
the accurate optimization of exchangers and often
confuses the choice of alternative design conditions
due to the unknown influence of operating variables.
In spite of the absence of practical data, the application
of chemical engineering techniques has been
shown to provide not only fairly reliable estimates of
heat transfer coefficients but also to indicate the influence
of the many variables upon which these
coefficients depend. As an example, the overall heat
transfer has been estimated from the physical properties
of the juice, steam and pipe wall using the Nusselt
method. Taking fouling factors from overseas tests,
calculated estimates have been made with a fair
degree of confidence.
It has been shown that the assumption of 30 per
cent loss in overall coefficient to account for fouling
outside the tubes would enable even closer agreement
between the calculated and measured overall coefficients.
This suggests that this loss of 30 per cent
due to fouling outside the tubes after one year's
operation (as measured overseas) could also occur in
local heaters. For this reason chemical cleaning outside
the tubes would probably allow for the installation
of smaller heaters.
The Nusselt equation indicates that the most
important single variable determining the inside
(juice) film coefficient is the juice velocity. For this
reason and to reduce scaling, the juice velocity should
be maintained at the recommended value of 5 to 6 ft
per sec 10 — unless sufficient information is available
to show by means of economic balance calculations
that the increased pumping costs prove the optimum
economic velocity to be lower.
The dependence of the overall coefficient on the
juice velocity for a typical fouled primary heater may
be calculated using equation (1) and the data in table
1 as
where u is the velocity of the liquid through the tubes
in ft/per sec. Thus, for velocities of 3 and 6 ft per sec
the respective overall coefficients would be 185 and
210 Btu/(hr)(sq ft)(°F) or an increase of 14 per cent.
On the other hand, if the tubes were perfectly clean
then the above equation would become
and for velocities of 3 and 6 ft per sec the corresponding
overall coefficients would be 375 and 500 Btu/(hr)
(sq ft)(°F) the increase being 33 per cent.
From the above calculation it is clear that the effect
of juice velocity on heat transfer is only really appreciable
when the heater is reasonably clean. An
important inference from this conclusion is that only
clean heaters have an appreciable amount of potential
self regulation. It was shown earlier that only a small
(13 per cent) difference exists between the coefficients
of fouled vertical and horizontal heaters, the difference
becoming significant (28 per cent) for clean heaters.
An important conclusion from the above is that
specification of the maximum heating surface required
to perform a given duty is quite simple provided
sufficient safety margin is allowed so that the heater
may operate when fully fouled and at low velocities.
Such variables as: vertical or horizontal, high or low
juice velocity, etc., may then be conveniently neglected
and the heater manufacturer may justly claim that
his heaters are equal in performance to any others
on the market. However, by taking this line of least
resistance it is quite possible that the resulting oversized
heaters are operated at a relative economic loss.
Furthermore, the tendency would be to allow the
accumulation of an abnormal degree of fouling
(particularly outside the tubes) before cleaning. This
in turn would result in reduced controlability of the
juice temperature. Since the majority of local heaters
have necessarily been installed without a substantial
basis of practical data on heat transfer coefficients
and a knowledge of the effect of juice velocity and
fouling, it may be assumed that they are generally
designed with a generous margin of safety. There is,
therefore, every reason to believe that the initiation
of a programme for the tabulation and correlation of
relevant data would facilitate the reduction of juice
heater costs and the elimination of such anomalies as
the fruitless operation of heavily fouled heaters at
excessive velocities.
Regarding the recovery of waste heat, the very
existence of the economic relationships expressed by
equations (12), (13) and (14) indicates that recovery
of heat is economical up to a point and thereafter the
cost of recovery increases beyond the marginal steam
cost. In South Africa, maintenance costs are relatively

98 Proceedings of The South African Sugar Technologists' Association—March 1966
low and consequently the economic recovery limit
may be higher than in other countries. For this reason
the optimum recovery point must be determined from
a knowledge of local conditions and not based on
empirical data from overseas. This again would
necessitate a more elaborate system for process data
retrieval. The resulting rationalisation of the design
and operation would logically lead to a significant reduction
in production costs.
Summary and Conclusions
It has been shown that the application of general
chemical engineering techniques to heat transfer
problems associated with the sugar industry provides
data which agree with practical experience. Detailed
calculations based on these techniques have shown
that many interesting conclusions may be drawn
regarding the most economical design and operating
conditions for heat exchangers.
It is suggested that in. the absence of both detailed
practical performance data and calculated estimates,
heat exchangers are necessarily oversized to allow a
margin of safety for the unknown effect of numerous
operating variables. To substantiate this remark it
has been calculated that the heater used as an example
in this report had a clean heat transfer coefficient of
450 Btu/(hr)(sq ft)(°F) but due to fouling (inside and
outside) that determined by measurement was only
157. In spite of this, the required juice temperature
was still attained.
Such heavy scaling has been shown to detract from
the self regulation of the heater. For example, an
increase of 3 to 6 ft per sec in juice velocity results in
33 and 14 per cent increase in overall coefficient for the
clean and fouled heater, respectively.
Calculations relating to horizontal and vertical
heaters have shown that 28 per cent additional
heating surface would be required on a vertical heater
when clean but only 13 per cent when fouled, since the
outside film coefficient becomes less significant as
fouling factors increase. Fouling masks the effect of
operating variables and hence it is easier to design
for fouled performance than for optimum performance.
Calculations have been presented for determining
the economic limit of waste heat recovery. A certain
degree of control is necessary for the maintenance of
optimum economic conditions. The economy of heat
transfer equipment is under constant investigation as
shown by the abundance of literature. Various
methods are being tested for increasing both inner
and outer film coefficients and eliminating scaling. The
design of heat exchangers for accurate optimum
conditions is now processed by programmed computers
in ten seconds. In order that the sugar industry
take full advantage of recent developments towards
increased heat transfer economy it is essential that a
system be established for the retrieval of operating
data to provide the basis for optimum design and
operation.
Nomenclature

Proceedings of The South African Sugar Technologists' Association—March 1966 99
Greek Letters
Δt Overall temperature drop, t h - t c , °F; Δt i
between wall and fluid outside tube
λ Latent heat of condensation, Btu/lb
μ Viscosity, lb/(ft)(hr); μ f , at mean film temperature;
μ w , at wall temperature
ρ Density, lb/cu ft; ρ f , at mean film temperature
ф Ratio (μ/μ w ) 0.14
Dimension/ess Groups
N Nu = Nusselt number, hD/k
N Pr = Prandtl number, cμ/k
to correct for wall temperature
NRe = Reynolds number, Du ρ
/μ = DG/μ
1
2
3
Bibliography
Anon., "Heat Transfer Coefficients and Scaling
in Juice Heaters", Research Inst. Mackay Tech.
Rep. No. 19, 1954, 4.
lbid.,p 10-11.
Brown, G. G., "Unit Operations", 1st ed., p 442,
Wiley, New York, 1950.
4 Chapman, A. J., "Heat Transfer", 1st ed., p 290-
291, Macmillan, New York, 1960.
5 Erf and Thelen, "Dropwise Condensation",
paper presented to First International Symposium
on Water Desalination, Oct.-1965, U.S.A.
6 Hanson, C. and Ingham, J., "Direct Contact
Liquid-Liquid Heat Exchange", Br. Chem. Eng.
10 (6) 1965, 391-399.
7 Happel, J., "Chemical Process Economics" 1st
ed., p 189-191, Wiley, New York, 1958.
8 Honig, P., "Principles of Sugar Technology",
Part 1, 1st ed., p 27, Elsevier, Neth., 1953.
9 Hugot, E., "Handbook of Cane Sugar Engineering",
1st (English) ed., p 313, Elsevier, Neth.,
1960.
10 Ibid., p 315.
11 Ibid., p 320.
12 Kirkley, D. W. and Momtchiloff, I. N., "Heat
Exchanger Design Programme for Optimization
Studies", Br. Chem. Eng., 10 (1) 1965, 31-33.
13 Landt, H. E., "Zur, Viscositat von reinen Zucketlosungen",
Zucker, 6 (23) 1953, 558-563.
14 McAdams, W. H., "Heat Transmission", 3rd ed.,
p 188, McGraw-Hill, New York, 1954.
15 Ibid., p 345-347.
16 Ibid., p 418.
17 Ibid., p 430.
18 McCabe, W. L., and Smith, J. C, "Unit Operations
of Chemical Engineering," 1st ed., p 438-
444, McGraw-Hill, New York, 1956.
19
Ibid., p 473-477.
20
21
22
23
24
25
25
27
Osment, B. D. J., Tudor, D., Spiers, R. M. M.,
and Rugman, W., "Promoters for the Dropwise
Condensation of Steam", Trans. Instn. Chem.
Engrs. 40(161) 1962, 152-160.
Perry, J. H., "Chemical Engineers Handbook",
4th ed., Chap 10, p 14, McGraw-Hill, New York,
1963.
Ibid., p 22-105 to 22-107.
Pidoux, G., "Formal Zur Berechnung der
Viscositat im Bereich," Zucker, 14 (20)1 961,
523-532.
Ryley, J. T., "Controlled Film Processing", Ind.
Chem., 38(448) 1962, 311-319.
Stoever, H. J., "Heat Transfer—Conduction,
Radiation and Convection", Chem. and Met.
Eng., 51 (5) 1944, 98-107.
Ten Broeck, H., "Economic Selection of Exchanger
Designs", Ind. Eng. Chem., 36 (1) 1964,
64-67.
Webre, A. L., "Duplex Unit Juice Heaters",
Sugar y Azucar, 56 (1) 1961, 35-37 and 58.
Appendix
Viscosity of Sucrose Solutions
The viscosity of pure sucrose solutions at various
temperatures and concentrations may be represented
according to Pidoux 23) by a series of linear plots on a
logarithmic ordinate and special temperature as
abscissa. Consequently, only two values of viscosity
for one concentration are required to determine all
values at various temperatures from the plot. Such a
plot is shown in fig. 5 for which the data of Pidoux 23)
and Landt 13) were used. The temperature coefficient
ф whose values correspond to the temperatures (t)
on the abscissa is calculated from
ф = tx 10 3 /(t + 273.16) 2
Thermal Conductivity of Sucrose Solutions
The thermal conductivity of sucrose solutions at
various temperatures and concentrations has been
tabulated by Honig. 8 ) However, due to the presence
of several obvious errors and the fact that extreme
accuracy is unnecessary for industrial scale calculations,
a linear multiple regression analysis was carried
out in order to express the thermal conductivity (k)
in terms of concentration (B) and temperature (t). The
small inaccuracy incurred due to the slight nonlinearity
of the t vs. k relationship is not significant
for industrial calculations. The value of k in Btu(ft)/
. (sq ft)(hr)(°F) may be calculated from t in °F and B
in weight per cent from:
k = 3.61 X 10 -4 t - 1.96 X 10 -3 B + 0.322
Specific Heat of Sucrose Solutions
The specific heat changes very little with temperature
variations. For industrial calculations in sugar
factories Hugot has proposed the following formula
for the calculation of specific heats of sugar liquors
(c) in Btu/(lb)(°F) from the brix (B):
c = 1—0.006 B

100 Proceedings of The South African Sugar Technologists' Association - March 1966
FIGURE 5. Graph for the evaluation of sucrose solution viscosities at various temperatures and concentrations 23 .

Proceedings of The South African Sugar Technologists" Association—March 1966
Mr. Hulett: What effect do stripping plates on the
tubes have on vertical juice heaters'?
Mr. Buchanan: I have no figures from practical
measurements, but obviously the condensate accummulation
as it runs down the tubes will be reduced
and hence the condensate film resistance will be reduced
by the use of stripping plates.
There is a calculation to show that for the outside
film coefficient of vertical and horizontal tubes to be
equal
hov = 793 X (12)1/4 = 1,453
whence L— 1.066
This indicates that for a vertical heater to be equal
in efficiency to a horizontal unit, stripping plates would
be required at intervals of one foot along the length of
the tubes.
Mr. Wagner: Is it actually possible to control juice
temperature by controlling the condensate flow? We
have tried unsuccessfully to use this type of control at
Pongola. Can you cite an example of the application
of this system?
Mr. Buchanan: As stated in this paper, this type of
control is unstable and is not regarded as good
practice. I think Mr. Gunn may provide a further
practical example.
Mr. Gunn: We have not had satisfactory results
from juice temperature control by condensate throttling
controllers. I would like to add that I feel that
the calculations for the factors in this paper appear
more academic than practical. After all, a heater
installed in a factory must be able to maintain temperatures
for a week's run under the fouling conditions.
Mr. Buchanan: You are quite right in that a heater
must be designed to cope with fouling over a certain
period, however, many heaters are allowed to scale so
badly and still attain the required temperatures that
one wonders if they have not been oversized. It is the
turn-around time between cleanings that I am questioning
and I feel that some benefit could be derived
by investigating the economic optimum cleaning
frequency against the installed heating surface.
Concerning the practical aspect of the empirical
calculations for individual and overall heat transfers
coefficient I disagree entirely that these are of academic
interest only. I have pointed out their limitations
but one of the purposes of this paper has been
to show by calculation that these coefficients compare
well with measured data in practice. The empirical
formulae are based on practical data from a wide
field and are used for design purposes in the absence
of such practical data. These formulae provide an
essential basis for the prediction of coefficients under
different operating conditions. In the absence of
specific performance data from practical measurements
these formulae provide the only means for
resolving controversies regarding vertical and horizontal
heaters, etc.
Mr. Young: Were the figures for thermal conductivity
of sucrose solutions taken from Honig?
Mr. Buchanan: They were taken from a table in
Honig's book and were subjected to a multiple regression
analysis in order to provide the formula
relating thermal conductivity at different temperature
and concentration levels.
Some of his figures were inaccurate, possibly due to
printing errors, and the formula eliminates these
errors.
101

102
Proceedings of The South African Sugar Technologists 1 Association - March 1966
PROCESS STEAM
PRODUCTION USE AND CONTROL
Introduction
In the short time available, it is not possible to go
into a lot of detail about Process Steam, or for that
matter, to cover the whole field of the Production Use
and Control of Process Steam, so I shall confine
myself to a few significant and interesting points.
Now, it is inevitable that at some stage most of us
wonder: "Why do we use steam for Process heating,
are there not other substances better suited to carry
out this job?" Well, there might be better substances
but if there are, they will have very little chance of
ousting steam by now, when one considers the time,
thought, study and money that has, for many years,
gone into perfecting the use of steam for process
heating. It would have to be a very far-seeing person
who would be prepared to spend millions of rand and
years of time developing some other substance to the
point where it can be shown to be better than steam.
Thus, we can concentrate all our energies making the
best use of steam, safe in the knowledge that our
energies will not be wasted by the sudden appearance
of something better.
Steam has a number of advantages—firstly, it is the
vapour stage of water which is common in all industrial
areas and, as found, is harmless and therefore
easy to handle. In changing from the liquid to the
vapour stage and back again it absorbs and gives out
large amounts of heat reasonably easily. In its liquid
stage it occupies a comparatively small volume and,
with a reasonable amount of care, can be heated to
very high temperatures without danger of dissociation.
This latter attribute should not concern the average
mill or factory engineer because these very high
temperatures are only used in the case of the generation
of large amounts of electricity and a few other
specialist applications.
Production of Process Steam
Steam for industrial processes can be obtained in
three main ways, namely:
(1) Directly from boilers at the temperature and
pressure required.
(2) From pass-out turbines.
(3) From reducing stations.
Basically, of course, all process steam originates
from boilers of one kind or another—the types of
boilers are many and varied, the actual design being
influenced by individual firms, application, type of
fuel, etc. There are a few unusual sources of steam
such as the natural hot springs in New Zealand and
Italy where the steam has been harnessed for the
generation of electricity, or the patented steam generators
burning oxygen and hydrogen which, although
they are only about 40 cubic feet in total volume, can
produce their full output of up to 100,000 lbs hr at
By D. T. O. GRIFFITH
500 p.s.i. or more within a few seconds of being
started up.
However, as in the first case the steam is very wet
and contains large quantities of corrosive substances
and, in the latter case, the fuel costs are about 100
times the fuel costs of a normal boiler, neither are
likely to concern the average production engineer who
is interested in efficient and economic production
methods. Although boilers would appear to be the
most obvious of all sources of process steam, the
choice of a boiler of suitable pressure and temperature
is not always straightforward. I shall deal with this
aspect at a later stage.
Pass-out turbo-generators are used in many
industries which require both electricity and steam in
their manufacturing processes, even in places where a
public electricity supply is readily available. The reason
for this is mainly economic.
Public electricity utilities can only use part of the
heat put into the steam for the generation of electricity,
the larger proportion being thrown away in the
condensers. For example, in a generating station using
steam at 1,500 p.s.i.g. absolute and 1,000° F., the total
heat of each pound of steam entering the turbine is
1,488.5 B.T.Us. while the total heat of this steam
when it enters the condenser at 28 ins mercury
vacuum is 1,105.7 BTUs. Thus only 383 B.T.Us. of
heat are used for generating electricity from each
pound of steam supplied, the rest being dissipated to
atmosphere or the local river through the condenser
cooling water. Now if we assume that the temperature
of the condensate extracted from the condenser is
101° F. and this is the starting temperature of the
cycle, then as this water contains 69 B.T.U.s/lb the
total quantity of heat supplied to the steam by the
boilers is 1,419.5 B.T.Us. Therefore the 383 B.T.Us.
used for the generation of electricity amounts only
to 27% of the heat supplied by the boiler. By means
of bleeding steam from the turbines for feed water
heating and also the use of re-heat cycles, this figure
can be increased to 33 or 34%. Naturally when one
purchases electricity from a public utility company
one has to pay for all the heat used up, not only for
that small portion used for the generation of electricity.
Now, on the other hand, if a pass-out non-condensing
(back pressure) turbo-generator is installed in a
factory, the steam, after giving up some of its heat for
generating the electrical power needs of the factory, is
available for other purposes such as heating and
drying, thus the cost of electricity is directly proportional
to the heat absorbed and is therefore about 34 %
of the cost of electricity generated by a public utility
company. I am referring here, of course, only to
marginal costs. Such things as the interest and
redemption of the capital required for the installation
of the equipment wouldhave to be taken into consideration
together with the fixed charges and demand

Proceedings of The South African Sugar Technologists' Association—March 1966
charges made by the supplier of electricity before one
could determine whether it was an economical proposition
or not. The large number of pass-out turbines
installed in mills and factories throughout the world
would indicate that it very often is an economic
proposition.
Naturally the steam and electrical demands of a
mill would have to be balanced at all times for such a
unit to be workable and as this is a virtual imposibility,
pass-out turbines are fitted with either an atmospheric
relief valve or, which is more economical, an L.P.
stage and a small condenser, to dispose of excess
steam. To compensate for shortage of steam in the
pass-out main, reducing stations are generally fitted
between the H.P. steam main and the pass-out main.
A reducing station normally consists of a double seat
valve as illustrated in the sectional view above, or in
the case of a small difference in pressure between the
two steam mains, a butter-fly valve operated by a
pneumatic or hydraulic motor.
The pass-out steam from the turbo-generator is
controlled from a sensing element in the pass-out
main, the control being a series of valves or a disc
valve mounted on the steam chest to which the passout
main is connected. These valves control the steam
flowing to the lower pressure stages of the turbine
or to the atmosphere, and not, as is often believed,
Air Operated Double Seat Reducing Valve
the steam to the pass-out. Thus if the pass-out pressure
rises, due to a fall in steam demand, the valves would
open allowing steam to escape down the turbine to
the condenser or directly to atmosphere. It will be
realised that in the case of a condensing turbine, as
more steam enters the L.P. cylinder, and provided the
electrical load is constant at the time, the speed of the
turbine will tend to rise. This speed change is countered
by the speed governor which will reduce the steam
supply to the turbine sufficiently to control the speed.
As a result of this reduced amount of steam entering
the turbine there will be less steam available for
pass-out and therefore the pass-out pressure falls,
which in turn causes the valves controlling the steam
to the L.P. cylinder to close partly thus increasing the
pass-out pressure and reducing the steam to the L.P.
cylinders which causes the turbine to slow down.
This is again compensated for by the governor and the
reverse process takes place. If this is not properly
controlled, severe hunting of the turbine speed and
pass-out pressure can take place. Various means are
adopted to prevent this hunting, all of which basically
consist of means of damping the pass-out control
valves to such an extent that they are sufficiently slow
acting to prevent hunting. In this manner equilibrium
conditions are reached fairly quickly although slight
fluctuations of the pass-out pressure have to be
tolerated.
103

104
Proceedings of The South African Sugar Technologists' Association-March 1966
(a) (b)
Actuating Mechanism for Pass-out Control Valves on a 10 M.W. Double Pass-out and Condensing Turbine.
The response rate of reducing stations which boost
the pass-out pressure must similarly be carefully
adjusted to prevent hunting between the turbine
pass-out control valves and the reducing station.
Reducing stations are anathema to many engineers,
as no useful work is done when changing the steam
pressure from a higher value to a lower value. The
attitude of some engineers to reducing stations is
summed up very succinctly by Sir Oliver Lyle who
says that "A reducing valve is, from the thermodynamic
point of view, an invention of the devil. It
sets out to degrade good heat, to dissipate the good
high potential. It performs its vile task to perfection
until it goes wrong. The use of reducing valves might
be called an admission of defeat—it is the easy way
out". 1 I personally cannot agree completely with his
attitude as these valves do perform a useful and in
many instances a necessary function.
I would now like to refer back to the question of
the choice of pressures for process steam. The mill or
factory engineer has little say in this matter normally,
as operating pressures are generally specified by the
manufacturers of the equipment installed. Within
small limits the manufacturer can vary the design
pressures to fit in with steam supply equipment in an
existing plant, but frequently the design pressure is
based upon the temperature required for a particular
process, so a steam pressure having a saturated
temperature of the desired degree is chosen. Thus it is
frequently necessary to generate and supply steam at
one pressure to suit a particular process and then
reduce and desuperheat the steam to supply another
process.
The question of desuperheating is one that frequently
arises. It is an axiom that superheated steam is
unsatisfactory for heat exchange processes. The
reason for this is that the superheated steam has to
(a) H.P. Valves — 160 p.s.i.
(b) L.P. Valves — 45 p.s.i.
fall in temperature to give up its heat, and where the
temperature difference between the steam and the
substance to be heated is small, this fall in temperature
of the steam reduces the temperature differential thus
slowing down the heat transfer rate. Far more important
is the fact that dry steam is a bad conductor of
heat and therefore the steam at the heat transfer
surface, which has already given up some of its heat
and has dropped in temperature acts as an insulating
blanket preventing the higher temperature steam from
giving up its heat, thus slowing down the heat transfer
process even further. On the other hand dry saturated
steam gives up a large quantity of heat—latent heat—
without change in temperature, i.e. the temperature
differential between the steam and the product is
maintained throughout the transfer of this heat.
The difference between the rate of heat transfer of
saturated steam and superheated steam can be clearly
seen if one looks at two cases of steam at 292° F. In
the first case 45 p.s.i.g. dry saturated steam with a
saturation temperature of 292° F. is used. When
condensing to water of the same temperature this
steam gives up 915 B.T.U. for each lb. of steam. In
the other case using 40 p.s.i.g. steam superheated to
292° F. (i.e. about 5° F. of superheat) only 1 B.T.U. of
heat will be given up by each lb. of steam for every
2° F. drop in temperature until the saturation temperature
of 287° F. is reached after which it gives up its
latent heat in the normal way without further change
of temperature, but by this time the temperature is
5° F. below the temperature required for the process.
Therefore when the supply steam is superheated, it
is necessary to desuperheat it by injecting water into
the steam line until the temperature has been brought
to near saturation temperature for the particular
pressure being used, before it enters the process unit.
In practice it is often found that a small amount of

Proceedings of The South African Sugar Technologists' Association—March 1966 105
superheat in the steam being supplied to a unit gives
the best results, especially where rate of running has
to be considered side by side with efficiency. The
explanation for this apparent contradiction lies
mainly in the fact that by having a small amount of
superheat, all lines and valves, etc. are kept dry and
no energy is used up in moving unnecessary condensate
around the system and the condensate removal
system is handling the minimum amount of condensate.
This reasoning applies equally well to long steam
supply lines, as there is always a certain amount of
heat loss with even the most efficient lagging, so by
ensuring sufficient superheat at the supply end to have
dry steam at the process end, the trouble of having to
drain the steam line at frequent intervals with the
consequent loss of good condensate, or expense of
recovery, is avoided.
Methods of Control
Control of process steam consists basically of
measuring some parameter of the finished product,
sensing any change from a required standard and as
a result causing a steam valve to open or close to
increase or decrease the amount of steam flowing to
correct the deviation from the set standard.
Measurements such as density, viscosity, temperature,
electrical conductivity, pH, etc. can be used
depending on the results required. Where this can be
done the control can be reasonably simple, but often
it is difficult or impossible to take measurements from
the product, or there is no suitable apparatus on the
market to take the measurements required. In cases
like this it is necessary to turn to other—indirect—
means of determining what changes in the rate of
flow of process steam are required.
As an example, on modern paper machines where
a continuous wet sheet of paper has to be dried by
passing it around a number of rotating steam heated
drying cylinders, it is obvious that a reliable means of
measuring the average moisture over the whole width
—more than 200 inches on many modern machines—
of the fast moving paper sheet, would be a virtual
imposibility. As an alternative one thinks of temperature
control as a means of ensuring a uniform drying
rate. This, however, does present difficulties because
if the temperature chosen is one above the saturation
temperature of the steam being supplied, then superheated
steam would have to be used and no condensation
would be permitted, as this would imply a lowering
of the temperature, so the steam would have to be
blown in and out of the cylinder without making use
of the latent heat. If, on the other hand, the actual
saturation temperature is chosen, then the control
becomes insensitive, as large quantities of heat can be
transferred without change of temperature.
Pressure control is another method which can be
adopted to maintain a constant drying rate of the
paper and in this case provided that the pressure
required is below the pressure of the steam supplied,
a reasonably precise control can be maintained. Where
such a system is used, should the moisture content of
the paper in contact with the cylinder suddenly
increase, it will absorb more heat from the cylinder
thus lowering its surface temperature. As a result the
steam in the cylinder will condense at a greater rate
and cause a drop in pressure. This drop in pressure
is felt by the pressure controller which opens the
automatic steam valve, permitting a greater flow of
steam which restores the pressure. Where a paper
machine has 60 or 70 drying cylinders, it would
theoretically be correct to control each cylinder
pressure individually, but in practice this presents
engineering difficulties in addition to excessive costs,
so the machine is divided into four or five groups of
cylinders, all the cylinders in one group being supplied
from a common, pressure controlled, steam header. A
refinement of this type of control is to measure the
differential pressure between the common steam
header and the common condensate header of a group
of cylinders, and by maintaining a constant differential—
achieved by bleeding steam to adjacent drying
sections— a drying rate proportional to the moisture
content of the paper can be maintained.
The instruments of control can be electrical,
electronic, hydraulic or pneumatic, with pneumatic
generally predominating where an extremely high rate
of response is not required. Some of the reasons for
this is that where compressed air is used for transmitting
a signal, the same compressed air can be used
for the actual mechanical operation of the necessary
valves, etc. and in addition the compressed air is not
toxic or dangerous to those working with it, nor do
leaks cause damage to the final product or mess up
machinery.
Air Motor as Fitted to Automatic Valves

106 Proceedings of The South African Sugar Technologists' Association—March 1966
Steam Trapping, Draining and Venting.
A major aspect of all process steam control is
proper draining and steam trapping. It is obvious that
if the steam space of a piece of apparatus becomes
partly filled with water, the heat transfer area available
to the steam is lessened, resulting in a drop in efficiency
and a reduced running rate. Conversely if steam is
allowed to blow straight through the apparatus, then
it is not giving up all its heat, and this is inefficient
and wasteful. To prevent a build-up of condensate or
loss of steam, automatic steam traps are fitted which
will ensure that condensate is allowed to drain freely
without permitting the passage of steam.
(a)
In conjunction with steam trapping is the question
of air venting. The engineer dealing with process
steam should never forget that all steam contains a
certain amount of non-condensable gases, and are
grouped together under the name "air" in this
connection. In many instances it is actual air that has
been drawn into steam spaces through glands and
traps during shut-down time when the condensing
steam creates a vacuum. Now, as is well known, static
air is a bad conductor of heat. This feature is used
frequently in every-day life, for example, warm
clothes are made from material which traps and holds
static numerous small pockets of air. Now when air
becomes trapped in a steam space, it can reduce heat
transfer in two ways. Firstly it can form a thin film
all over the heat transfer surface. This thin film is very
difficult and sometimes impossible to remove. Steam
inlets should be arranged in such a manner that the
Two Typical Steam Traps
(a) Mechanical — (Ball Float)
(b) Thermal — (Thermostatic)
Steam traps are divided into two main classes—
mechanical and thermal—and are probably the most
maligned of all the pieces of apparatus used with
process steam. Over and over again they are blamed
for the poor performance of a process when, in fact,
they are being expected to do a job of work for which
they are not suitable—because of a wrong choice in the
first instance. Yet if properly installed and correctly
sized they are reliable and long lasting. They are in
fact a simple piece of apparatus which seldom fails.
A careful study of the correct application of steam
traps in process steam control will be well rewarded
by increased efficiency and reduced costs.
(b)
steam enters at high velocity in a direction which will
scour the heat transfer surfaces, and by this means
any air that tends to cling to these surfaces is removed.
Secondly, pockets of air can form in the steam space.
This is as a result of the slow accumulation of air in
an area which is not properly vented. Trapped air
pockets prevent steam from coming in contact with a
portion of the heating surface and although the air
does transmit a small amount of heat, the effective
heating surface is reduced, which reduces the capacity
of the equipment. It is, therefore, essential that all air
entering a steam space is properly vented. It is often
difficult to determine where these pockets will form—
generally, but not always, low down and remote from
the steam inlet.
Sir Oliver Lyle states that "The essence of air venting
is finding this remote point, or better still, to make
the steam follow a certain path to a pre-arranged

Proceedings of The South African Sugar Technologists' Association—March 1966
remote point. Once the remote point has been found,
automatic venting can take place which will ensure
that the maximum use is made of the heat transfer
surfaces." 1
Steam traps incorporating air venting are available,
although, in many instances it is necessary to install
separate automatic air vents as the condensate path
before the steam trap can prevent the removal of air.
Conclusion
No talk in the use of Process Steam would be
complete without mention of heat losses. This can be
one of the biggest efficiency destroyers on any steam
system. All apparatus which carries or controls steam
at a temperature above atmospheric will radiate heat
to some degree, so if steam is generated at Point "A"
and the heat is required at point "B", why waste
some of the heat by heating up the space between
point "A" and point "B"? Besides wasting heat, the
problems of wet steam lines and condensate removal
are accentuated. Good lagging is essential throughout,
and not, as is so often the case, only in positions
where people are likely to come into contact with the
hot surfaces. Although no lagging is 100% efficient,
if the tables obtainable from most lagging manufacturers
are consulted, and the recommended thickness
and type of lagging for any specific condition is
used, a happy balance between excessive lagging
costs and acceptable heat losses is obtained.
Finally, I would like to mention the safety
regulations contained in the Factory, Buildings and
Works Act and Regulations of 1952, to which all
apparatus installed in mills in this country have to
conform. Consulting these regulations can be a
considerable help to engineers in the design and proper
maintenance of equipment. Also as these are formed
to ensure the safety of the people who have to work
on or around the equipment, breakdowns are less
likely to take place on equipment which conforms to
these regulations, so owners should not begrudge any
money spent in this connection, they should rather
look upon this expense as a form of insurance.
References
107
1. Oliver Lyle, "The Efficient Use of Steam", H.M. Stationery
Office, London. 1947.

108 Proceedings of Tlie South African Sugar Technologists' Association—March 1966
BOILER OPERATION, MAINTENANCE AND TESTING
Introduction
Boilers are akin to wives and their treatment is
therefore a highly controversial subject. Although
there are many of various shapes and sizes, no two
react alike, but with encouragement, some are efficient.
Others however, are simply temperamental and if
neglected, all become dangerous and are liable to blow
off steam. Providing sufficient care is taken of them,
there is a fair chance of obtaining reasonable results,
but the fellow who claims to know all the answers will
surely get his fingers burnt.
Boiler Operation
Combustion
Good combustion depends on Temperature, Time
and Turbulence, commonly known as the three T's.
Anything combustible will burn if given sufficient
time and temperature, but in this modern age time
appears to be in short supply. With the advent of
bigger and better boilers, assistance in the form of
mechanical firing equipment, fans, economisers and
air pre-heaters is necessary. The temperature is
increased by preheating the combustion air and
turbulence is created by the introduction of over-fire
air known as "Secondary Air". These two features in
conjunction with each other reduce the Time for the
fuel to reach ignition temperature and burn to ash.
When we consider combustion in a furnace, we find
that various practical aspects make it impossible to
burn a pound of fuel with only the theoretical quantity
of air. Losses occur due to the construction of the
grate, furnace, and the fact that in practice the air and
fuel cannot be sufficiently intermixed to ensure that
every atom of carbon and of hydrogen combines with
its complement of oxygen. In many instances, because
of uneven fuel grading, the air takes the line of least
resistance, with the result that some of the air passes
into the furnace without combining and appears in
the form of free oxygen in the flue gas. Where insufficient
air has passed through the fire, the fuel is not completely
consumed and there is a loss in efficiency
arising from the unburnt carbon rejected with the ash.
While the application of excess air gives better
mixing and reduces the unburnt gas loss and unburnt
carbon loss, there is a limit to the quantity of excess
air which can be used without introducing a further
serious loss. Air introduced to a boiler unnecessarily
will not only fail to serve any useful purpose but it
will carry away an amount of heat depending on the
difference between the chimney gas temperature and
the ambient air temperature.
The art of efficient combustion is to find the best
compromise between chimney loss and unburnt fuel
loss.
The quantity of air used for combustion depends on
the compromise made between chimney and unburnt
fuel losses, but the pressure of the air required is
By S. G. HOLTON
closely related to the grading of the fuel. When
burning coarse fuel a low undergrate air pressure is
sufficient due to lack of fuel bed resistance, but a high
air pressure is necessary when burning fines.
This feature was illustrated during my earlier days
at sea on a "coal burner". The boilers were double
ended Scotch Marine and the stokers were tough
Liverpool Irishmen, who appeared to work only when
broke. During the outward bound voyage the leading
hand of each watch had the coal trimmers select the
larger pieces of coal from the bunkers. This enabled
the stokers to shovel huge quantities of coal onto the
grates until the fuel bed thickness was anything from
10" to 12". The stokers were then able to sit on their
shovels and relax whilst the fires burnt down with
forced draught pressure rarely exceeding J" w.g. to
maintain steady steam conditions. The last lap of the
homeward voyage was a very different story. By this
time the coal in the bunkers was mainly fines, which
required to be thinly and evenly spread over the fires
at frequent intervals. Regular slicing and raking of
the fuel bed was essential to permit ingress of the
forced draught now at 21" w.g., the maximum available
from the F.D. fan. From this point on, a constant
battle ensued between the engineers and stokers.
Possibly it was only the king-size thirst and the
thought of getting back to their local pubs that gave
the stokers strength to produce sufficient steam to get
us back to port! Luckily Father Neptune wasn't
interested in smoke abatement, otherwise we would
not have made it.
Fuels
When steam is used for process in a factory, the
cost of the fuel greatly influences the price of the end
product, and when fuel costs are high the steam
generating plant must maintain the highest possible
efficiency.
The efficiency aspect of the boiler plant however,
becomes less acute when using bagasse or spent bark,
as quantities of these products in excess of the normal
requirements, can prove to be an embarrassment due
to storage problems.
Although the cost of coal per ton delivered to the
bunkers is constant, the amount of steam generated
per ton of coal varies in accordance with the condition
of its presentation to the boiler. Mechanical stokers
can be even more temperamental than their Liverpool
counterparts when faced with poor grading, bad
conditioning and segregation.
A growing demand, with increased mining costs
causes difficulty in obtaining a consistent size of coal.
The quantity of explosives used per ton of coal extracted,
has increased considerably over the past ten years,
and this, together with mechanisation tends to increase
the proportion of fines. A mechanical loader at the
colliery has no discrimination in what it loads. It

Proceedings of The South African Sugar Technologists' Association—March 1966 109
loads coal and shale and fines, and it loads them
altogether.
Separation of the larger from the smaller coal is
caused by movement. In general, it may be stated
that when coal is in free movement the larger coal
will move a greater distance. When a pile is formed
either on the ground or in a boiler hopper from a
central feed, the larger coal will be found at the outside
edges of the hopper. Unless this is prevented, the
zones of the fire which have received the coarse coal,
burn rapidly with a long flame caused by the large
quantity of excess air. The severe heat at the edges of
the grate accelerates the furnace side wall deterioration.
The zones of the fire which have received the finer
coal burn more slowly due to air starvation and
generally form clinkers containing large quantities of
unburnt coal. Under these conditions, whilst the total
air flow should be sufficient to give a good percentage
of carbon dioxide in the flue gas, the size segregation
of the fuel results in a double loss; i.e., low percentage
of CO2 and unburnt carbon.
Dry coal containing a high percentage of fines
cannot be efficiently burnt on a stoker due to its high
resistance to air flow. The addition of free moisture
to the coal influences the facility with which air passes
through the smalls because the surface moisture tends
to agglutinate the fines, thereby reducing the fuel
resistance to aeration. This is a physical action and
probably independent of chemical properties.
The many complexities of coal preparation are
beyond the scope of this paper.
Opera tion—General
The two most important factors in good operation
are knowledge of the equipment and an interest in its
performance. It is essential for an operator to observe
the proper operating sequences and to understand
the operations which are carried out automatically by
the equipment. If an operator acquires a sound
knowledge and understanding of the plant he automatically
develops an interest in the operating records,
and should a variation occur with the data being
recorded he may possibly be able to take corrective
action and avoid damage to equipment.
It is essential that any adjustments to the firing
equipment are made gradually in order to maintain
steady conditions. For instance, the outcome of any
alteration to a fuel bed on a travelling grate will not be
apparent for approximately half an hour, therefore
any heavy-handed action could take up to an hour to
correct.
For records, it is recommended that the following
data be recorded hourly:
1. Water level.
2. Visual observation of the fire.
3. Steam and feed water temperatures, pressures
and flow.
4. Temperature and pressure of air entering and
gas exit from the principle portions of the boiler.
5. Percentage of CO2 or O2.
The term automatic control on a boiler is frequently
mis-interpreted to infer that the unit is completely
self-regulating thus rendering the services of an
operator unnecessary. Unfortunately this condition
only applies, and then only remotely, to such fuels as
oil and gas and to a lesser degree pulverized coal where
the conditions of moisture, calorific value and grading
or viscosity of oil, remain constant with each consignment
received. The use of automatic control can only
be justified in the regulation of all functions which do
not require the judgment of an operator, as the
control equipment can only monitor the weight or
volume of a fuel and then regulate the ratio of air for
combustion to suit the steam demand. Variables in
the fuel, such as grading, moisture and calorific value
can only be corrected by visual observation and the
controls then manually trimmed accordingly.
Care of Supheaters
The superheater drains should be fully opened
prior to lighting fires.
Non-drainable superheaters require sufficient heat
to boil the condensate out of the loops but in order to
prevent abnormal temperature differences between
the portion of the tube containing water, and the
other, the gas temperature must be regulated not to
exceed 900° F. Alternatively, the firing rate may be
controlled to increase the saturation temperature
120° F per hour.
The most reliable method of determining the time
required to raise pressure within safe limits is by the
installation of a few thermo-couples on the outlet legs
of the superheater. Since no steam will flow through
a tube partially filled with water, the temperature
recorded will read saturation temperature until the
tubes are cleared. When a flow of steam is established
the temperature will rise to approximately 75° F above
the saturation temperature.
The superheater drains should be left open until the
boiler is on range when a flow of steam through the
superheater is assured. When a boiler is being brought
up to pressure the air-vent should be closed only when
the pressure has risen to 25 p.s.i. This will ensure the
expulsion of all air.
Steam Quality and Purity
The amount of moisture in steam after separation in
the steam drum is defined as quality and the amount
of solids in it, as purity.
The water in a boiler accumulates suspended and
dissolved solids, which unless checked will lead to the
contamination of the steam. Steam is always contaminated
to a greater or lesser degree and even if the
water delivered to a boiler is chemically pure, it is
possible for droplets of water to be carried with the
steam into the superheater. Excessive amounts of
moisture in the steam will cause loss of superheat,
corrosion, scaling and eventual failure of the superheater.
It is most important to prevent "carry-over"
and whereas the design of the boiler drum internals
takes care of the usual sources of moisture and impurity
with extremely high efficiency, the system cannot

110
cater for severe overloading. These internals must be
given an opportunity to function and if they become
submerged by a high water level or choked with
chemicals, superheater failure will occur.
A brief explanation of what happens in the steam
drum may assist in stressing the importance of maintaining
correct chemical condition and level of the
water.
As steam bubbles travel from the riser pipes into
the drum and reach the water level, they burst into
fragments causing drops of moisture to be projected
into the steam space. The height to which these drops
rise above the water level depends upon their size and
velocity. As the velocity of the droplets is less than
that of the steam rising from the surface, the carryover
varies from small amounts of atomised mist to
heavy carry-over, when the capacity limits of the
drum internals are much exceeded. The size and
velocity of the droplets is greatly increased when
steaming the boiler in excess of its designed rating, or
when there is a sudden drop in steam pressure. A drop
in pressure leaves the water substantially superheated
above the boiling point associated with the higher
pressure. This results in a greatly increased quantity
of steam bubbles causing the boiler contents to swell
considerably above the former level. Under severe
conditions of priming, steam may carry as much as
50 % of moisture by weight which represents less than
1 % by volume. Slugs of water carried over can cause
serious damage to the plant as they contain dissolved
and suspended solids, which become deposited in the
superheater tubes or on the turbine blades. The main
danger is that solids trapped in the superheater form
an insulation between the steam and the metal,
resulting in eventual failures of the metal through
overheating.
Additional factors which tend to promote priming
are: too high a water level; excessive alkalinity of the
water; excessive total solids; organic matter and
sludge.
The water gauge does not accurately indicate the
true water level in the drum. For instance, when a
boiler is under load the water level in the steam drum
is always higher than that indicated by the gauge
because the solid column of water in the glass has a
higher density than the steam/water mixture in the
drum against which it is balanced.
Blowing Down
Most plants use the chemical analyses of the water
from the boiler as a guide to determine the amount
and frequency of blowing down. Where such analyses
are not made, the boiler should be blown down at
least once every 24 hours. The amount of blowdown
will depend upon the class of feed water and quantity
of steam generated. Economisers and water cooled
furnace walls should never be blown down whilst the
boiler is in service as it is possible to create steam
pockets or even reverse the circulation which could
result in tube failures. Blow down valves for the above
mentioned equipment should be padlocked whilst the
boiler is in service.
Proceedings of The South African Sugar Technologists' Association — March 1966
Balanced Draught
All large boilers are equipped with forced and
induced draught fans and in most instances secondary
air fans.
Balancing the draught is effected by adjusting the
induced and forced draught fan pressures until the
furnace draught gauge indicates no pressure difference
between the inside of the furnace and the outside
atmosphere. In actual practice the furnace is always
maintained under a slight suction thus permitting a
slight ingress of air which prevents the setting from
overheating and provides a precaution against possible
blow-back and injury to the operating personnel.
On modern boiler plant equipped with instrumentation
a change of CO2 or O2 in the flue gas indicates
to the operator that there has been a change in
combustion conditions and this, with visual inspection
of the fire, will help him to adjust the draught suitably.
Economisers
Economisers with a recirculating connection should
have the valves open whilst pressure is being raised
preparatory to operation. This is to ensure that the
tubes are at all times completely full of water. When
the boiler is on range and the economiser has been
taking feed water for approximately 5 minutes, only
then should the recirculating valve be closed.
Water-washing
The constituents of the flue gases which cause
fouling of the economiser are dust, fly-ash particles and
acid forming substances. Loose dust collects initially
on the down stream side of the bank of tubes mainly
because of the reduction in the velocity of the gases
leaving the bank. The worst fouling occurs during the
period the boiler is being sootblown. These deposits
are effectively removed by mechanised waterwashing.
The apparatus consists of a motor driven oscillating
nozzle lance installed above the top bank of tubes and
provides a slowly moving curtain of water directed at
right angles to the horizontal tubes. 2 to 3 gallons per
minute per sq. ft. of projected area should be used
during this process to effectively dilute the acid solution,
and also to ensure that sludge or solid matter,
displaced from the upper banks does not accumulate
in the lower bank.
The duration of the washing period will vary with
operating conditions and characteristics of the fuel,
but washing should be continued until the wash water
effluent is clean.
Cast iron airheaters are waterwashed in a similar
manner.
Soot-blowers
The frequency of soot-blowing depends upon the
nature of the fuel but generally systematic cleaning
should be a regular feature of operation. Normally
once a day is sufficient. The soot-blowing sequence
should commence at the front of the boiler, working
toward the rear. This is to drive the loosened deposits

Proceedings of The South African Sugar Technologists' Association — March 1966
out of the boiler system. During soot-blowing routine
the furnace draught should be increased to give
sufficient suction to prevent a furnace pressure through
the released steam and to assist the flight of the fly-ash.
Boiler Maintenance
Boiler maintenance falls into two distinct groups of
activities. One is the checking and overhaul of
mountings, boiler auxiliaries, firing equipment and
control gear. The other is the cleaning and inspection
of pressure parts, supporting structures, brickwork,
baffles, casings and ducting.
Depending upon their size and cost, boiler mountings
are often, as a routine measure, replaced with
reconditioned spares during short outages so that
many of the smaller items on a boiler are always in a
good state of repair and need no special attention, when
the unit is shut down for its periodic overhaul. The
same applies to certain auxiliary equipment and
control gear.
The checking and overhaul of major auxiliary plant
and boiler equipment is usually carried out only
when the boiler is shut down for overhaul, and, if this
is the case, it is necessary to ensure that the checking
and overhaul programme is so thorough and comprehensive
that the equipment concerned can be
relied upon for good service until the next scheduled
outage.
Boiler pressure parts are subject to statutory
inspection and testing and internal and external
cleaning of the boiler is required by the Government
Inspectors to enable them to carry out their
examinations satisfactorily.
In regard to internal cleaning, it is, of course,
possible using modern plant and techniques to
approach the ideal in steam raising by using feed
water which will not give rise to any deposits on the
internal surfaces of boilers. In many cases, however,
where the control of the water conditioning plant is
faulty or inadequate, deposits of various sorts occur
inside boilers and high maintenance costs are incurred
in removing them. Apart from the desirability of
removing deposits for inspection of pressure parts,
there is also an imperative need to remove them
because of their effect on heat transfer through tube
walls or other pressure parts and the possible failure
of these items through local overheating. In addition
to sludge or scale formation other characteristics of
the feed water may give rise to corrosion of boiler and
superheater tube surfaces, and the serious problem of
caustic embrittlement has also sometimes to be faced
when feed or boiler water conditioning has been
neglected. The internal cleaning and inspecting phase
is therefore vital, and must be geared to suit the
circumstances existing in each case. Much the same
applies to external cleaning and inspection.
External corrosion of pressure parts, ducting, airheaters
and fans, fouling of gas-passes by fireside
deposits of various types, damaged brickwork, fire
damaged parts and sometimes erosion by fly-ash are
among the unpleasant discoveries one can make during
an inspection.
411
Inspections must be carried out by competent
personnel, and an appraisal of all causes and effects
should be made the responsibility of senior technical
staff.
Boiler Efficiency Testing
Some boiler plants are so fully instrumented and
staffed that it is always possible to ascertain without
difficulty what efficiency is being attained. At the
other end of the scale there are boilers equipped only
with the instruments required by law.
When fuel costs are a serious consideration, boiler
plant manufacturers are nearly always faced with
contractual obligations to meet specified efficiencies
on tests after the erection of new plant, and the user
during the subsequent life of the boiler is also likely
to be interested in the results of periodic efficiency
checks.
In some cases the efficiency of the steam consuming
section of the plant is constant and the total weight
of product obtained from the steam consuming section
of the plant can be directly related to the amount of
steam delivered to it from the boiler house. Should
this condition apply it is possible that some continuous
check on the efficiency of the steam raising plant can
be obtained readily by comparing the weight of the
final product obtained during a period of time with
the amount and calorific value of the fuel supplied
during the same period.
When this is not possible there are two other ways
in which the efficiency of a boiler can be assessed.
Both methods require a great deal of care if any
reliance is to be placed on the results.
The direct method of boiler testing entails the
measurement of the weight of the fuel used during a
specified time, and its calorific value, the measurement
of the weight, temperature and pressure of the steam
supplied by the boiler in the same time and the
measurement of the feed water temperature. In a
variation of this the quantity of feed water supplied
to the boiler is measured instead of the steam quantity.
These measurements are sufficient for the purpose of
assessing the efficiency of the boiler proper, as it is
clear that the nett heat imparted by the boiler to the
steam can be calculated as can the total heat supplied
by the fuel to the boiler in the same time. In some
cases, however, the efficiency of the boiler plant as a
whole is required, and an allowance must be made in
the calculations for the consumption of power by
essential auxiliary plant such as fans and feed pumps.
The indirect method of boiler testing is carried out
by examining the channels by which heat is rejected,
and therefore wasted by the boiler. It entails careful
sampling and an ultimate analysis of the fuel,
collection and weighing of ash, riddlings and dust
and the determination of the percentage combustible
material remaining in such refuse. A careful analysis
of the flue gas and measurement of flue gas temperature,
and that of the ambient air is also required.
Certain losses such as radiation losses, cannot be
measured and it is necessary to assume a figure for

112
these based on a heat balance compiled from the
results of a complete boiler test by the direct method.
Many tests have been failures and of no real value
through insufficient care in obtaining or recording the
data obtained. To illustrate the importance of being
meticulous and obtaining correct data, Alfred Cotton,
an authority on boiler testing, summarises the possible
errors that occur even with greatest care.
Coal
Weighing ± 0.5%
Estimating the amount of fuel in
the stoker hopper at the start and
end of the test ± 0.5%
Error in Coal weighing ± 1.0%
Failure of moisture sample to represent
bulk ± 1.0%
Failure of C.V. sample to represent
bulk ± 0.5%
Error in analysis ± 1.5%
The total error in the coal is
therefore ± 2.5%
Water
Weighing or metering between
starting and stopping ± 0.5%
Failure of steam sample to represent
bulk steam as to entrained water ± 0.5%
Now the total error in water
evaporated is ± 1.0%
These values could vary the BTU value of the coal
from 97.5 % to 102.5 % and that of the water evaporated
from 99.0% to 101 %. If all these errors combine
in one direction, the report may show an efficiency of
75.4% or 80.8% for an efficiency that is actually
78 %. Under these conditions it is clear that regardless
of how carefully a test is conducted, the efficiency
could not be guaranteed closer than + 3 %.
Summary
Boiler Operation
Any combustible substance will burn if its temperature
is high enough for ignition, and if given
sufficient time to ignite. Turbulence of the air and
increased temperature will reduce the time for
complete combustion. Badly graded dry coal contaming
a high percentage of fines must be conditioned
prior to presentation to a boiler to avoid patchy fires,
poor combustion and loss of efficiency.
If hourly readings of the boiler performance are
logged, a reasonably consistent standard of operation
can be obtained. Superheater tube failures occur due
to raising pressure too rapidly after lighting up, and
also by carry over of impurities entrained with the
saturated steam. Carry over can be caused by operating
with the boiler water level too high, steaming in excess
Proceedings of The South African Sugar Technologists' Association — March 1966
of the rated capacity of the boiler, or a high concentration
of impurities in the boiler water. The high
concentration of impurities can be reduced by regular
blowing down. Fireside deposits on the boiler tubes
are removed by soot-blowing and from the economiser
by water-washing.
Boiler Maintenance
It is better to be sure than sorry, should be the
slogan during boiler overhauls. The external and
internal cleaning must be thorough and inspections
carried out by experienced and competent persons.
Boiler Efficiency Testing
The performance and efficiency of a boiler can be
determined in three ways:
(a) By the overall method whereby the weight of
the manufactured product is compared with the
weight and C.V. of the fuel consumed over a certain
period to produce the process steam used.
(b) By the Direct Method, by measurement of the
weight and C.V. of the fuel consumed and measurement
of the quantity, pressure and temperature of
steam produced during a specified time.
(c) By the indirect method (losses method) in
which the heat rejected in the solid refuse and in the
flue gas is ascertained by analysis and calculation.
All methods require a great deal of care if accurate
results are required.
Mr. Hulett (in the chair): Mr. Holton has rightly
pointed out the great importance of fitting instruments
to boilers.
Mr. Heslop: What is the effect of intermittent
boiler operation, as practiced in the sugar industry
compared to continuous operation with an annual
shut-down?
Mr. Holton: Frequent heating up and cooling down
causes spalling of the brickwork or refractories and
unless correct lighting up procedure is adhered to on
every occasion, thermal stresses will eventually cause
failure of the pressure parts. Superheater tubes are
particularly vulnerable to damage during this period.
Mr. Griffiths: Is it possible to waterwash superheaters?
Mr. Holton: We have off-load washed superheaters
prior to examining them for erosion. Rust caused by
water on the tube surfaces clearly indicates any portion
that has been subjected to erosion. Great care
must be taken to avoid wetting the brickwork and the
grate. We normally cover the grate with tarpaulins.
Mr. Steffen: Do you recommend waterwashing
economisers whilst steaming, or during the weekend
shutdown?
Mr. Holton: Definitely during the weekend and
preferably before the boiler has cooled down. This
will give the metal surfaces time to thoroughly dry
out prior to the unit returning to service.

Proceedings of The South African Sugar Technologists' Association—March 1966
STEAM TURBINES-THEIR CONSTRUCTION,
SELECTION AND OPERATION
Introduction
The first steam engine built by James Watt in the
year 1769 was the advent in substituting the low
energy rates produced by wind, water, man and beast
for the higher mechanical power produced by a
machine. A further milestone was in the year 1866
when Werner von Siemens invented the principle of
producing electricity from a rotating machine, the
so-called electro-magnetic principle. Coupled to the
steam engine this had the advantage of producing
power centrally and making it available at a large
number of points.
The steam engine and also to a latter extent the
diesel engine had a limited capacity in producing
power, due to the inherent disadvantages of reciprocating
machinery. This led to the introduction of
rotating machinery to produce the steadily increasing
needs for electricity; already well known in the harnessing
of water energy the principle of blading was
adopted in the steam and at a later date in the gas
turbines.
I. Theory of Steam Turbines
The steam turbine obtains its motive power from
the change of momentum of a jet of steam flowing over
a curved vane. The steam jet, in moving over the
curved surface of the blade, exerts a pressure on the
blade owing to its centrifugal force. This centrifugal
pressure is exerted normal to the blade surface and
acts along the whole length of the blade. This fundamental
is shown diagrammatically in Fig. 1. The
resultant of these centrifugal pressures, plus the effect
of change of velocity, is the motive force on the
blade.
This introductory paragraph regarding the transfer
of energy from the fluid to the blade can also be
shown mathematically.
FIGURE 1: Pressure on blade
By W. B. JACHENS
Fig. 2 shows the general view of a rotor mounted
on a shaft, Figs. 2a and 2b the side and end elevations.
V1 is the absolute uniform velocity of the fluid entering
the rotor passage, while V2 is the absolute uniform
velocity leaving the rotor passage. Components of V1
and V.2 are conveniently considered in 3 directions,
i.e. axial, radial and tangential.
This leads to the so-called Euler equation for
turbo machinery.
Equation 2 may be transformed in terms of blade
absolute and relative velocities determined from the
physical conditions, i.e. flow area, flow volume and
blade angles. We shall consider a 2-dimensional flow
as occurs in practically all turbo machinery.
From Fig. 3-
FIGURE 3: 2-dimensional flow
113
A similar expression can be derived for the rotor
blade exit.

114
FIGURE 2(a): Side elevation
Substituting in Equation 2:
This ultimate form of the fundamental equation is
broken down into 3 components:
This represents the absolute kinetic energy change
in the fluid as it passes through the rotor (velocity
head).
This represents the change in static head due to the
centrifugal effect (axial flow U1 = U2).
Change in static head due to diffusion or expansion
process in the flow passages, i.e. area increasing in the
direction of flow, therefore, relative velocity decreases
and static pressure increases.
Types of Turbines
There are two types of steam turbines; impulse and
reaction. There is a distinct difference in the working
of these two types, and manufacturers of steam
turbines usually specialise in the production of one
Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 2: Rotor mounted
on shaft
FIGURE 2(b): End Elevation
of these types only. The main distinction is the manner
in which the steam is expanded as it passes through
the turbine. In the impulse turbine, the steam is
expanded in nozzles and remains at constant pressure
when passing over the blades. In the reaction turbine,
the steam is continually expanding as it flows over the
blades.
The original steam turbine, the De Laval, was an
impulse turbine having a single-blade wheel. Other
impulse turbines are known as Curtis, Zoelly and
Rateau.
The reaction turbine was invented by Sir Charles
Parsons and is known as the Parsons turbine.
In all turbines the blade velocity is proportional
to the steam velocity passing over the blade. If the
steam is expanded from the boiler pressure to the
exhaust pressure in a single stage, its velocity is
extremely high. If this high velocity is used up on a
single-blade ring, it produces a rotor speed of about
30,000 r.p.m. which is too high for practical purposes.
There are several methods of overcoming this high
rotor speed, all of which utilise several blade rings.
The following are the four principal methods used:
(a) Compounding for velocity. Rings of moving
blades separated by rings of fixed blades, are
keyed in series on the turbine shaft, Fig. 4.
The steam is expanded through nozzles from
the boiler — to the back-pressure, to a high
velocity, and is then passd over the first ring

Proceedings of The South African Sugar Technologists' Association—March 1966
of moving blades. Only a portion of the high
velocity is absorbed, the remainder being
exhausted on to the next ring of fixed blades,
which change the steam direction without
appreciably altering the velocity. The jet then
passes on to the next ring of moving blades,
the process repeating itself until practically all
the velocity of the jet has been absorbed. It
will be noticed that, due to the pressure remaining
constant as the steam passes over the
blades, the turbine is an impulse turbine.
FIGURE 4: Compounding for velocity
(b) Compounding for pressure. In this type, the
total pressure drop of the steam does not take
place in the first nozzle ring, but is divided up
between all the nozzle rings. The steam from
the boiler is passed through the first nozzle ring
in which it is only partially expanded. It then
passes over the first moving blade ring where
nearly all of its velocity is absorbed. From this
ring it exhausts into the next nozzle ring and is
again partially expanded; this absorbs a further
FIGURE 5: Compounding for pressure
115
portion of its total pressure drop. It then passes
over the second ring of moving blades, the
process thereby repeating itself. As the pressure
remains constant during the flow over the
moving blades, the turbine is an impulse
turbine. This method of pressure compounding
is used in Rateau and Zoelly turbines.
(c) Pressure-Velocity Compounding. In this type
of turbine, both of the previous two methods
are utilised. This has the advantage of allowing
a bigger pressure drop in each stage and,
consequently, less stages are necessary, resulting
in a shorter turbine for a given pressure
drop. It may be seen that the pressure is constant
during each stage; the turbine is, therefore,
an impulse turbine. The method of
pressure-velocity compounding is used in the
Curtis turbine.
(d) Reaction turbine. In this type there is no
sudden pressure drop; the pressure drop is
gradual and takes place continuously over the
moving and fixed blades. The fixed blades
correspond to nozzles; they change the direction
of the steam and, at the same lime, allow it to

116 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 6: Pressure-Velocity Compounding
expand to a higher velocity. The pressure of the
steam falls as it passes over the moving blades;
the turbine is, therefore, a reaction turbine.
The Thermo-dynamics of the Steam Turbine Elements
Nozzles:
The steam nozzle is a passage of varying crosssection
by means of which the heat energy of steam is
converted into kinetic energy. The nozzle is so shaped
that it will perform this conversion of energy with the
minimum loss. The flow of steam through a nozzle
may be regarded, in its simplest form, as being an
adiabatic expansion. The steam enters the nozzle with
a relatively small velocity and a high initial pressure;
the initial velocity is so small compared with the final
velocity that it may be neglected.
Let Is1 = total heat of steam entering nozzle.
Is2 == total heat of steam at any section
considered,
v = velocity of steam at section considered
in ft. per sec.
u = heat drop during expansion.
= Is1 - Is2
Then, assuming a frictionless adiabatic flow and
considering 1 lb. of steam,
Gain of kinetic energy = heat drop.
FIGURE 7: Reaction Turbine
In practice, there is a loss in the nozzle, due to
friction, of about 10 to 15% of the total heat drop;
the effect of this is to reduce the value of u in equation
4.
The effect of the friction of the nozzle is to reduce
the velocity of the steam, and to increase its final
dryness or super heat.
Weight of Discharge through Nozzle:
The adiabatic flow of the steam through the nozzle
may be approximately represented by the equation
pv n = constant
where n = 1.135 for saturated steam.
= 1.3 for super-heated steam.
Now, the work done during the cycle will be given
by

Proceedings of The South African Sugar Technologists' Association—March 1966 117
Then, gain of kinetic energy = work done during
the cycle
Substituting this value in 6:
From Equation 7:
Substituting the values of v 2 and V in 9:
The equation 10 can be used to obtain the flow of
steam through the turbine because:
A = cross-sectional area of nozzle.
n = either 1.135 or 1.3.
p1 = initial or live steam pressure.
V 1
= initial or live steam volume per lb.
p 2 = discharge or wheel-chamber pressure.
v 2 = discharge or wheel-chamber volume per lb.
Further relationships as regards nozzles are obtainable,
i.e. throat pressure for maximum discharge,
but these are nozzle details with which we are not
directly concerned.
Theory of Blading:
The most important turbine elements are the
blades. The following thermo-dynamic approach is
meant to briefly show the different efficiencies obtainable
with certain blade combinations, and the importance
of the so-called speed ratio.
(a) The Simple Impulse Stage:
The velocity diagram indicates, as shown in
Fig. 8, the velocities involved.
FIGURE 8: Impulse stage velocity diagram
Notations used:
= nozzle jet velocity.
= relative steam velocity at inlet.
= absolute steam velocity at exit.
= relative steam velocity at exit.
= mean peripheral velocity of blades.
= nozzle jet angle.
= inlet angle of blade.
= outlet angle of blade.
= absolute angle of steam leaving blade.
The useful tangential propelling force in the direction
of the blade motion:
Expression 11 can be expressed in function of
thus substituting ρ and v w in Expression 11:
To obtain the efficiency of the blading the energy
E (see Equation 2) may be divided by the kinetic energy
of the jet issuing from the nozzle.

118
Substituting the value of v w obtained above:
i.e. for a simple impulse stage or turbine having one
impulse row of blades, the maximum blade efficiency
can be determined at a corresponding speed ratio.
Impulse Turbine Staging or Compounding:
The simple impulse turbine is limited in its application
when a large pressure drop is necessary, because
the high nozzle velocities resulting implies high blade
speed with related blade and disc stress problems.
Four already discussed methods are available to
cope with large pressure drops at reasonable blade
speeds. These are:
(a) Compounding for velocity.
(b) Compounding for pressure.
(c) Pressure-velocity compounding.
(d) Reaction turbine.
As determined above for the simple impulse stage,
ρ opt. can also be theoretically determined for the
blade configuration mentioned under (a) to (d). As the
principle is the same, only the results shall be stated.
The results are based on the assumption of frictionless
flow.
A comparison of the above results shall help to
explain certain features of a definite design.
Comparing velocity and pressure staging, i.e. (a)
and (b) shows the following:
Proceedings of The South African Sugar Technologist*' Association—March 1966
Velocity Compound:
To reduce u to 1 we require m stages,
mu
Efficiency lower than pressure staging because of
higher residual kinetic energy and friction losses.
Advantage of taking large pressure drop in one
step and thus reducing the pressure and temperature
at which the blades run.
Cheaper (shorter) construction.
Pressure Compound:
To reduce u to 1 we require m 2
stages,
mu
Lack of simplicity and construction but advantage
of uniform distribution of work among stages, K.E.
loss = 1 of velocity compound turbine,
m
More expensive construction, thus higher initial cost,
due to longer turbine (more stages required).
To correct the defects of the Curtis and add the
qualities of the Rateau, (c) may be used, but in
practice is very seldom found due to the complicated
design and a relatively low efficiency.
A more common arrangement is to provide on the
high pressure side one or more Curtis stages, followed
by Rateau or Reaction staging. The Curtis stages
reduce the pressure and temperature of the fluid to a
FIGURE 9: Impulse-Reaction Comparison

Proceedings of The South African Sugar Technologists' Association—March 1966
moderate level with a high proportion of work per
stage, and then the more efficient Rateau or Reaction
staging absorb the balance of the energy available.
Comparing frictionless conditions, the maximum
efficiencies for the simple impulse, Curtis and Reaction
blading are equal; however, when friction is taken
into account, the reaction stage is found to be the
most efficient, followed by Rateau and Curtis in that
order. The reason that friction losses are less significant
in the reaction stage lies in the fact that the flow
velocities are lower.
Fig. 9(a) shows how F, varies with p for both an
impulse and reaction turbine. Both types develop
the same force at p = 0, the force dropping to zero at
twice p for the reaction turbine.
Fig. 9(b) compares blade efficiency for impulse and
reaction turbines for the given conditions. Maximum
efficiency developed at slightly less than ρ = 1 for the
reaction turbine, just twice that for the impulse
turbine.
Developed forces for both turbines are equal at
their maximum blade efficiencies.
II. Selection of Steam Turbines
In the section, Selection of Steam Turbines, the
different types of steam turbine shall be discussed on
a more practical level so as to be of use to the works
and planning engineer.
The Differences between the two most widely used
Industrial Steam Turbine types
The two most widely used steam turbine types are
the reaction turbine and the pressure-compounded or
Rateau turbine. The high pressure part for both types
consists of a simple or two-stage velocity-wheel,
depending on the live steam conditions at the turbine
entry. The impulse stage, known as the governing
stage, is generally applicable since it alone permits
partial steam admission to the moving wheel.
Both turbine types have inherent features which
may be termed an advantage or disadvantage, as the
case may be.
Characteristics of the Two Types
Rateau
The rotor is built as a disc rotor with a slender
shaft. Between the individual discs, inner labyrinths
or seals are arranged on a shaft of small diameter.
Due to the slender rotor construction, the lack of
rigidity causes the critical speed or speeds of the rotor
to be below the rated speed. In starting up or stopping,
the rotor must pass through the critical speed ranges.
Although, when running up to speed the critical speed
ranges can be rapidly passed by quick opening of the
throttle valves, there is no way to accelerate the rate
of speed decrease when shutting down. The measure
against excessive vibration at the critical speed is to
provide adequate clearance between diaphragm and
shaft.
The stationary blades and the shaft seals are inserted
into half split diaphragms which are, in turn, attached
to the casing.
Stationary and moving blades have different
profiles. They impose a considerable change in
direction of the steam flow.
Since the steam, as it passes through the turbine,
is subjected to a heavy rotational motion around the
rotor, the steam path at the root and tip of the blades
is not uniformly filled with steam.
The friction heat generated at the shaft seals, due
to rubbing, is carried off by the leakage steam only.
Reaction
The rotor is of the so-called solid type. The pressure
difference between the inlet side and outlet side of the
stationary blade is half that of an impulse stage. The
sealing thus presents less difficulty.
The rotor of the reaction turbine is rigid, the critical
speeds lying above the rated speed. Starting and
shutdown require no special precautionary measures
against excessive vibrations.
The stationary blades are directly inserted into the
casing.
Stationary and moving blades have similar profiles.
They impose a smaller change in the direction of the
steam flow.
The steam path is uniformly filled with steam, since
the steam particles pass mainly in axial direction
through the turbine, the rotational motion around the
rotor being less.
The friction heat generated at the seals is carried
off by the whole working steam quantity.
A few of the above differences between the two
types of turbine should not propagate any one design,
but should point out inherent differences in construction
of both types.
Steam Turbine Selection
In the previous sub-section, the internal characteristics
of the two most widely applied turbine types or
turbine constructions, were briefly discussed. In this
sub-section, the type of turbine refers not to the
internal construction, but to the turbine as a whole
unit.
Four basic types of turbine are available:
119
1. Back-pressure turbines expand the live steam
supplied by the boiler to the pressure at which
the steam is required for the process. The overall
plant efficiency of a back-pressure turbine
exhausting to a process is high, due to the
considerable heat losses through the condenser
being eliminated. The electric power generated
by the back-pressure turbine is directly proportional
to the amount of process steam required.
To avoid the direct relationship between backpressure
steam and power, the alternator would
have to be connected to the grid, or a by-pass
valve installed.

120
2. Extraction back-pressure turbines are employed
where the process steam is used, at two different
pressure levels. The operating of this type of
turbine is similar to the back-pressure machine,
a high overall plant efficiency being obtainable.
Difficulties exist in the control of such a turbine
type, two by-pass valves being necessary to
maintain the process pressures if the alternator
is not coupled to the grid.
3. The condensing turbine is used where process
steam is not necessary. This turbine has a lower
overall efficiency due to the loss of heat in the
steam condenser, and it is difficult to obtain
cheaper station cost per kilowatt-hour than
buying electric power, due to the inherent disadvantages
of a relatively small condensing
turbo-alternator set.
4. The extraction-condensing turbine is used where
the power required is in excess of the process
steam. The efficiency of this type is lower than a
back-pressure turbine due to the partial loss of
heat in the condenser. However, this turbine has
the advantage that the power generated is not
proportional to the process steam required and,
furthermore, power can be obtained whilst the
process in the factory is shut down.
Using this short introduction, the next logical step
is to describe the different turbine types in a thermodynamic
language using the Mollier diagram, this
showing the relationship between the enthalpy drop of
steam and the power obtainable. The turbine shall
also be integrated into a steam cycle, showing the
employment of reduction valves and related plant
material.
Proceedings of The South African Sugar Technologists' Association—March 1966
1. Back-Pressure Turbine
Fig. 10(a) shows the back-pressure turbine incorporated
into the steam cycle. Points 1 and 2 signify
the live steam and the back-pressure steam condition
respectively. Due to various losses in the turbine, the
actual work done h u will be less than the isentropic
or theoretical work h s. The relationship is called the
internal efficiency and may be written thus:
The heat equivalent of the actual work per lb. of
steam may be calculated as follows, knowing that 1
kW-h = 3,416 B.T.U.
For a turbine installation, h s and η i can for any
particular load be considered constant, thus N is
proportional to G, the quantity of process steam used.
Knowing G and assuming η i it is, therefore, possible
to calculate the power obtainable for a specific live
steam condition. If N obtainable is lower than N
required, the balance can be obtained by coupling the
alternator to the grid; vice versa, the excess can be
supplied to the grid.
When coupled to the grid, the frequency, i.e. the
turbine speed, is maintained by the grid, the backpressure
being maintained by the governor which
opens or closes the nozzle valves in the live steam
main admitting more or less steam. The steam quantity
is thus always adjusted to the requirements of the
FIGURE 10(a): Steam cycle FIGURE 10(b): Mollier diagram
for the back-pressure turbine

Proceedings of The South African Sugar Technologists' Association—March 1966
process, the power being balanced due to the grid
connection.
In most cases, the turbine is not coupled to the
grid, in which case, the governor has to maintain the
speed, i.e. frequency of the turbine, at a constant level
by adjusting the steam quantity to the power requirements.
Here, the opposite is true, the power being the
dominating factor, in which case the pressure of the
back-pressure steam is maintained at a constant value
by installing a by-pass and a blow-off valve.
2. Extraction Back-pressure Turbine
FIGURE 11(a): Steam cycle
FIGURE 11(b): Mollier diagram
for the extraction back-pressure turbine
Fig. 11(a) shows an extraction back-pressure turbine,
Fig. 11(b) the steam expansion in the Mollier
diagram. Let η i1 and η i2 be the internal efficiencies
ot the high and low-pressure parts respectively, and
G 1 and G 2 the steam flows; then
Generally, the amount of steam to be supplied to
the process is known, let these amounts be termed
E 1 and E 2.
Substituting in Equation 15:
The above equation shows that, for definite
pressure drops and efficiencies, the amount of power
obtainable is proportional to the process steam
required.
FIGURE 12: Steam consumption diagram
121
The relationship between the power obtainable
and the steam quantities flowing, is shown in Fig. 12.
This is a strongly simplified diagram, but may be used
to explain the principles involved. Take, for instance,
point 1. This point indicates that maximum power is
obtainable by passing the steam quantity G 1 through
the high-pressure part, and the quantity G 1- E 1
through the low-pressure part. At point 2, the maximum
power is still obtainable, the high-pressure steam
flow dropping to G 2 , the low-pressure steam flow
increasing to (G 2 - 3/4 E t ); the reduction in live steam
is, therefore, balanced by the increased steam quantity
flowing through the low-pressure part.
The process pressures are maintained, when the
alternator is coupled to the grid, by the turbine
governor receiving impulses from both the extraction
and the back-pressure line. When the alternator is not

122 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 13(a): Steam cycle FIGURE 13(b): Mollier diagram
for the condensing turbine
coupled to the grid, the turbine governor has to maintain
a constant turbine speed, the extraction and
back-pressures being maintained at the desired values
by a system of by-pass and blow-off valves.
3. Condensing Turbine
Condensing turbines are seldom used in industry
due to the lower efficiency as compared with large
power stations. Large power stations use an elaborate
system of regenerative feed heating which is not
possible with a small industrial condensing steam
turbine. The steam is expanded to condenser pressure,
the pressure of the exhaust steam being dependent on
several factors, as shown in the basic equations given
below:
Cooling surface A =
Q
loge ts-ti ... 16
K ts-t0
Where A = cooling surface in ft 2 .
Q = cooling water quantity in lbs./hr.
ts = steam exhaust temperature in ° F.
ti = cooling water inlet temperature in
°F.
to = cooling water outlet temperature in
°F.
K = heat transfer rate in B.T.U./ft. 2 °F.hr.
Outlet cooling water temperature
to = t, + WH . . . 17
Q
Where W = Steam flow into condenser in lb/hr.
H = Heat rejected by steam in B.T.U./lb.
Generally, in calculating the condenser pressure,
i.e. the value of H, the cooling water inlet temperature
t1 is known; the relationship of Q may be assumed to
W
be approximately 70 and. to approximately 10° F
greater than the inlet temperature of the cooling water.
These results would give a value of H and, knowing
the approximate expansion line of the steam, the
exhaust point can be plotted on the Mollier diagram.
With a certain amount of trial and error, the correct
expansion end point is obtained. The condenser cooling
surface can then be calculated, the value of K
obtainable from suitable heat transfer curves.
To calculate the power N, the same basic formula
as derived for the back-pressure turbine can be used:
As against back-pressure turbines where the enthalpy
drop remains constant at all loads, due to the
back-pressure being regulated, the enthalpy drop for
a condensing turbine changes with load. This may be
explained by considering Formula 17. Assume to, ti,
and Q remain substantially constant, then, as W
decreases with decreasing load, H would increase, i.e.
the condenser pressure rises as the load decreases.
The regulation of condensing turbines presents no
problem. If coupled to the grid, the governor can be
used to keep the live steam pressure constant; if
operated independently, the governor maintains the
system frequency.
4. Extraction-condensing Turbine
The extraction-condensing turbine is a combination
of the back-pressure and condensing types, the formulas
obtained for the latter types may be employed
on the extraction-condensing type.

Proceedings of The South African Sugar Technologists' Association—March 1966 123
FIGURE 14(a): Steam cycle FIGURE 14(b): Mollier diagram
The extraction pressure is fixed by the pressure of
the process steam required, the exhaust pressure
determined as for the condensing turbine.
The extraction-condensing turbine is widely used,
the principal advantage being that due to the condensing
part, the power obtainable is not proportional
to the process steam quantity required. In a factory
where a considerable secondary load exists, or where
power is required when the process part of the factory
is shut down, the condensing part of the extractioncondensing
turbine is able to supply the load. A
typical example is a factory with a large irrigation
load, the latter being secondary. The irrigation load
generally has to be supplied throughout the year.
The relationship between power and steam is
basically identical to that shown in Fig. 12 and shall
not be repeated. The regulation would also be a
repetition of previous explanations and need not be
commented on.
Turbine Reduction Gears
The turbine ratings for industrial power plants lie
between 1,000 and 10,000 kW, the lower and upper
limit depending on the steam conditions. In the
theoretical part of this paper, it has been mentioned
that the large pressure drop in a single-blade ring
would produce a rotor speed of about 30,000 r.p.m.
To decrease this speed, compounding, i.e. employing
a larger number of stages, is necessary.
for the extraction condensing turbine
The universal speed at which large two-pole alternators
operate is 3,000 r.p.m. For industrial turbines, an
economic solution has been found in an intermediate
speed, generally ranging between 5,000 and 12,000
r.p.m., the latter speed being for turbines of small
output. The reasons are obvious, but shall be enumerated
as follows:
1. High speeds with the same stage-heat drops
result in longer blades of smaller diameter.
2. The flow ducts have favourable dimensions.
3. The peripheral losses are reduced.
FIGURE 15: Helical spur reduction gear

124 Proceedings of The South African Sugar Technologists' Association—March 1966
4. The peripheral clearance is small and, therefore,
the clearance losses are correspondingly low.
5. The use of lower speed alternators, generally of
1,500 r.p.m., which are cheaper than the two-pole
type.
The reduction gears are generally of the doublehelical
spur gear type. The turbine and the reduction
gear are connected by a flexible coupling, the reduction
gear and the generator by a rigid coupling. Planetary
gearing is also used; the advantage of co-axial shafts
must be weighed against the complications introduced
in connection with inspection work.
Turbine Standardisation
No manufacturer who hopes to offer an economically
competitive and, at the same time, reliable
turbine, can afford to disregard the value of standardisation.
For all types of turbines, a definite range of
standard sizes are designed. Each standard size
represents a casing size, the size or standard casing
used depends on the steam conditions and the quantity
of steam flowing. The power output, for instance,
may vary considerably for any one standard casing
size.
The advantages are obvious; due to the completed
designs and available drawings, the delivery time
may be reduced considerably. The standard parts
FIGURE 16(a): Turbine cross-section
have proved themselves, so that a higher reliability
may be expected. The problem of obtaining spare
parts at short notice is non-existent.
Taking a standard turbine, it is possibly to classify
the turbine parts into three groups:
1. Parts which have to be adapted to the thermodynamic
conditions:
nozzle valves; nozzles and blades of the impulse
wheel; and reaction or Rateau blading.
2. Parts which are mass produced and may be used
on several standard sizes:
operating cylinder for valve gear; labyrinthglands;
thrust bearing; bearing pedestal; journal
bearings; main oil pump, shaft driven; emergency
stop valve; steam strainer; governing
system; and long end-blades for condensing
turbines.
3. Parts which are only usable for the standard size
considered:
turbine casing with integral steam inlet chamber;
emergency stop valve casing incorporating the
live steam inlet flange; exhaust casing incorporating
the exhaust flange; turbine rotor; and
coupling on alternator or reduction gear end.
The above parts list is by no means complete but it
indicates the method of producing a standard turbine.

Proceedings of The South African Sugar Technologists' Association—March 1966 125
A: Servo-motor F: Emergency stop-valve G: Steam strainer
FIGURE 16: Standardisation of a condensing turbine
FIGURE 16(b): Turbine longitudinal cross-section

126
Proceedings of The South African Sugar Technologists' Association—March 1966
B: Labyrinth glands D: Bearing pedestal I: End blades
C: Thrust bearing E: Oil-pump casing H: Journal bearing
III. Steam Turbine Operation
As in all sections, the wide field of steam turbines
has been limited to the most essential features in order
to grasp certain basic features and ideas which may
lead the way to a more intensive study, if necessary.
The same principle has been adhered to in this
section.
Turbine Performance at varying loads
The performance of a turbine when operating at
loads different from the designed or economic load,
depends on the particular method employed for
controlling the supply of steam to the turbine, so that
the speed of rotation will remain sensibly constant,
irrespective of the load. The principal methods of
governing are:
1. Throttle governing.
2. Nozzle governing.
3. By-pass governing.
4. Combination of 2 and 3.
1. Throttle Governing
The primary aim is to reduce the mass rate of flow.
[unreadable] reducing the mass rate of flow, the steam
[unreadable] increasing pressure drop across the
[unreadable] and, consequently, a throttling or
constant enthalpy process, with an increasing entropy
and a corresponding decrease in available energy.
This condition is shown in Fig. 17:
FIGURE 17: Throttling governing
The relationship between load and steam consum¬
tion for a turbine governed by throttling is given by
the well known "Willans Line", which is a straight

Proceedings of The South African Sugar Technologists' Association—March 1966
line between no load and most economic load, as shown in Fig. 18:
Equation of a straight line G = Go + m.K. The
specific steam consumption is found by dividing by
K:
2. Nozzle Governing
Ideal governing would be obtained if all nozzles in
each and every stage in a turbine could be controlled.
The Willians Line for such a turbine would be a
FIGURE 18: Willans line diagram
straight line, as indicated for the throttle governing,
however, with a considerably better specific steam
consumption at part loads. In an actual turbine,
nozzle governing must be restricted to the first stage
nozzles for construction reasons, and even here,
groups of nozzles are governed, rather than each
nozzle. With nozzle governing, the pressure and
temperature entering the first stage nozzles are
constant with load. Fig. 19 and Fig. 20 indicate the
better specific steam consumption obtainable with
nozzle governing.
FIGURE 19: Steam flow FIGURE 20: steam consumption
127

128 Proceedings of The South African Sugar Technologists' Association—March 1966
Fig. 21 indicates the lines of expansion:
FIGURE 21: Steam expansion for nozzle governing
It is important to note that the absolute pressure
of the steam entering the second stage nozzles is in
direct proportion to the mass flow through the turbine.
The significant feature of nozzle governing is that
considerably less throttling of steam occurs than if a
single valve were used. Fig. 22 shows a typical
arrangement of nozzle valves:
FIGURE 22: Nozzle governing with by-pass
3. By-Pass Governing
This is generally used for the overload valve which
passes the steam directly to the steam chest, thereby
by-passing the impulse wheel. The normal amount of
nozzle valves up to economical load are used.
4. The combination of 2 and 3 has already been
outlined under 3 and may also be seen in Fig. 22.
Lubrication
The problem of lubrication resolves itself into three
parts which may be enumerated thus:
1. Maintenance of a sufficient supply of oil to the
bearings.
2. Cooling of oil to remove the heat generated in the
bearings and limit the temperature reached by the
oil.
3. Maintenance of the physical properties of the oil
at a sufficiently high standard to ensure safe and
satisfactory lubrication.
Fig. 23 indicates a typical lubrication system for a
turbo-generator.
In all cases, the main oil pump, which is driven
directly from an extension to the turbine shaft,
performs a dual function; it supplies oil both for
lubrication of the bearings and for operating the
governor relays. For the latter, the oil is uncooled.
The pressure in the relay system is maintained by valve
5, and that in the lubrication system by valve 11. The
auxiliary oil pump is arranged to supply the full
quantity of oil and to start automatically when the
pressure from the main oil pump falls below a predetermined
value. A further flushing oil pump is
provided for flooding the bearings when starting and
stopping. The oil purifier equipment is generally of
the centrifugal type and is operated continuously
FIGURE 23: Turbine lubrication system
1. Drain from bearings
2. Oil tank
3. Oil strainer
4. Main oil pump
5. Pressure sustaining
valve
6. Oil to relay
7. Oil coolers
8. Oil to bearings
9. Spring-loaded by-pass
valve
10. Fine strainer
I I. L.P. relief valve
12. Oil purifier
13. Auxiliary oil pump
14. Flushing oil pump
15. Priming connexion

Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 24(a): Linkage un-loading FIGURE 24(b): Linkage un-loaded
while the turbine is running, and should have a capacity
not less than one-tenth of the total quantity of oil in
the turbine lubrication system.
Governors and Governor Gear
As a flow-pressure type machine operating at high
speed, the steam turbine is fitted with a rotor having a
very low moment of inertia. This means that its speed
responds immediately to any changes in load. The
turbine governor whose duty it is to maintain constant
speed must, therefore, be sensitive and operate very
rapidly and reliably.
Fig. 24(a) shows a ball-type governor acting
through means of a pilot valve and servomotor on the
steam valve. When the load decreases, the speed of
the machine increases; the flyballs lift the sleeve M
to M l . Point K on the connecting rod linking the
power piston remains at rest, forming the fulcrum
FIGURE 25: Static regulation of speed governor
129
of the motion which forces the pilot valve into position
S. 1 The pilot valve now allows high-pressure oil to
flow above the power piston, which moves downwards
and closes the regulating valve.
Fig. 24(b) shows the position of the governor
linkage on completion of unloading. Since M 1 is now
the fulcrum, the pilot valve returns to its original
position as K moves to K 1 , thereby shutting off the
high-pressure oil.
The difference in speed between no load and full
load referred to the rated speed, is designated as the
proportional range or static regulation of the governor.
It is determined by the characteristic of the speeder
spring.
A turbine with a normal speed of 3,000 r.p.m.,
which operates as an isolated unit independent of
other prime movers, and which is fitted with a
governor having 4% static regulation runs, for in¬

130
stance, at 3,060 r.p.m. at no load and 2,940 r.p.m. at
full load in accordance with the continuous line in
Fig. 25. If the governor sleeve is additionally loaded,
e.g. by further compression of the speeder spring, the
flyweights must produce a greater force in order to
maintain the sleeve in its position. This can take place
only with rising speed, as shown by the dotted line in
Fig. 25.
Whereas, with an isolated unit, operation of the
speed adjusting device changes the speed at constant
Many types of governing systems have been
employed. A more modern development is the
hydraulic speed governor which eliminates the
centrifugal balls and their associated linkages. This
type is shown in Fig. 27:
FIGURE 27: Hydraulic speed governor
Proceedings of The South African Sugar Technologists' Association- March 1966
FIGURE 26: Loading of the alternator
load, with the alternator coupled to the grid, it is the
load that changes, the speed, remaining constant.
Assuming 4% steady state regulation, on adjustment
of the speed changer by an amount which would
correspond to a speed increase of 60 r.p.m. (2%)
with an. isolated unit, the machine which operates at
point C, see Fig. 26, with 25 % of full load, now takes
on. additional load and operates at point C 1 with 75%
of full load.
The hydraulic governor utilises as regulating
impulses the pressure difference produced by a small
oil gyro upon a change in the turbine speed. The main
oil pump itself is not used as impulse transmitter
because of the non-uniform high pressure oil requirements
of the governing system.
Supervision and Instrumentation
Thermometers for local indication are fitted to all
bearings and to the oil coolers. If necessary, a contact
thermometer for alarm signals can be fitted behind
the oil coolers. All water inlets and outlets at the oil
coolers and condensers are also fitted with thermometers.
A thermometer at the steam admission section
serves to measure the live steam temperature. The
pressure gauges for steam pressures ahead of and
within the turbine, as well as vacuum gauges for the
condensers, pressure gauges for pressure oil, trip oil,
governor oil and bearing oil, are conveniently arranged
in a pressure gauge pillar. Such a pillar for nine
pressure gauges in shown in Fig. 28:
Measuring devices for monitoring the axial shaft
position are fitted to the bearing pedestals. The
measuring device at the front bearing pedestal serves
to supervise the thrust bearing in order to detect any
wear in good time. The measuring device at the rear
bearing pedestal indicates the axial differential expansion
of shaft and casing. Both measuring devices
are provided with vernier scales for adequate
measuring accuracy.
A condensate drain controller maintains the
condensate at a constant level. This can be read off
on a level gauge.

Proceedings of The South African Sugar Technologists' Association—March 1966 131
FIGURE 28: Pressure gauge console
Turbine Troubles
The causes of failure naturally divide themselves
into two groups, namely, those inherent in the design
or in the material used in the construction of the
turbine, and those which are related to the operating
conditions. Some of the former are inter-related to
some of the latter; for instance, blade erosion might
be due to unsuitable material or to bad steam
conditions.
The causes of failure inherent in design and
materials of construction may be classified as follows:
(a) Shaft vibration.
(b) Disc vibration.
(c) Blade vibration.
(c) Faults in machining.
(e) Incorrect design of casing, faulty arrangement
of steam pipes, causing distortion of the casing.
(f) Materials of construction.
To discuss each of the above points would be far
beyond the scope of this paper.
IV. Summary
The installation of the industrial turbine in a
factory where both process steam and electrical
power are required in the right proportions, can only
be of economical advantage.
References
Lewitt, E. H. (1959). Thermodynamics applied to heat engines.
Kearton, W. J. (1964). Steam Turbine Operation.

132
STEAM TRAPPING AND CONDENSATE CONSERVATION
Savings
For efficient heat economy it is important not to
forget the value of steam after it has done its work, in
other words, look after the condensate. Its heat value
alone is important.
Consider a factory producing 200,000 lbs. of steam
per hour. If the boilerfeed make-up is 40 per cent due
to loss of condensate then 80,000 lbs. of cold water
at 65° F. must be added to the hot well, and heated
to the normal temperature of return condensate which
may be 200° F.
Heat to be supplied = 80,000 (200— 65) Btus.
11,370 lbs. steam per hour.
If coal is being used in the factory this is equivalent
to 1,420 lbs. per hour of coal at R5.00 per ton for
144 hours per week = R511.
In a 40 week season this amounts to R20,440 per
season, so it behoves us to look carefully at our condensate
collection systems. With care the make-up
can be cut to 10 per cent and under, saving R15,330
per annum in fuel bills alone, plus the saving in not
having to treat the equivalent amount of make-up
water which in the above example would be a further
saving of R6,880 if we assess the cost of pumping,
clarifying, softening and conditioning feed water
make-up at 20 cents per 1,000 gallons.
So, in our hypothetical factory producing 200,000
lbs. of steam per hour and using supplementary coal,
there is a potential saving of R22,210 per season if
we cut make-up from 40 per cent to 10 per cent.
For any factory using coal this works out at R85
per week for every 1,000 gals. per hour of make-up
water saved; worth putting one member of the staff
full time on condensate collection.
Steam Traps
For the efficient removal of condensate from steam
users and distribution we have firstly to choose the
right trap for the job. There is a bewildering selection
of traps to choose from as the following list shows.
Group 1. Mechanical Traps
Examples of these are:
1. Plain float traps.
2. Trip action float traps.
3. Open bucket traps.
4. Inverted bucket traps.
5. Relay operated, or steam assisted traps.
Proceedings of The South African Sugar Technologists' Association- March 1966
By J. M. CARGILL
Group 2. Thermostatic Traps
Examples of these are:
1. Metallic expansion traps.
2. Liquid expansion traps.
3. Balanced pressure traps.
Group J. Thermo-Dynamic Trap
Examples of these are:
1. Impulse traps.
2. Thermo-dynamic disc type traps.
3. Multi-stage nozzle type traps.
No consideration will be given to "U" lubes as
steam and water separators as its use is confined to
very low pressures. It cannot vent vapours without
losing its seal, and its use is fraught with inconsistency.
If one imagines hot condensate rising in the outlet
leg of a "U" tube one can readily see the danger of
flashing occurring and the probable loss of the "U"
seal due to the lack of head exerted by the outlet leg.
It can sometimes be used with success but its use
should be confined to cases where venting is no
problem, and the loss of seal is of little consequence.
Generally speaking, its use should be avoided.
There is no room in this paper to go into the pros
and cons of each type of trap; the technicalities of
trapping have been covered in standard books of
reference, Oliver Lyle's "Efficient Use of Steam" being
one of the best.
Condensate Collection
It may not always be economic to collect all condensates
from steam traps. We should only consider
doing so if:
(a) There is enough of it to bother about — a
highly superheated steam main will give very
little condensate under normal operation yet
still require trapping.
(b) We want the water for further use, such as
boiler feed, mill maceration, process water,
and/or
(c) We want the sensible heat out of the condensate.
Generally speaking, in sugar mills it is hardly
worthwhile collecting water from the steam traps of
steam mains scattered all over the factory. Goodness
knows we have enough pipes, a goodly number of
them extraneous as it is, around the place without

Proceedings of The South African Sugar Technologists' Association — March 1966 133
cluttering it worse with a host of small bore condensate
returns. The money would be better spent on a bit
of lagging here and there.
Boiler Feed
Let us consider first of all the use of condensate as
boiler feed. If it is "first generation" condensate, i.e.
direct from H.P. steam and not a condensed juice
vapour, it will normally be very good boiler feed. A
bit of deaeration, some chemical conditioning and it
is ready to go back to the boiler. "Second and third
generation" condensates are not so easy. They are
normally contaminated from the process. They do not
make as good a boiler feed, or at the least, they are
more risky, and do not lend themselves to condensate
contamination control by simple conductivity controllers.
So, from a boiler feed point of view we
should concentrate on the "first generation" condensates.
The sources available are many, depending on the
plant arrangement. The obvious ones are condensing
turbines, first effects of evaporators, vapour cells,
juice heaters, pans — the latter process equipment is
often fed by direct exhaust steam.
Ignoring turbine condensate as its use is so obvious,
the ideal set-up is to condense all exhaust steam as it
is generated in one step. The advantage of the use of
one large, or a series of parallel vapour cells, or preevaporators,
to fulfil this function is outstanding. This
matter will be, or has been, discussed at this symposium
and will not be gone into detail here.
If it is necessary to use second and third generation
condensates for boiler feed, choose the units where
the vapour pressure is greater than the juice pressure,
such as the vessels of a quad rather than a juice heater.
The risk of contamination from a burst tube is less
in a quad than in a juice heater.
The essential point is to collect sufficient hot condensate
to satisfy the boiler requirements, and to
reduce the risk of contamination to a minimum.
Process Use and Heat Recovery
Any surplus to boiler feed condensate should now
be channelled to process. The biggest user is mill
imbibition. But we must remove the heat from the
condensate by the use of a heat exchanger with mixed
juice as the coolant. This cools the condensate for use
as imbibition and imparts the surplus heat to the
mixed juice on its way to the primary heaters. If there
is a surplus of cooled condensate after imbibition it
could be cooled further and used as make-up water
for bearing and crystallizer cooling water.
The other uses are centrifugal and filter cake
washing, lime dilution, and pump gland sealing some
of which recycle in juice or syrup evaporation. Tank
washing and floor hosing account for a relatively small
amount, and are usually run to waste.
Basically, once the factory is running, it can be made
self sufficient as far as water is concerned by judicious
collection of condensates, and. recirculation of cooling
water. It is essential to critically examine the collection
and use of condensate throughout the plant if the
ideal heat balance is to be found.
Mr. Bentley (in the chair): I would like Mr. Cargill
to give us further information about the condensate
used for maceration and whether hot water was used.
Mr. Cargill: The condensate is cooled and sent to
maceration via a liquid/liquid heat exchanger. The
total maceration is condensate. The condensate inlet
temperature to the liquid/liquid heater is approximately
150° F. and it is cooled by mixed juice to
approximately 100° F. At this temperature it is
applied to the mill as maceration. Surplus condensate
is available for process use.
Mr. Gunn: How did you solve the problem of high
conductivity readings which do not contain sucrose?
Mr. Cargill: We tried fitting an ion exchange column
in the sample line to the meter to remove ammonia
from the sample (the cause of high conductivity).
The resins used were specially imported from America
and were such that ash due to sugar contamination
would not be removed. This was not successful as
the sample column ran for only a few hours before
regeneration was required, and attempts to regenerate
were not good. The resin would not return to its
original condition. We have not solved the problem
of high conductivity.
Mr. Angus: There is apparatus available that will
prepare condensate prior to conductivity measurements
by removing ammonia and carbon dioxide.
Dr. F. Straub of the University of Illinois developed
the "Straub degasser" which condenses the steam and
then reboils it, removing ammonia and carbon dioxide
so that what passes to the conductivity cell is the
condensate plus non-volatile material contained in the
"carry-over".
The Powell-McChesney scheme for condensate
testing has a long tube which is calculated so that you
get pressure reduction and a controlled amount of
cooling. Ammonia and carbon dioxide are removed
to a low constant value before the condensate is
passed to the conductivity cell.
There is also the Larsen-Lane analyser which passes
the condensate through an acid regenerated cation
resin and gets rid of ammonia by ion exchange. We
are left with carbon dioxide which is removed by
reboiling. Acids which are very much more sensitive
to conductivity measurements than salts are formed
from the inorganic material in the "carry-over".
Mr. Cargill: We tried unsuccessfully to boil the
sample to drive off the ammonia present. To condense
the steam from this boiled sample and check its
conductivity would not be a check for contamination
of the original sample.
Mr. Phipson: For detecting ammonia we have obtained
electrodes with ten times the conductivity of
the usual electrodes. Lower-powered electrodes are
used to detect sugar.

134 Proceedings of The South African Sugar Technologists" Association- March 1966
Mr. Cargill: Sugar contamination is normally
detected at 9 micro-mhos. The ammonia contamination
raised this to 20 micro-mhos thus rendering the
instrument useless as a sugar detector.
Mr. Chiazzari: Why do you object to the use of "U"
legs instead of Gestra traps on pans?
Mr. Cargill: We have found "U" legs very erratic.
They cannot vent incondensible gases and gas or air
locks are a source of trouble. Perhaps if the upflow
leg was large enough they would work better but the
cost of all this large bore piping nullifies its use—
better to fit a trap.
Mr. Young: What is your feeling about injecting
cold water into the upward leg of the "U" leg to
avoid flashing.
Mr. Cargill: 1 hate the thought of wasting good
heat and water by doing this.

Proceedings of The South African Sugar Technologists' Association—March 1966 135
RESIDUAL FUEL OIL AS A SUPPLEMENTARY FUEL
Abstract
The need for supplementary fuel in local sugar mill
steam raising plant has long been accepted. While
solid fuels have always been used for this purpose, the
alternative liquid petroleum fuel oil may well find a
place for such a service. The basic needs, requirements
and techniques when considering the use of heavy fuel
oil for this service are discussed in this paper.
Introduction
To appreciate that supplementary fuels are extensively
used in our sugar industry, it has been reported
that during the 1961-62 milling season 130,150 tons
of coal and wood were used as additional fuel to
bagasse at an average cost of R0.60.3 per ton of sugar
produced. For various reasons shortages of bagasse
occur from time to time necessitating the use of an
additional fuel to maintain steam for factory needs.
Despite frequent enquiries and discussions relating to
the use of such fuels they continue to be used. Coal,
and to a lesser degree, wood, have been the supplementary
fuels used. Economics and supply have always
been the reason for their selection. Whether convenience
has played an important part in their selection
is immaterial, for they have enjoyed widespread use
and apparently continue to do so.
More recently residual fuel oils have become readily
available locally and some interest has already been
shown in the use of such a fuel for this purpose. It
is the intention to provide some basic information
such as characteristics of this type of fuel oil, the
storage and handling of such a product and some
operating techniques regarding its preparation for
firing. Some thoughts on equipment and conversion
problems are included.
The Economics of using Fuel Oil
It will be appreciated that the economics will vary
largely from mill to mill when considering the use of
a specific supplementary fuel.
A direct thermal comparison will reveal that coal
is more economical to use than fuel oil, but this is
not conclusive, as many other factors bear consideration.
The prospective users' present needs, e.g. existing
firing arrangements, location, labour force, etc., etc.,
can and will influence the selection of the particular
type of fuel to be used. Directly related to these factors
should be production and profit losses due to shut
down and/or change over from one fuel to another.
The easiest and most rapid method is obviously advantageous.
Only by considering the complete operation
can a true and factual assessment be made of the
economical necessity for incurring the capital expense
to convert to supplementary fuel oil firing or continue
with one of the more conventional methods practiced
today. Factors such as basic fuel costs, handling and
storage, ash disposal, operating costs, capital expen­
By J. GUDMANZ
diture and labour charges have to be thoroughly
investigated to determine the most desirable arrangement.
The Need for Oil Firing
Boilers originally designed for bagasse firing present
certain physical problems when oil burning equipment
has to be installed. It should be appreciated that
the conversion to oil is of a supplementary nature and
is to be used when the normally available fuel, i.e.
bagasse, is in short supply or is unavailable due to
breakdown or stoppage.
The question of bagasse shortages have been the
subject of many thorough and probably controversial
investigations in the past, but despite this such shortages
persist to a varying degree necessitating the
provision of supplementary fuelling arrangements.
A question asked is whether a bagasse fired boiler
can operate satisfactorily on fuel oil. The prospective
user may even expect a boiler efficiency of up to 80
per cent under such conditions, because this can be
readily obtained with fuel oil equipment. As the proposed
conversion is intended to cater for emergency
periods when the normal fuel — bagasse — cannot be
provided the user will be pleased to be in a position
to supply steam for his requirements, even at an
increased cost, thus maintaining production and not
reducing or even entirely losing profits through a
complete shut-down. Under such conditions the
question of efficiencies will be of secondary importance.
Therefore, bearing in mind that although economics
reveal boiler operating costs to be high, when comparing
oil against coal, and that such conversion will
probably not result in the high efficiencies of oil-fired
systems, or that the boiler manufacturer may consider
it impractical to convert because of either of the above
reasons, basic economics demand that the mill continues
to operate and produce.
Legitimate practical objections presented by the
boiler manufacturer can probably be overcome by
modification of their normal conversion techniques
and thus assist the user to maintain the continuous
supply of steam so necessary for his plant.
Key to Fig. 1:
A. Cold Oil Filters — Coarse.
B. Fuel Oil Pumps.
C. Fuel Oil Preheaters.
D. Hot Oil Filters — Fine.
E. Oil Feed Line to Additional Burners.
F. Burner.
G. Oil Return Line from Burners.

136
Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 137
Residual Fuel Oils
The specification shown below is typical of a heavy
residual fuel oil obtainable here and is suitable for a
wide range of applications including that of a boiler
fuel.
Table 1
Typical Physical Characteristics
Under certain circumstances, these relating to some
specialised applications, it may be possible to obtain
a residual fuel oil basically the same as shown above,
but having a lower viscosity. However, a fuel oil of
the type shown is entirely suitable for steam raising
purposes as now primarily under discussion.
Internationally, the Redwood Viscosity at 100° F.
is extensively used to describe a fuel oil and that shown
here would refer to a "950 second" fuel oil.
Storage and Handling
Because of the physical properties of residual fuels
certain conditions and facilities are needed for their
satisfactory handling and preparation for service. The
storage facilities necessary may vary and are dependent
upon consumptions and the location of the mill. An
indication of what may normally be required is shown
below.
Storage Tanks — the capacity of these would be
dependent upon estimated consumptions, the distance
from supply point and the method of delivery, normally
rail tank cars. Their size and location is determined
by available space at the mill, always considering
rail facilities and distance to the boiler house.
Outflow Heater — this is fitted to the tank and is
either electric or steam/electric and is sized to maintain
the oil at a suitable viscosity for pumping so as
to supply maximum operational requirements.
Tank Valves and Piping — suitable piping and valves
for filling and discharge lines are needed and
should be traced and lagged where necessary.
Transfer Pumps — these may be required and they
too should be sized to supply the maximum plant
requirement from the main storage tank to the service
or day tanks.
As this is a viscous product, the ambient temperature
plays an important part in the handling of the
fuel oil. Minimum temperatures should be observed
and the pumps used must be capable of handling the
fuel at the lower temperatures, thus outflow heaters
are installed to increase the temperature and reduce
the viscosity that this may be achieved.
It is not intended to enlarge on the storage aspect
but to conclude by noting that safety should never be
overlooked. Location, relative to thoroughfares and
fire hazards are important, and the design of such
storage facilities should always be based on sound
engineering principles and practice.
This aspect is normally handled by the supplier of
the fuel, who is able to advise on the best practical
arrangements.
Fuel Firing Equipment
The type of burner selected will determine the
method of supplying fuel oil to them.
On boilers of the type and size under discussion,
the distribution of the fuel oil to the boiler by means
of a ring main is preferred. In this manner the fuel
oil can be supplied to the burner at a constant pressure
as well as at the desired temperature for efficient
atomisation and burning. The pressure in such a ring
main is dependent upon the burner selected, but can
be adjusted to suit varying conditions should they
arise. A typical system is illustrated in Fig. 1, which
shows the various items of equipment that should be
provided. Some comments on the type of equipment
used on a typical ring main system handling heavy
fuel oil are given.
Pumps
Positive displacement pumps are recommended and
readily obtainable. The most common types used are
rotary pumps which can handle high viscosity fuel
oils. They are reliable even when pumping hot oil
continuously and are not oversensitive to the presence
of vapour in the oil. They can be either electric motor
or steam driven, but as the power requirement is not
excessive, electrically driven pumps are preferred.
Built in relief valves are necessary and avoid any
damage to the system, particularly when starting up
from cold. The pump size should be such as to supply
oil in excess of what is intended to be burnt in the
boiler. A pump capacity of 50 per cent to 100 per cent
in excess of the maximum hourly consumption rate of
the boiler would be considered satisfactory. It is
normal to provide a dual pumping system for such a
service.
Preheaters
It may be considered desirable to install a steam
and electric fuel oil preheater, but capital and operational
costs may decide against this. Steam preheaters
of suitable capacities can be obtained and should be
fitted with a reliable thermostat control system. Excellent
thermostatic controls are obtainable which will
maintain the oil temperature within a range of 5° F.
Although atomising temperature may be known, it
sometimes occurs that under actual operating conditions
it is found necessary to vary this temperature
for improved atomisation and more satisfactory
burning. Bearing this in mind the preheater should be

138
Proceedings of The South African Sugar Technologists' Association- March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966
able to provide the maximum quantity of fuel oil at
required temperature and in addition be so designed,
to supply at a higher temperature, as it is frequently
necessary to vary these oil temperatures until the
most satisfactory burning conditions are determined.
Slight variations in the viscosity of the fuel oil as
supplied may occur. If a particular burner is sensitive
to viscosity and will not atomise the oil satisfactorily
a change in preheat temperature will be needed. This
is another reason that some flexibility is required.
With regards to electric preheaters these should be
so designed that element loadings do no exceed 10
watts per square inch of heating surface. Oxidised fuel
oil deposits on the elements may occur if this is
exceeded. The recommended design factors for oil
preheaters are given in B.S.S. 799. Fig. 2 shows the
temperature-viscosity relationship for residual fuel oils.
The curve is that of a typical fuel oil and from it the
desired preheat temperature can be obtained when
knowing the viscosity at which the oil should be supplied
to the burner. The curve is for a "950 Second"
fuel oil. Viscosities of other fuel oils can be determined
by drawing a line parallel to that shown after plotting
the known viscosity at 100° F.
Filters
In order to protect the fuel oil pumps, coarse filters
are fitted in the supply line from the storage tanks to
the pumps. In addition, fine filters are located on the
hot oil side. This serves to protect any equipment
having fine clearances and minimises the possibility of
burner deposits. Effective straining of the oil before
it is circulated to the burners is essential as any
restriction or clogging of the burner nozzle will impair
atomisation causing poor burning conditions. Basket
perforations can be 1/16 in. and 1/32 in. diameter for
the cold and hot lines, but burner design features may
necessitate the provision of different sizes to these.
The strainer area should be in the range of 300 per
cent larger than pipe cross-sectional area.
It is usual to install dual filters in order to maintain
continuous operation when filters have to be opened
for cleaning.
Pressure Sustaining Valves
It is necessary to maintain the circulating fuel oil
pressure at a constant value. To attain this a pressure
sustaining valve is located in the pipe system at a
suitable point. Such a valve should be sized to handle
the maximum flow of oil and may be spring loaded or
of the diaphragm type. It is considered essential to
provide a hand operated bypass valve to be used under
cold start up conditions or should the pressure regulating
valve fail.
Lagging of Fuel Pipes
Ring mains should be suitably lagged in order to
reduce the heat loss and maintain the circulating oil
at the required temperature during operational periods.
Under certain conditions it may even be necessary
to provide steam or electric tracing over which the
lagging is fixed. Piping containing static fuel oil should
be traced and lagged to enable rapid start-up when
necessary. When installed and operating the system
should be able to handle and supply to the burners
mounted on the boiler the desired amount of fuel oil
at the required temperature. In designing the system
dual arrangements for filters, pumps and preheaters
should be incorporated, that continual operation is
assured. It will be appreciated that actual ring main
design is determined by the type of burner installed.
Therefore in Fig. 1 the filters, pumps and heaters
shown may be suitable for many systems, but the
piping arrangements from the fine filters, D, to the
burners, together with any return piping and oil
pressure controls would vary from system to system.
Oil Burners
Because of excellent flame radiation, improved evaporation
rates may be expected providing the combustion
chamber construction with respect to refractories
and insulating bricks is suitable for the higher
thermal load when operating on fuel oil.
It is probable that steam temperatures may be
reduced due to lower gas temperatures at the superheater.
If it is not possible to maintain the required
steam temperature, the quantity of excess air can be
increased thereby reducing the flame temperature,
and flame radiation, thus allowing higher gas temperatures,
and improving the superheat conditions. This,
however, results in high backend and stack temperatures
with a corresponding reduction in efficiency.
Boiler design with respect to these items, i.e. combustion
chambers, gas flow rates and fans will have to be
examined and the necessary modifications provided
which will ensure acceptable operating conditions.
The selection of any particular type of burning
arrangement would largely depend upon the recommendation
of the boiler manufacturer. As he is fully
aware of the specific design features of his boiler and
the need to operate satisfactorily on fuel oil, his
recommendations should be conscientiously considered.
He, in turn, should bear in mind that any potential
conversion to fuel oil is done to maintain steam supply
during comparatively short emergency periods when
the normally used fuel, i.e. bagasse, is in short supply.
He will then be providing an essential service to his
customer even should his converted boiler not operate
at the efficiency rate that his regular oil-fired units do.
Modern fuel oil burners are designed to supply oil at
a suitable degree of atomisation and vaporisation to
ensure rapid ignition and continuous clean burning
characteristics. To do this the fuel oil must be supplied
at the correct temperature and pressure. This means
that the equipment installed, i.e. filters, pumps, heaters
must be complementary to the burners selected. It is
not intended to discuss burners in great detail but
rather to provide a brief description of certain types.
Several types are used on larger steam boilers, the
more frequent being:
1. Steam Atomising Burners.
2. Pressure Atomising Burners.
3. Rotary Cup Burners.
139

140 Proceedings of The South African Sugar Technologists'' Association—March 1966
1. Steam Atomising Burners: These are designed for
use on medium to large boilers where several burners
are installed to supply the required quantity of fuel oil.
Oil is supplied at the comparatively low pressure of
30-60 p.s.i., and steam at a higher pressure of 80-140
p.s.i., or approximately double the oil pressure. A
steam consumption of below 2 per cent of boiler
evaporation is acceptable but a figure of 1 per cent
is desirable on large units and should be aimed at.
Turn-down ratios of up to 10-1 are obtainable on
steam atomising burners.
2. Pressure or Mechanical Atomising Burners: Atomisation
of the fuel oil is achieved by supplying heated
fuel to the nozzle at pressures in excess of 300 p.s.i.
An advantage of this type of burner is that operating
costs are lower as no air or steam is needed for atomisation.
In addition maintenance costs are usually
found to be lower. Mechanical atomising burners can
be of the return flow or straight through flow type
and also have turn-down ratios of up to 10-1.
3. Rotary Cup Burners: Rotary cup burners are
supplied as a complete unit which includes motor, fan,
oil pump, burner cup and oil preheater. In addition
ignition devices and flame failure controls are included.
The initial cost of such a unit is somewhat high but
large units are available and can be used for conversions
of the certain types of boilers.
Rotary cup burners achieve atomisation from the
centrifugal action of the cup into which oil is fed.
Air from the fan mixes with the spray of oil resulting
in fine atomisation.
Ignition Systems
For the type of operation under discussion a sophisticated
ignition system using gas, oil and electric spark
is not considered necessary. Torch ignition has certain
disadvantages mainly concerning safety or fire hazard,
but proper procedures can eliminate any dangers thus
reducing installation and operating costs. Rotary cup
burners have an ignition system incorporated.
Operating Conditions
To ensure smooth change over to fuel oil the operating
personnel should be fully instructed in the new
technique. They should be aware of the fuel oil
temperature requirements, the layout of the system
and method of operation.
It will be readily appreciated that no plant can
operate satisfactorily unless the operators are properly
trained and the system is kept well maintained and
ready to be used when needed.
Much has been said of corrosion and slagging problems
experienced with oil-fired boilers. Many theories
have been advanced regarding the causes of high and
low temperature corrosion of metal surfaces in boilers.
The same applies to slag formation in high temperature
zones of boilers and this can be related to certain
fuel oil characteristics.
To overcome and reduce these problems certain
operating techniques have been suggested, including
the use of fuel oil additives. This aspect could well be
the subject of a paper itself and will not be elaborated
on here. Suffice it to say that despite these problems
fuel oil is enjoying wider use and slagging and corrosion
is under continual investigation to determine
the most suitable means of reducing and overcoming
their effects.
Conclusion
As the supply and use of coal and wood have presented
problems during recent years, some thought
towards the use of an alternative supplementary fuel
must be given. The use of residual fuel oils as an
established supplementary fuel in place of bagasse will
require much investigation to determine the economics
as well as the suitability for conversion. Assistance in
this direction will be forthcoming from interested
parties, namely, the boiler manufacturers, the suppliers
of oil burning equipment and fuel oil suppliers.
In future it may well be more economical to use a
supplementary fuel and divert bagasse for the manufacture
of paper or similar products, which could
prove to be a profitable decision.
Diesel Engines
Complementary with the use of residual fuel oils in
boilers as an additional fuel is that of Diesel engines
operating on such a product. A brief discussion on
this is given.
The use of residual fuel oils as a Diesel engine fuel
has been practiced for many years. This practice was
established on marine engine applications, but, today,
many large stationary engines use this type of fuel.
The reason of course is economics, as the basic cost
of residual fuel oils is considerably lower than the
normally used automotive diesel fuel. This practice is
limited to slow-speed units where high centane numbers
of fuel oils are unnecessary.
With the light fuel the main requirements before
supplying the fuel to the engine fuel pumps is filtration.
With heavier residual fuel oils a more involved
arrangement is needed. The normal procedure is to
cold filter the oil, heat and centrifuge it, then store it
in a service tank. There it is held in a temperature of
100-140° F. before further filtration and heating to an
operating temperature of 180-200° F.
Certain operating problems must be guarded against.
The higher cylinder and ring wear rate on engines is
directly attributed to the ash and sulphur content of
the residual fuels, ash causing abrasive wear and
sulphur corrosive wear. Since the advent of this
practice much development has taken place with
special high additive oils for cylinder lubrication.
These 'Super' oils are of high alkalinity, thus combating
the ill-effects of corrosion. In addition, their
excellent dispersant ability keeps the ring area free of

Proceedings of The South African Sugar Technologists' Association—March 1966 141
undesirable deposits. Today, with the use of such
lubricants in these large slow-speed units, wear rates
approaching those of Diesel engines operating under
conventional methods are obtained.
The initial cost of providing for the satisfactory
preparation of residual fuels is costly as special filters,
centrifuges, oil heaters, pumps and service tanks with
heating arrangements must be installed. However, with
proper preparation, control and operation this can be
justified. This information is given for the interest it
affords, but it is not anticipated that such systems will
be extensively used in the sugar areas of Natal, unless
comparatively large Diesel powered electric generator
plants are needed.
References
A.S.T.M. 396 — Fuel Oils.
Buck W. (1962) —The Problem of Additional Fuel, S.A.
Sugar Tech. Assoc.
B.S.S. 799 — Oil Burning Equipment.
B.S.S. 2869 — Oil Fuels.
Mr. Bentley (in the chair): Sugar factories on the
whole probably use more additional fuel than they
should. I personally prefer fuel oil to coal as a supplementary
fuel but it is expensive and I wonder if Mr.
Gudmanz can tell us what price fuel oil can be supplied
to a factory to make it competitive with coal.
Mr. Gudmanz: Owing to railage and other charges
fuel oil for the sugar industry is probably at present
twice the cost of coal but it is possible that its price
might come down.
Mr. Gunn: If the price of oil fuel was reduced
sufficiently to make its use economical would there
be sufficient available in South Africa to supply the
industry's needs?
Mr. Gudmanz: The change-over would be gradual
and therefore the oil fuel suppliers would have sufficient
time to organise supplies.
Mr. Fokkens: When costing oil fuel against coal
in Cape Town, taking into account all charges, oil
fuel was R19.00 per ton against R12.50 for coal.
Mr. Hurter: What would be the cost of generating
power using oil instead of coal?
Mr. Gudmanz: I do not have any figures available.
Mr. Ashe: At Umfolozi we are running a 1,000 kW
alternator on residual fuel. The average load is
600 kW and we are saving R4 per hour. One disadvantage
is that the fuel has to be centrifuged and the
centrifuges require cleaning twice a day.
Mr. Hulett: There is no reason why molasses
should not be used as fuel if you adjust your burners
properly.

142 Proceedings of The South African Sugar Technologists' Association—March 1966
STEAM ECONOMIES AT DARNALL
The period 1957 to 1964 is the topic for discussion
as it was over these years that large economies were
made in the steam consumption at Darnall with very
little capital expenditure on additional plant to achieve
it. Table I shows the statistics relevant to the steam
economy of the factory and a study of this table shows
how the amount of fuel used has decreased notwithstanding
a large increase in water evaporated in the
factory.
In 1957 the boiler plant consisted of four boilers,
3 WIF Type B. & W. boilers of 7,322, 11,080 and
11,000 square feet heating surface respectively,
each fitted with air heaters and V & P type bagasse
furnaces and one C. E. boiler fitted with a C. E.
spreader stoker and of 50,000 lbs. per hour rated
capacity. These boilers at that time were continually
hard pressed to meet the steam demand of the factory
and coal was fed regularly to the C. E. unit. As can
be seen from Table I, by the end of 1964 these same
boilers were managing a far greater factory load
and the only extra fuel burnt was to start and stop
the plant as inadequate bagasse storage facilities
existed at the mill. In the meantime surplus bagasse was
being burned on the hillside.
In 1957, the factory was set up as is shown in
Chart 1. The evaporation was accomplished in two
Quadruple effect evaporators; juice heating and all
pan boiling was done on 10 p.s.i. exhaust steam.
Chart No. 2 shows the first step in the economy
sequence when the primary heaters were resited and
steamed with Vapour 1 from the spare 4,500 sq. ft.
evaporator vessel. This move necessitated the cleaning
of the evaporator over the week-end, but since the
mill stopped each week-end in any case, this was not a
serious drawback. The primary heaters at this stage
absorbed about 40,000 lbs. of steam per hour and
this resulted in a steam saving of about 10,000 lbs.
per hour and, of course, the brix of the syrup improved
as the evaporator became more powerful.
This saving of 10,000 lbs. per hour of process steam
resulted in a direct saving of fuel as at the time the
accumulator was passing in the region of 60,000 lbs.
per hour to the process main.
During the 1963 season some money became available
for the conversion of the factory to the remelt
system for the manufacture of J. A. sugar so further
alterations to the evaporator and the purchase of
D. J. L. HULETT
two additional heaters became necessary. Chart No. 3
shows how the heaters were installed and the arrangement
of the evaporator. The five vessels of the 22,500
square feet evaporator were connected together in
parallel so as to operate as one large pre-evaporator
supplying vapour for the pans, secondary juice heaters
and the first vessel of the 32,000 sq. ft. quad. The exhaust
back pressure was raised to 15 p.s.i. so as to
supply 9 p.s.i. Vapour I to the pan door. Primary
heating to 150° F. was done by a 2,300 sq. ft. liquidliquid
exchanger on the condensate and a 2,300 sq.
ft. heater operating on second effect vapour. The
old primary heaters were converted to pre-heaters for
the clear juice and were operated on exhaust steam.
This scheme of operation resulted in an exhaust
steam saving in the region of 60,000 lbs./hr. over the
original scheme and coupled with the increased steam
consumption of the prime movers due to the 15 p.s.i.
exhaust caused a surplus of exhaust over the process
steam requirements. To minimise this surplus more
water was applied to the imbibition but the surplus
remained even at 60 per cent imbibition on cane and
at 210 T.P.H. However, notwithstanding the fact that
Darnall mill continuously had a cloud of steam hanging
around the exhaust relief valve, the set-up was
healthy and the available bagasse storage space could
be filled in a day and a bagasse fire was kept going on
the nearby hillside.
1965 and the Future
During 1964 it was decided that all the "Hulett"
group of mills should be provided with stand-by
plant both at the boiler station and the power station
and so an additional boiler and turbo generator had
to be purchased for Darnall mill. It was considered
wise to use this opportunity to balance the process
steam demand with exhaust steam supply and so a
450 p.s.i. 750° F boiler was ordered together with a
Topping turbine of 2 MW. capacity exhausting at
200 p.s.i. This turbine would relieve 2 M W. of electrical
load from the existing 200 p.s.i. generating sets
and consequently their steam consumption would
fall by a corresponding 50,000 lbs./hr. In addition a
de-aerator feed water heater was required for the new
boiler and this increased the exhaust steam consumption
by approximately 10,000 lbs. per hour. All this
would result in a deficit of exhaust production over
process demand and it was visualised that should
further economies or increased evaporator capacity

Proceedings of The South African Sugar Technologists' Association — March 1966 143
be required, use could be made of thermo-compressors
using the live steam of 200 p.s.i. to re-compress
V1 back into the exhaust line at a slightly higher exhaust
pressure. Chart No. 4 shows this proposed
layout.
Once the extra fuel is eliminated and imbibition is
Year
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
Tons Cane
per hour
183.61
188.15
189.24
184.25
191.83
199.03
188.60
200.73
209.40
192.07
Fibre per cent
cane
15.57
15.01
16.04
16.53
15.09
14.81
16.23
16.12
15.86
15.69
*Records not available for these years.
Imbibition
per cent cane
47.24
44.78
41.87
48.02
55.71
55.15
62.00
59.89
59.00
58.88
Table I
increased to its maximum and still surplus bagasse
exists, this bagasse could become a nuisance. However,
bagasse is a useful raw material for paper and
board and so when a guaranteed large surplus of this
substance exists, it is likely that a plant will be installed
to process this and turn it into money.
Sucrose
per cent cane
13.26
12.93
13.01
13.65
13.93
14.03
13.28
13.54
13.99
13.19
Brix Syrup
57.26
55.76
56.15
58.08
58.82
61.86
61.65
61.94
60.74
60.58
Moisture
per cent bagasse
54.94
54.46
54.69
54.04
53.61
52.12
51.38
52.54
52.19
51.96
Coal & Wood
Rands
*
*
25,000
6,660
3,180
1,675
1,950
2,468
3,578
8,481

Proceedings of The South African Sugar Technologists' Association — March 1966 147

148 Proceedings of The South African Sugar Technologists'' Association — March 1966
Mr. Bentley (in the chair): It can be seen from Table
1 that the use of extra fuel diminished from 1958 to
1961 but from then on, despite additional throughput,
it started to increase again.
Mr. Renton: The chief reason for the increase last
year was the number of stops the factory made
because of lack of cane.
Mr. Jones: When exhaust steam was short from the
prime-movers what was the percentage and was it
sufficient to allow for variations in throughput and
avoid blowing-off ?
Mr. Hulett: At the beginning we were very short
and there was no blowing off. But until just recently
when we commissioned a new boiler, we were always
blowing-off vapour.
Mr. Jones: So when you commissioned the new boiler
you again had a shortage of prime-mover exhaust.
Mr. Renton: Yes, however, we have not had full
load on our topping turbine yet as only one boiler
has been operating on high pressure.
Mr. van Eck: Would you say that your steam accummulator
helps to maintain your steam balance?
Mr. Hulett: It helps to even out the load on the
boiler but I can't say it helps the steam balance. If
we were starting from scratch I would not buy an
accumulator.

Proceedings of The South African Sugar Technologists'' Association — March 1966 149
HOW TO MEASURE AND EXPRESS SUGAR MILLS
EFFICIENCIES
It is natural for the sugar industry to think and
speak in terms of sucrose: we purchase cane and sell
sugar on the basis of sucrose content; through the
different phases of process, whether milling or boiling,
we follow, analyse and check the sucrose to minimise
losses as much as possible.
All the sucrose of the cane is not available for
crystallisation as there are two factors governing
extraction and recovery of sucrose:
(a) The fibre content of the cane
(b) The purity of the juice.
However efficient our milling and boiling techniques
may be, some sucrose will be immobilised and
retained by the fibre of the cane and by the impurities
of the juice.
Bearing in mind that the aim of the milling process
is to separate the sucrose from the fibre and that of
the boiling process is to separate and crystallise the
sucrose from the non-sucrose of the juice, sugar
technologists have formulated yardsticks to measure
and express efficiencies so that guidance is provided
to all concerned. Such yardsticks are based on practical
results.
The choice of yardsticks should be based on their
merits which means that they should measure and
express efficiency, accurately and correctly, hence
providing guidance clearly to all concerned.
Milling
For the milling process, we have the choice of two
yardsticks:
(a) The lost absolute juice % fibre;
(b) The extraction ratio.
Both yardsticks are based on the concept that the
Unit of Fibre is the Unit surface of Adsorption which
shall immobilise and retain the juice and sucrose.
Lost Absolute Juice % Fibre expresses the parts of
absolute juice lost % fibre. Apart from the fact that
the calculation of brix or solids in bagasse, hence lost
absolute juice, is based on the blind assumption that
the purity of the residual juice is the same as that of
the last expressed juice, the term'lost absolute juice'
is fallacious for the following reasons:
(a) Absolute juice lost is calculated by dividing the
brix or solids in bagasse by the brix % of the cane's
absolute juice. We should remember that the solids
of the cane's absolute juice and those of the
residual juice in bagasse have totally different
purities which means that they are of a totally
byT. H. FOURMOND
different nature, one containing a higher per
cent of sucrose than the other.
Is it logical to use the brix % of a raw material
of a high purity to calculate the corresponding
loss of such raw material, from another raw
material of a much lower purity? It would appear
that this is quite a pertinent question bearing in
mind that we are only interested in measuring
and expressing the sucrose lost.
(b) As during the milling process the brix free water
is never extracted and remains attached to the
fibre of the bagasse, we only extract undiluted
and diluted juices. Therefore, how can we use
an expression which is closely associated with
brix free water to measure the loss of a juice which
does not contain brix free water?
For all these reasons it would appear that lost
absolute juice % fibre is not a reliable yardstick
to measure accurately milling losses and does not
express clearly and correctly milling efficiency.
Extraction Ratio indicates the percentage of sucrose
lost in bagasse as a percentage ratio of the fibre in cane.
From experience we know that the determination of
sucrose in mixed juice and bagasse is far more accurate
than the determination of brix or solids in such
materials. Although the sucrose in bagasse is merely
a Pol determination, the difference between Pol
and true sucrose in bagasse is so small as to be
negligible. Bearing in mind that the aim of the milling
process is to separate and extract the sucrose from the
fibre, it would appear that such a yardstick is accurate
and representative, and therefore gives better guidance.
To illustrate how lost absolute juice % fibre can
be confusing and misleading, let us take the case of
two mills (A & B) crushing cane of the same quality:
Fibre % - 15.00
Sucrose % = 14.00
Solids % = 16.47
Purity % = 85.00
Mill A, being of smaller size, uses a higher imbibition
to achieve the same lost absolute juice as Mill B.
From experience, we know that in any particular
cane, the juices of higher purities are more readily
extractable and washable than the juices of lower
purities as proved by the fact that in the milling
process, sucrose extraction is always higher than brix
extraction, which means that during the milling
process, sucrose is more readily soluble or extractable
than non-sucrose. Therefore, we find that Mill A,
on account of a higher imbibition, has a lower purity of
last expressed juice than Mill B and the whole picture
is as follows:

150
MILL Moisture"/,,
A. 52.00
B. 48.80
Calculating the corresponding lost
extractions, we arrive at the following
Bagasse Solids in
MILL % Cane Bagasse
A. 33.33 I.00
B. 31.25 1.00
A. 33.33 1.00 0.70
B. 31.25 1.00 0.73
and we find that although both, mills show the same
lost absolute juice, hence the same absolute juice
extraction, their corresponding extraction ratio and
sucrose extraction differ somewhat: Mill A's Extraction
Ratio is 1.47% lower than Mill B and its sucrose
extraction higher by 0.22 %.
A pertinent question would be the following one:
In practice, does a higher imbibition lead to a higher
sucrose extraction than non-sucrose, hence a lower
purity of last expressed juice or is it a mere assumption?
The following figures taken from the summary of
Laboratory Reports period ending 30th October,
1965, will prove that it is no assumption but a plain
fact.
Lost Abso- Imbibition Extraction
lute Juice % % Fibre Ratio
Fibre
Darnall 29.22 379 29.82
Tongaat 29.26 224 31.61
Such practical findings prove that a higher imbibition
will extract sucrose more readily than nonsucrose,
hence a lower extraction ratio. Theoretically,
we should expect the same extraction ratio for the
same corresponding lost absolute juice as both are
correlated to the unit of fibre. However, we find that
in practice it is not true.
This illustrates clearly that lost absolute juice
% fibre is not a reliable yardstick to measure and
express correctly milling efficiency.
As the aim of the milling process is to separate
and extract as much sucrose as possible from the
fibre, it would appear that Extraction Ratio is a far
better yardstick to measure and express milling
efficiency.
Milling Performance
For the benefit of all concerned, it is preferable to
express Efficiency as a per cent of what is available
in practice, and efficiency figures are usually related to
standards which have proved realisable in practice.
Therefore, we suggest that a milling performance
figure be introduced to express milling efficiency.
For instance, Natal Estates has proved that the
milling process can achieve an Extraction Ratio of 22.
Is there any objection to creating a standard of 20 as
an incentive to the mill engineer's creative mind?
Proceedings of The South African Sugar Technologists' 1 Association — March 1966
Bagasse Analysis
Fibre % Solids % Sucrose % Purity last expressed
45.00 3.00 2.10 70.0
48.00 3.20 2.34 73.0
absolute juice % fibre, extraction ratio, absolute juice and sucrose
figures:
Sucrose in Lost absolute Extraction Absolute juice Sucrose
Bagasse Juice ";', Fibre Ratio Extraction Extraction
0.70 34.40 33.33 93.93 95.00
0.73 34.40 34.80 93.93 94.78
Should we agree to this, then the milling performance
yardstick can be expressed by the following
formula:
Mill Extraction X 100
100—0.20 F
Where F actual fibre of the cane.
It is a simple and easy calculation and the meaning
is so obvious that it would convey proper guidance
to all concerned as it expresses the sucrose extracted
% sucrose available for extraction.
As a matter of interest, Natal Estates Milling
Performance for last crushing season was 99.22%.
Boiling
Boiling House Performance
So far, the most suitable and accurate yardstick to
measure and express boiling efficiency is the Boiling
House Performance as it expresses the crystallisable
sucrose recovered in sugar % crystallisable sucrose
available in mixed juice.
This yardstick is based upon the practical finding
that for every part of non-sucrose present in the mixed
juice a certain corresponding amount of sucrose will
be retained in the molasses when properly exhausted.
The retention factor for calculating the non-crystallisable
sucrose in mixed juice will vary according
to the purity of the mixed juice as juices of lower
purity (provided such low purities are not due to
cane deterioration) are usually associated with a
higher reducing sugar ash ratio which, as we know,
helps to achieve better exhaustion as salts have more
affinity for reducing sugars than for sucrose.
Bearing in mind that the retention factors represent
the averages obtained from practical results, the table
adjusting the retention factor according to the purity
of mixed juice is as follows:
Purity Mixed Juice Corresponding Retention
Factor
82 0.460
83 0.470
84 0.480
85 0.489
86 0.498
87 0.507
88 0.515
89 0.523
90 0.530

Proceedings of The South African Sugar Technologists' Association — March 1966 151
The interpretation of this can easily lead to confusion,
and be misleading. For instance one could
say that as one part of non-sucrose will retain 0.460
part of sucrose, then the expected purity of the final
know, is not realisable in practice. This would be a
fallacious interpretation of the Boiling House Performance
yardstick. The concept only means that for
every part of non-sucrose present in mixed juice we
must expect a certain corresponding amount of sucrose
to be considered as non-crystallisable as such sucrose
will be found in the molasses. However the same ratio
of non-sucrose—sucrose cannot be expected in the
molasses for the simple reason that some 18 % of the
non-sucrose is eliminated during clarification process.
This means that for every 100 parts of non-sucrose
available in mixed juice only 82 parts will be present
in the final molasses and this changes the whole picture.
It means that the expected purity of the Final
realistic.
The following figures will clearly demonstrate the
vast difference between the correct and incorrect
interpretation of Boiling House Performance for 82%
recovery of non-sucrose in molasses.
Purity
Mixed
Juice
82
83
84
85
86
87
88
89
90
Retention
Factor
.460
.470
.480
.489
.498
.507
.515
.523
.530
Correct In- Ii icorrect In-
terpretation t erpretation
for expected fi ir expected
Molasses Molasses
purity purity
35.9
36.4
36.9
37.4
37.8
38.2
38.6
38.9
39.2
31.5
32.0
32.4
32.8
33.2
33.6
34.0
34.3
34.6
This is a totally different picture of Boiling House
Performance as we know from practical experience
that such corresponding purities for final molasses
are obtainable in practice from such purities of mixed
juice. In the light of such figures we can, therefore,
conclude that Boiling House Performance represents
the best yardstick to measure and express Boiling
House Efficiency.
It could be said that sometimes such standards
have been surpassed. We should remember that
standards are based on the law of average. Any fresh
juice associated with unusual high reducing sugars
and low ash contents is bound to show a higher boiling
house performance than the one expected from the
created standard, and the application of Douwes
Dekker's formula for expected purity will confirm the
fact and show the green light.
It has also been found that after droughts, when the
reducing sugars ash ratio is exceptionally high, through
-—March 1966 151
cane deterioration such juices do not respond as above.
This is quite in order because crystallisation of sucrose
and exhaustion of molasses are not the simple result
of a hypothetical chemical reaction where one part
of non-sucrose will combine with so many parts of
sucrose but the physical properties of the juice
(viscosity) will also play an important part. Although
viscosity does not prevent crystallisation, it, nevertheless
retards the rate of crystallisation tremendously
and as the boiling house of any sugar mill is bound
by the limitation of pans' and crystallisers' capacities,
such deteriorated and viscous juices will lead to poor
boiling house performance, however high may be the
reducing sugars ash ratio.
Obviously, unusual high losses in the filter cake
or in the undetermined losses, whether through
entrainment or inversion, will also lead to poor boiling
house performance. Excessive destruction of the reducing
sugars during clarification by a too high pH
would also lead to poor boiling house performance.
Overall Performance
At present, the only formula for expressing the
overall work performed by sugar mills is the overall
recovery. However, this yardstick is not a true
reflection of a sugar mill efficiency as Mill Extraction
and Boiling House Recovery are closely related to the
fibre content of the cane and to the purity of the juice.
It is obvious that canes of low fibre content and of
high purity juice are bound to yield a higher overall
recovery than canes of high fibre content and of low
purity juice.
As the policy of the sugar industry is to recover as
much sugar as possible from any quality of cane, we
must look forward to a yardstick which will give us a
true picture of our sugar mills' efficiencies. If that
school of thought is correct, then the overall performance
of our sugar mills must come into the
picture.
It is sometimes said that mill engineers are only interested
in the performance of their mills and process
managers in the performance of their boiling house.
This is a fallacious school of thought because without
close co-operation between these two technical men,
no team work can be expected to achieve the highest
recovery of sugar.
But what about mill managers, general managers
and directors? We would imagine that their main
interests lie in the highest recovery of sugar and it is
obvious that the only yardstick to express this is the
Overall Performance.
Therefore, it would appear that for the benefit of
both technical and financial control, an overall picture,
reflecting the true efficiency of sugar mills, appears to
be necessary.
An overall performance yardstick could be obtained
by multiplying Milling Performance by Boiling House
Performance. It would indicate the quantity of crystallisable
sucrose recovered in the form of commercial
sugar as a percentage ratio of the crystallisable sucrose
available in cane.

152
Proceedings of The South African Sugar Technologists' Association—March 1966
ALTERATIONS AND IMPROVEMENTS TO
MOUNT EDGECOMBE MILLING TANDEM
Introduction
To enable the throughput of the Mount Edgecombe
Mill to be increased from 170 T.P.H. to 200 T.P.H.,
many major alterations and additions were undertaken
during the 1964 Off-Crop.
Needless to say sis with all new plant, many unforeseen
problems were experienced during the first crop.
In overcoming these problems however, valuable
experience was gained and the purpose of this paper
is to record all this information, and show how it was
later used to improve the performance of the mill.
For simplicity this paper is presented in 4 sections
as follows:
Section I
A summary of the original plant and the alterations
and additions which were effected.
Section II
Modifications which had to be made during the
first crop (1964) to maintain production,
and
Alterations made during the 1965 Off-Crop, based
on observations and experience gained after the first
season's operation of the new plant.
Section III
The 1965 Season and how the performance was
improved by constant attention and, where required,
further modifications.
Section IV
Conclusions and opinions based on a successful
1965 season.
Section I
Summary of Equipment
Comparison of Sketch No. 1 and Sketch No. 2 in
conjunction with the following summary gives an
indication of the magnitude of the alterations which
were undertaken and which had to be completed
within the 3-month period of the 1964 Off-Crop.
The 45 ft. long by 15 ft. wide cross carrier sited
below the ground level and previously used for the
Tramway System was removed and replaced with a
"Loading Table" measuring 130 ft. long and 25 ft.
wide. This table has a staging capacity of 150 tons of
cane and is capable of comfortably feeding 200 T.P.H.
of cane into the main cane carrier.
A second 15 ton 75 ft. Span Gantry was added to
the existing off-loading equipment to feed this table
and facilitate the change-over from Tramway System
to Road Transport.
The main cane carrier was split just beyond the
first set of knives and a new head shaft and drive
fitted, the purpose of this alteration was to enable the
By R. C. TURNER
second set of knives to cut on the turning point of the
main cane carrier.
A Drag-type slat carrier was installed to elevate the
cane from the discharge of the second set of knives
into the new "Grueddler" shredder, which was placed
in front of the Crusher and which replaced the 84"
Searby which had previously been between the crusher
and the first mill.
As with the shredder a Drag-type slat elevator was
installed to convey from the shredder and feed the
crusher.
All three of the above drives: Main Cane Carrier,
Shredder Elevator and Crusher Elevator were fitted
with Electro-Magnetic couplings for speed variation
and to facilitate the Automatic carrier system which
we installed.
The alterations which had to be made to the actual
milling train were:
(a) The replacement on five mills of the Apron-type
intermediate carriers with "Drag-type" slat
elevators and "Vertical chutes".
(b) The replacement of all the existing feeding devices
which we knew as "Apron Pusher Carriers"
with Underfeed Rollers.
(c) The removal of the Searby Shredder and the
re-routing of all mill platforms, catwalks and
maceration piping.
(d) At the same time we decided much to our regret
later on in the season, to do away with the
cush-cush sieves and elevator and to pump the
first expressed juice and mixed juice direct to the
Peck-strainers and Vibro screens. The cush-cush
being returned to the mill via a launder carrying
the first mill maceration.
These then were the alterations which were undertaken
and completed prior to the start-up of the
1964-65 Crushing Season.
Section II
Cush-Cush
The first equipment to give trouble after the 1964
start-up was the maceration and cush-cush pumping
system.
In designing the mounting of the bearing supports
for the Underfeed Rollers, use had been made of the
brackets which were part of the casting of the "Mill
Cheeks", and had previously supported the dead-eyes
and side-plates of the Intermediate Apron Carriers.
Mounting the Underfeed Rollers on these brackets
however, restricted their adjustment and prevented
them being inter-meshed with the grooving of the
front Roll of the mill.
This was a bad mistake as this left a gap between
the rolls of as much as 2 in., consequently when any
of the mills jibbed even momentarily, a blanket of

Proceedings of The South African Sugar Technologists' Association — March 1966 153

V'dfcl
154
Proceedings of The South African Sugar Technologists' Association — March 1966

Proceedings of The South African Sugar Technologists' Association — March 1966 155
bagasse 84 in. wide and up to 2 in. thick would be
discharged into the Juice Gutters.
The conventional type of chokeless pumps in use
at the time were never designed to handle this quantity
of cush-cush and bagasse, and even after repeated
attempts at using different types and shapes of
Impellors and Casings, we eventually had to make
our own pumps to do the duty.
These pumps together with the external wooden
trash-plates which we had fitted between the Underfeed
Rolls and the Front Rolls, succeeded not in curing
our problem, but in transferring it one stage further
down the line to the Peck-strainers and Vibro-
Screens.
Although this station was very much overloaded
and inefficient operating under these conditions, we
managed to keep it going until the close of the season.
During the 1965 Off-Crop, knowing where the
trouble originated we modified the Mill Cheeks,
burning off the original brackets and welding on new
ones, this enabled the Underfeed Rollers to be repositioned
to allow the grooving to fully intermesh
with the front Rollers of the mills.
To ease the load on the Peck-strainers and Vibroscreens
and at the same time improve this station we
installed three 6-ft. wide D.S.M. Screens with 1.5 mm.
screen apertures as pre-screeners.
This system has worked exceptionally well and
trouble free throughout the 1965 Season as shown by
the following comparison of Mill stoppages and time
lost:
Number of mill stops and time lost due to cush-cush
pumping and screening plant:
1964-65 Crop 22 Stops 14 hrs. 2 mins.
1965-66 Crop Nil.
Mill Elevators
When these elevators were designed it was arranged
for the slats to be angled forwards on their attachment
links to prevent the carry-over associated with slat
elevators, and also to facilitate rapid and clean fall-off
of the bagasse when the slat passed over the chute
opening.
In testing the elevators however, we found that
with the slat inclined at this angle, they would tend to
compress the bagasse (see Sketch No. 3) when travelling
around the tail shaft. To obviate this the chain
was turned end for end, the edge of the slat now leading.
This however was not entirely successful at
crushing rates above 190 but worked well with high
elevator speeds and lower tonnages (150 ft. per
minute).
It was during this troublesome period that we
established that: "Providing the centre of radius of
the boot itself is established, on the tail-shaft centre
and minimum clearance is allowed between the plating
of the elevator boot and the slats (see Sketch No. 4)
then the elevator will perform efficiently irrespective
whether the bagasse is discharged into the elevator in
'front or behind' the tail-shaft."
After we altered the elevator boots to comply with
the above, very few major stoppages were recorded.
We were however, continually troubled with minor
stoppages due to broken slats, choked attachments,
worn-out brushes, etc. All these faults were analysed
and the following modifications were made during the
1965 Off-Crop to rectify them:
(a) The return flight "Runner bars" were removed
from under the chains and repositioned under
the slats, the rollers now only had to rotate
while passing over the head and tail shaft
sprockets. This has had the desired effect and
has arrested the rapid wear which was taking
place between stainless bush and steel roller.
(b) The wooden slats 6 in. wide and 2 in. thick were
replaced with 6 in x -j in. M.S. Flat Bar backed
with a 2{- in. angle bar for rigidity, these in
addition to being cleaner were lighter and less
likely to break.
(c) With the elevator boots working effectively
the speed was reduced to 110 ft. per minute further
contributing to the life of the chain.
(d) The attachment links and wearing plates have
also been altered, but neither of these have been
in operation long enough to warrant comment.
Comparison of Mill Stops on the mill elevators is:
1964 178 Stops 404-hrs. Lost
1965 12 Stops 5i-hrs. Lost
Shredder Elevator
This Drag-type elevator between the second set of
knives and the shredder was the cause of many mill
stoppages, the partially knifed cane travelling on the
bottom of the cane carrier would either bridge across
the lower slats and choke around the headshaft, or
bridge at the discharge from the knives and choke the
boot of the elevator.
Only after repeatedly unsuccessful attempts to
modify this elevator did we reach the conclusion that
for a Drag-type elevator to work effectively on unsh
redded cane, the preparation by the cane knives
has got to be of a very high standard. Unable to
improve our preparation due to the H.P. available
being marginal we decided to replace the elevator
during the 1965 Off-Crop with a conventional aprontype
carrier smilar to our existing cane carrier.
Comparison of Mill Stops due to this elevator:
1964 68 Stops 28-^ hrs. Lost
1965 18 Stops 8 hrs. Lost
Mills
In spite of the extensive arcing which we did on the
rollers, considerable difficulty was experienced
throughout the 1964 season in getting the mills to
feed properly.
This condition was not particular to one mill only,
but was general throughout the milling train, and at
. no time was it possible to load up the mill "Hydraulics",
and obtain good performance. Improvements
were certainly maae to "Feed chutes" (see Sketch
No. 5) settings, etc., but the results obtained were
never entirely satisfactory.
From observations made and the comprehensive
records kept of each choke or breakdown we were
however, able at the end of the season to determine

156 Proceedings of The South African Sugar Technologists' Association — March 1966

Proceedings of The South African Sugar Technologists' 1 Association — March 1966 157
Proceedings of The South African Sugar Technologists' 1 Association — — March 1966 157
that our feeding problems throughout the milling train Underfeed Roll Driving Chains
were due to:
In common with the other mills using this type of
(a) Insufficient adjustment on the vertical chutes. feeding device, frequent mill stops were recorded
(b) Feathering of the mill trashplates due to exces­ owing to broken driving chains, and as a result the
sive plate wear, the result of intensive arcing of time lost due to these stops was extremely high
the rolls.
during the 1964 season.
(c) Heavy arcing damage to rollers,
To ensure that the best possible use was being
(c/) Incorrect mill, chute and underfeed roller set­ made of the chain fitted, the P.C.D. of the sprockets
tings and ratios.
on all these drives was increased to the maximum per­
(

158 Proceedings of The Smith African Sugar Technologists' Association — March 1966
crusher, this being particularly noticeable when the
elevator was running fast.
To rectify this the headshaft of the elevator was repositioned
well back and clear of the opening to the
feed chute, this reduced the velocity of the bagasse as
it left the slat and allowed it to tumble freely into the
chute opening. (See Sketch No. 6.)
This alteration has proved so successful that the
rest of our elevators are being modified to this design
during the present off-crop.
With the Crusher (first Unit) extraction at 68 percent
the overall extraction rose steadily until the
eighth week when a 96.51 per cent was recorded.
The high moisture of ±53 per cent however,
plagued us right through until the 11th week, when we
were at last able to fit a new discharge roller "with
messchaert grooves" into the last unit. This improved
the drainage and cured the second fault. (None of our
rollers had been machined with messchaerts for the
1965 crop.)
The moisture in final bagasse immediately dropped
by 2 per cent to 51.3 per cent and the unit extraction
rose to 41.3 per cent.
Encouraged by these results and of the opinion
that still better performance could be obtained from
this unit, a new chute was designed to keep the pressure
feeder full under all conditions, and allow the mill
to be run at a more constant speed.
This chute was designed and fitted with the added
facility of automatic speed control of the mill. Both
were an immediate success, reducing the speed of the
last unit by 20 per cent with a subsequent reduction
in moisture of 1 per cent and enabling the moisture
to be maintained at 4:50 per cent for the rest of the
season.
The alterations made to this unit alone had the
effect of reducing the final moisture by a full 3 per cent
and emphasised the importance of messchaert grooves
in the discharge roll of the last (drying) unit of a
milling train.
Several other smaller alterations were undertaken
during the season, viz:
(i) The conventional type of maceration discharge
trays were replaced with our "Weir" type of
tray to ensure an even distribution of maceration
across the carriers. (See Sketch No. 7.)
(ii) Special nozzles were fitted onto the imbibition
water supply to assist the water to knife into,
and penetrate the bagasse.

Proceedings of The South African Sugar Technologists' Association — March 1966
(iii) The mills were hydraulically loaded to such an
extent that we started breaking roller shafts,
stripping coupling boxes and tripping out mill
motors on overload.
All these factors were in some way contributory
to obtaining an extraction of 96.55 per cent and a
final moisture of 50.5 per cent for the 20th week of
the season, and an overall figure of 96.02 per cent for
the 1965 crop.
A comparison of Mill performance for the past two
seasons is:
1964
Crushing Rate . . . 189.43
Extraction 93.74
Moisture 52.43
Fibre 15.65
Imbibition % Cane . . 38.93
Sucrose % Cane . . . 13-69
Mcch. Eff. 92.2%
Season
1965
175T.P.H.
96.02%
51.8%
16.13%
43.7%
12.5%
95.32%
Section IV
Conclusions
The two most important factors which an engineer
must consider when he wishes to improve the performance
of a milling train are:
159
(i) The Mechanical Efficiency
(ii) The Overall Extraction
And the following information could possibly assist:
Mechanical Efficiency
(i) Keep records of every breakdown and stop
which occurs on the milling train or ancillary
plant.
(ii) Investigate each breakdown and ensure that
everything possible has been done to prevent a
re-occurrence.
(iii) Even new installations and designs are susceptible
to breakdowns and may require modifying.
(iv) Ensure that designs or alterations are kept as
simple as possible, it is always the simple
designs which are the most efficient.
Milling Efficiency
Here we list in their order of importance the requirements
the writer considers necessary to obtain good
milling results:
1. Even loading and continuity of cane supply.
(No bald patches on the carriers.)
2. Good cane preparation from the Knives and
Shredder.

160 Proceedings of The South African Sugar Technologists' Association - • March 1966
3. Good feeding from conservative chute/underfeed
and underfeed/mill ratios. (These are not
pressure feeders, neither are we dealing with
four roller mills.)
4. Low surface speed with high Hydraulic pressures
and floating top rolls.
5. Sufficient Front Roll drainage on all units.
6. Ample discharge roll drainage on the last unit.
(The gain when used on intermediate mills does
not compensate for the risk of roll damage.)
7. Even distribution of maceration and imbibition.
Too much emphasis cannot be put on the importance
of obtaining the correct ratios conducive to
good feeding and it is hoped that the following table
of ratios for a Six Mill Train preceded by a shredder
may be of assistance.
Ratios of Settings
Unit Fw/Dw uF/Fw C/uF
1
3
2.4
2.2
2
2.3
2.4
2.0
3
2.3
2.3
2.0
4
2.25
2.2
2.1
5
2.2
2.0
2.1
6
2.2
1.85
2.1
Fw — Work opening 1 Detween between front a and ind top rollers.
Dw — Work opening between disci' discharge and top
rollers.
uF — Set opening between underfeed and top rollers.
C — Chute discharge opening.

Proceedings of The South African Sugar Technologists* Association—
Mr. Dent: Tongaat has had trouble due to the nonintermeshing
of under slung feeder rollers.
Referring to sketch No. 5, Tongaat also had feeder
chutes as altered by Mount Edgecombe in the 1964
season but in order to make them work had to change
back to their initial design. Admittedly the chutes
were very narrow at the top and the cane would not
drop.
We are at present using the weir type maceration
trays as designed by Mount Edgecombe and we find
them excellent.
Mr. Turner: We originally had trouble getting cane
to drop into the very narrow opening at the top of
the feed chutes until we moved the head shafts back
and overcame this problem.
Mr. van Hengel: I am sorry that Mr. Turner did
not give his Mill Ratios on the basis of the discharge
work opening as we are now accustomed to this
standard method of the Mutual Milling Control
Project, decided upon at a meeting where all companies
sent their representatives. Such standardisation
facilitated the exchange of information which contributed
to the success of the project.
Referring to Mr. Turner's ratios and settings on
page 9 of his paper, the corresponding ratios on discharge
work opening for the feed rollers would have
been from unit No. 1 to 6—7.2, 5.5, 5.3, 4.9, 4.4 and
4.1 and the chute ratios referred to the discharge
would have been 15, 8, 11, 10.4, 10.3, 9.2 and 8.6.
Now the Mount Edgecombe tandem has a very high
performance and the ratios and settings used pretty
well coincided with the recommendations of the
Mutual Milling Control Project of the Sugar Milling
Research Institute.
The S.M.R.I, originally recommended that ratios
should vary from 7 to 5 on the last mill and in a paper
2 years ago by me it was shown that (theoretically)
ratios varied between 11 and 8. S.M.R.I, recommendations
are now between 13 and 11.
Mr. Turner: At the beginning of last season our
chutes were not set to any given ratio initially but
to a formula which had been used successfully in
Australia by Donnelly for setting vertical feed chutes.
Then by means of plotting a graph for the front roll,
back roll, underfeed roll and chute setting it was possible
to pick out immediately from the graphs any
irregularities or severe discrepancies between one
mill and the next and so establish what we consider
the optimum ratio.
Mr. Van Hengel: I know Donnelly's formula, but
there are no feeders like this in Australia and 1 cannot
see that they recommend the average.
The Donnelly formula supported by the University
of Queensland is K = \ (WOxD), where WO =-
Front Work Opening, D =•- Top Roll Diameter.
Mr. Turner: In Mourilyan, Australian underfeed
rollers were fitted to the second and third mills in
1960 and *Fitzmaurice writes that the moisture of
the No. 2 Mill was reduced from 57/59% to 52/54%
and the moisture of the 3rd Mill also dropped considerably
as it was possible to obtain much closer
settings of the feed and delivery rolls after fitting
underfeed rollers.
I agree that this formula was not derived to suit
underfeed rollers nor was it meant for use on feed
and top rollers but I assume that the actual formula
for the chute opening was derived to suit a given feed
opening and would apply to any two rollers provided
the maximum self feeding angle of these rollers was not
exceeded.
Mr. Hulett: At Darnall, throughout the mill, the
ratio of the underfeed roller setting is under 3 to 1.
Mr. Turner: You must have powerful drives and
must be using the underfeed roller as a 4th mill roller
and not as a simple feeding device.
Mr. Renton: That is true—we do use them as
4-roller mills.
Mr. Ashe: The idea of not having chevrons on the
top roller seems sound but in practice I have found
that choking of the mill occurs. Mr. Turner, with your
drives as shown how do you adjust your carriers to
take up wear on the chain?
Mr. Turner: The plumber blocks of the head shaft
are fitted to gusseted brackets at the top of the carrier
and by inserting packing plates of various thicknesses
between the plumber block and its supporting
bracket, the chain is either tensioned or slackened.
The drive consists of a shaft-mounted gear box with
the motor fixed underneath the whole assembly and
held in position by means of a torque arm.
Mr. Renton: I think it unnecessary to have chevrons
on top rollers if they are kept rough by arcing.
*FITZMAURICE, C. H. Some Notes on milling train improvements at Mourilyan. Q.S.S.C.T. 28th Conference 1961, p.p. 81-84,

162
Proceedings of The South African Sugar Technologists' Association March 1966
THE ELECTRICAL SUPPLY SYSTEM OF SUGAR
FACTORIES
When attempting to achieve the most economical
conditions of power supply in a sugar factory, it is
not only the power generation part of the system which
requires attention, but also the components of the
distribution network. The electrical supply system
may be said to consist of the following components:
Generators,
Main switchgear,
Transformers,
Motor-control-centres,
Drives and cable network.
The total power requirement of consumers in a
factory determines the choice of generation and
distribution voltages. When, for reasons to be dealt
with in the latter part of this paper, high tension is
chosen, the location of the power house in relation
to the factory and also the positions of transformer
stations and low tension main switchgear, have a
considerable influence on the final layout of the
system.
The transformer stations may be arranged either:
a) —at load centres as separate stations where two or
three units are installed together with their LTswitchboard.
b) —at a central location where all units and the
LT-main switchboard are installed.
For operation of the factory on an LT supply
system, case b) would apply.
Let us first discuss case a), where the stations are
placed in centres of power. It depends on the size
of the transformers whether their LT sides should
be connected in parallel or to separate busbars. In
view of rupturing capacity considerations, more than
2,500 to 3,000 KVA should not be connected in
parallel. Coupling facilities are, however, advisable,
so that the power supply of any section, where a
transformer fails, may be maintained. This however,
necessitates that transformers of load centres are
dimensioned accordingly. If, for instance, three
transformers are installed, any two of them must be
capable of carrying the total load connected to the
three. If there is a possibility of temporarily disconnecting
non-essential consumers however, the transformers
need not be so amply dimensioned. If it is
preferred to minimise the size of transformers another
method to ensure constant supply, even in the event
of a transformer failing, may be employed. This
involves an LT-ring cable system with isolators in the
stations, where all switchboards of power centres are
interconnected. It depends on the distances between
stations and the power to be transferred, which
solution is superior and more economical. An HT
ring system of course is also possible. This however,
necessitates HT switchgear in each load centre and
consequently represents the more expensive way.
Normally, transformers of load centres are radially
By A, GRADENER
fed by separate cables, which are run from the HT
switchboard of the power house.
In case b), where all transformers are located in one
station, a costly ring main is unnecessary as the LT
system has common busbars. In order to avoid heavy
short circuit currents, these busbais .must be divided,
into sections, with the possibility of coupling in the
event of a failure of any transformer.
Whether case a), or case b), is chosen depends on
the physical layout of the factory. The primary aim
is to have minimum lengths of cables between the
LT main switchboard and the motor control centres.
This is of the utmost importance as losses incurred
in an LT cable system are a multiple of those of an
HT system. The reason for this is simple to explain
as losses go into account by the square of the current
but only linear to the resistance—i.e. of the cross
section—of cables. For this reason, electrical losses
being economical losses, HT should be brought as
closely as possible to consumers. If the power house
is not situated close to the factory, it would be very
uneconomical to have the transformers there, necessitating
long runs of LT cables of considerable sizes
to motor control centres. In this case the power centre
system as described under a), lends itself.
If the power house is located in the centre of the
factory, the centralized transformer station as described
in b), is mostly the advisable solution.
Considerations of power are decisive for the design
of electrical equipment and plant. It is not only the
power requirement that is of interest, but also the
power factor involved. This defines the KVA rating
to which the electrical plant must be designed.
In Fig. 1, the conditions which prevail when electricity
flows are shown. There is hardly an electric
consumer which does not possess resistance, inductance
and capacitance simultaneously and it is only

Proceedings of The South African Sugar Technologists' Association—March 1966 163
the predominance of one or the other which defines
the final behaviour.
The electric current will either lag or lead the voltage
by an angle / between 0 and TT/2. This can be illustrated
either by drawing the sine curves of voltage
and current in relationship to one another or by
drawing a vector diagram which shows voltage and
current vectors. The two illustrations of Fig. I, both
show the same conditions, i.e. the voltage V and the
current I at an angle / to one another.
The lengths of the vectors are in direct proportion
to the r.m.s. values. It may now be imagined, that the
actual current (I) is made up of two components,
one of which is "in phase", with the voltage (active
current), while the other is out of phase by 90° (reactive
current). As, according to the rules of geometry:
the active current la -.- Ix cos /and
the reactive current Ir -••-: ix sin / the
resulting power may be arrived at by multiplying
both sides of the equations by the voltage across
phases (three phase system) and %/J"
active power Pa --•• V >; vT>' la " : V ;•: •v/J'x 1
:-: cos/
reactive power Pr V .: vT • If V vT : 1
x sin f
The maximum electrical active power of a turboset
is determined by the mechanical output of the turbine.
The required maximum KVA rating of the alternator
results from the turbine output and the power factor
at which the alternator supplies this power. The consumers
of a factory draw both reactive and active
power. Each consumer has a certain power factor
which indicates the ratio of Watts/VA.
The mean value of power factors of all consumers
is the power factor at which the alternator must
supply power. The power factors of individual
motors (being the main consumers in a sugar factory),
depend on their HP rating, their speed and their
loading. Tacle I shows the influence of these factors.
TABLE I
Influence of H.P. rating, speed and loading on the
power factor of an electrical motor
Influence of rating:
(all at same speed)
Power
p.f.
10 kW
0.85
100 kW
0.88
1,000 kW
0.89
Speed influence:
Speed 500 rpm (at (at same HP)
750 ipm
1,0001 l.OOOrpm rpm 1,500
rpm rpm
pf 0.80 0.86 0.87 0.88
Loading influence:
(at same
HP)
4/4
Loading
3/4
Loading
1/2
Loading Loading
at 500 rpm p.f. 0.80
0.76
0.65
at 1500 rpm p.f. 0.88
0.85
0.77
As may be observed, the influence of the rating is
not great. This also applies to the speed, not taking
very slow speeds into account. The loading however,
influences the power factor considerably. Consequently
overdesigned motors, which in effect are not
properly loaded, reduce the power factor to a considerable
extent. It is always good practice to design
equipment with a margin of safety but this does not
apply to the ratings of electrical motors where an
overdimensioning results in a poor power factor.
In turn the reactive current required must be generated
in expensive machinery. Apart from that, excessive
HP adds to the cost of the motors involved.
As a rule, the power factor of a sugar factory is
about 0.7, if power factor correction capacitors are
not installed. It is therefore advisable to choose a
KVA rating of the alternator, which is large enough
to allow full utilization of the given power of the
turbine. (It must be born in mind, that alternators are
normally designed for a power factor of 0.8). In
order to demonstrate the foregoing clearly, conditions
of a 3,000 kW alternator are investigated in the
following and demonstrated in Fig. 2.
Power Limitation Diagram of Synchronous Alternators
The kW rating is divided by 0.8, to arrive at the
KVA rating of the alternator. Therefore, this rating
is ' = 3,750 KVA. It is represented by the line
0.8
A — C, of the diagram in Fig. 2.
When ascertaining the power which the alternator
is capable of supplying at a power factor of 0.7, it
must be realised that the permissible power, in KVA,
of the alternator operating between a power factor of
0.8 to 0, decreases along the curve C-E-F. The
reason for this is that an increase of reactive power
requires additional excitation. The latter mentioned
however, has its limits as the permissible temperature
of the alternator must not be surpassed. At a power
factor of 0.7 the KVA would be represented by the
line A-E, corresponding to 3,500 KVA, i.e. to 93 %
of the original rating. As however, at this poor power

164 Proceedings of The South African Sugar Technologists'' Association- -March 1966
factor the reactive power has been increased, as per
line G-E, the active power available is also reduced
to 2,400 kW (line A-G), i.e. to 80% of its original.
Consequently, the turbine is not being fully utilized
under these conditions as part of the available power
is idle.
In order to comply with the required conditions a
larger alternator must be chosen, say 4,500 KVA at a
power factor of 0.8, corresponding to the line A-C
At a power factor of 0.7 this alternator will only be
capable of supplying 4,230 KVA, corresponding to
line A-E'. If full turbine power of 3,000 kW is utilized
at a power factor of 0.7, 4,280 KVA would result,
being represented by the line A-E." As the alternator
is not quite large enough to cope, the only other
possibilities which remain, other than employing a
larger alternator, are to utilize either the full power
of 3,000 kW but at a power factor of 0.71 or to
operate at only 2,940 kW at a power factor of 0.7.
As may be seen from the foregoing, the alternator
KVA must be adapted to the power factor as given by
the consumers, to ensure that it can generate the
active power corresponding to the turbine power and
at the same time supply the reactive power required.
It is of course possible to load the alternator with a
greater active power if the reactive power is generated
by another source, such as by capacitors. In the
following, a price comparison is given, in order to
ascertain which solution is more economical. Fig. 3
compares the situation between the alternators as
mentioned before, both working at 400 Volts.
In example 1, a 4,230 KVA alternator works at a
power factor of 0.71 and. in example 2, a 3,750 KVA
alternator works at a power factor of 0.8. In both
cases the load comprises 7 consumers each of which,
for the sake of simplicity, is 605 KVA. In the 2nd
example however, the reactive power is compensated
by 7 capacitors, one of 100 KVA r, at each consumer.
In this case therefore, the power which the alternator
must supply, is reduced to 3,750 KVA and consequently
a smaller machine is required. In both cases,
the alternator is connected to the main switchboard
by means of a busbar with a length of 60 feet. The
lengths of cables from this switchboard to the consumers
is 300 feet. The required cross section is
smaller in. example No. 2, as the reactive power is
generated at the consumers and therefore need not
pass through the conductors. The busbars of the
switchboard and the circuit breakers have been
neglected, as in both cases a 6,000 Amp alternator
breaker and 1,000 Amp circuit breakers are required
for the feeders. If the price of the smaller alternator
is represented by 100%, the following result is
obtained:
The price of the larger alternator is 106%. The
busbars in example I, are 14% and in example 2, 11 %.
The cable price is 36 % in example 1 and in example
2, 28%. In the 2nd example an additional 17% is
required for capacitors and power factor control
equipment. The total price for example 1 is then 156%
and for example 2, also 156 %. Both solutions are
equal in price and it is adviseable to chose the technically
simpler method, i.e. to use a larger alternator.
As a rule, it is more economical, in the case of low
power, to generate the reactive power in the alternator.
This is clearly shown in examples 3 and 4.
Both alternators of examples 4 and 5 generate 1,000
kW, the one of example 4, being for 1,250 KVA at a
power factor of 0.8 and the one of example 3, being
for 1,490 KVA, at a power factor of 0.68. Again, 7
equal consumers are supplied, each of 210 KVA. In
example 4, the reactive power is compensated in 7
capacitors of 50 KVA r each.
If the price of the smaller alternator is represented
by 100% the following result is obtained:
The price of the larger alternator is 113 %, the prices
of the busbars are 10% and 12%, while the cable
prices are 12% and 15%,. However, the price of the
power factor equipment is 35% in example 4, as
compared with 17% in example 2. The conclusion
is, that the plant with the larger alternator, totals up
to 140% while the plant with the smaller alternator in
combination with the power factor connection
equipment, totals 157%. The latter mentioned is
therefore 12% more expensive than the former.
Generally speaking therefore, it is more economical
and technically simpler to have the alternators laid
out for a power factor of 0.7. Wherever new alternators
are required, this should be taken into account.
With existing generating plant, the only solution is to
install capacitors which should be connected very
closely to consumers, in order to relieve circuit
breakers and cables. Motors of about 50 kW and

Proceedings of The South African Sugar Technologists' Association — March 1966 165
Proceedings of The South African Sugar Technologists' Association — March 1966 165
more, would be individually compensated, especially The short circuit current reaches its peak after the
when continuously in operation. Smaller motors first half wave, i.e. at 50 cycles, a period of 10 milli­
however, would be compensated in groups or all seconds (Length b, in Fig. 4). From cycle to cycle the
together. In both cases however, it is essential to amplitudes become smaller and the unsymmetrical
control automatically the switching of capacitors in behaviour becomes more and more symmetrical.
order to adapt the capacitances to the reactive power Within about 250 milliseconds the short circuit
requirements and thereby avoid over-compensation. current oscillates symmetrically. A steady state, the
Over-compensation bears the danger of self excitation "continuous short circuit current", is attained after a
and voltage increases which in turn endanger the number of seconds. As may be observed from Fig. 4,
insulation of electrical equipment and machinery. a fading DC component occurs in the first instance
With individual compensation of motors, the of a short circuit and displaces the amplitudes of the
capacitor is generally laid out to compensate about AC current asymmetrically. The AC component, as
90 % of the no-load reactive power. A power factor indicated with broken lines, attains its peak at the
of approximately 0.9 at full load and 0.95 at partial moment when the short circuit is initiated (length a).
load is thereby achieved. The capacitor is connected The amplitude of the AC component decreases
in parallel to the motor terminals and switched on rapidly at first and then decreases slower until it
and off together with the motor. With larger capacitors finally attains a constant value. This decrease of
damping reactances or resistances are required to current is caused by the armature reaction of the
limit the current resp. the steepness of the current short circuit current which weakens the alternator
increase.
field and therefore reduces the voltage at the terminals
of the alternator.
The question whether or not to generate power at It must be realized that the peak value of the short
high, or low voltage is of great importance for the circuit current is responsible for the dynamic stress of
electrical supply system. The limit to which 400 Volt electrical equipment. The dynamic power developed
alternators can be manufactured is 5,000 K.VA. The is in direct relationship to the square of the current.
rated current of this alternator is 7,250 amps for As the time constant of circuit breakers is generally
which no circuit breakers are available, the maximum between iOO and 250 milliseconds and the peak value
size of low voltage switchgear being 6,000 amps. of the short circuit current is attained within 10
It would of course be possible to cope with 12,000 milliseconds, there is practically no possibility of
amps by connecting two circuit breakers in parallel avoiding its full dynamic effect on electrical equipment
but the rupturing capacity cannot be doubled by this in circuit, other than by using HRC fuses. These have
measure. In the case of a fault, one of the circuit the ability to interrupt the circuit before the peak
breakers will always open a fraction of a second value is reached (this solution is however, only
earlier than the other, leaving only one breaker to possible with Amperages within values of fuses on
deal with the full rupturing power. The rupturing the market). The peak value of the asymmetrical short
capacity, as a matter of fact, is the primary consider­ circuit current is therefore decisive foi the conation
with regard to the choice of high or low voltage struction of switchgear, busbars, current trans­
for a system and before going on, a few basic details formers, etc. and consequently for the plant costs
on this subject are given.
involved.
The severest conditions prevail in the case of a 3phase
short circuit with the alternator fully excited,
if occurring at the moment of the voltage passing
through zero. These conditions are illustrated in
Fig. 4, where the behaviour of a short circuit current
over a period of time is shown.
As may be seen from the foregoing, it is not the
asymmetrical peak short circuit value, but the current
value, at the moment of interruption, which defines
the layout of circuit breakers. As the DC component
has faded away within 100 to 250 milli-seconds,
i.e. 10 to 25 half cycles, only the AC component
need be taken into account. The latter however,
has also decreased within this period and the greater
the time constant of the circuit breaker, the smaller
the current is. Consequently the actual current at the
time of separation, of contacts generally is smaller
than the maximum symmetrical short circuit current.
It can be calculated by multiplying the maximum
symmetrical value, with a correction factor. The
circuit breaker is then laid out to meet this value. By
multiplying the current which prevails at the moment
of separation of contacts, with the voltage across
phases and \/3, the rupturing capacity is arrived at.
This is only a theoretical figure however, as the two
values do not occur at the same time.
In the same way as the actual current at the time
of switching may be calculated, by multiplying the
maximum symmetrical short circuit current by a
correction factor (which is dependent on the conditions

166 Proceedings oj The South African Sugar Technologists* Association- March 1966
of the network system), the asymmetrical peak short
circuit current value can be calculated from the
maximum symmetrical short circuit current. It is not
necessary to go deeper into this subject and for the
purpose of this paper it suffices to mention that in the
severest practical case, the maximum peak value of the
asymmetrical short circuit current can attain 2\ times
the maximum r.m.s. value of the symmetrical short
circuit current.
As the foregoing shows, the max. symmetrical short
circuit current value can be considered as a key for
short circuit calculations. The factors which define
this value are the service voltage and the total impedance
of the system between the points of generation
and the fault. For rough calculations it suffices to use
only the reactances of the components involved in
the short circuit. These components are the alternators,
the transformers, the large motors, the
busbars and the cables. A simplified method of
calculating short circuits, is to use referred values
such as relative subtransient reactances. The lowest
value for turbo alternators, running at 3,000 r.p.m.
is 12 °0 and for salient pole alternators, running at
1,500 r.p.m., which are commonly used in Sugar
Factories, is 17%. For transformers up to about
1,600 KVA, 6 % is the usual.
In the following, the influence of both types of
alternators and also the influence of transformers and
cables on the maximum symmetrical short circuit
current is shown.. In Fig. 5, four different power
supply and distribution systems are given. In each
case the size of the alternator is 5,000 KVA.

Proceedings of The South African Sugar Technologists' Association—March 1966
In each case the conditions which would result
from supply by a turbo alternator on. the one hand
and a salient pole alternator on the other, are investigated.
In examples 1 and 2, the alternators feed the
busbars directly while in examples 3 and 4, transformers
are between the alternators and the consumers.
In examples 1, 3 and 4, the alternators supply
6 KV, while in example 2, the alternator generates
400 volts. In example 3, the 4 transformers of 1,250
KV each are connected in parallel on the 400 Volt
side, while in case 4 they supply separate systems. In
all 4 cases, 350 KVA is supplied to a consumer group
by a cable with a length of 300 feet. For examples 2,
3 and 4, a cable of 3 X 240 mm 2 is just big enough,
while for example 1, a 6 KV cable of only 3 x 10
mm 2 suffices.
It is now assumed that a short circuit occurs twice
in each case, once at the main busbars and once at the
end of the 300 ft. cable. The result is shown in Fig. 6.
The back columns represent supply by 3,000 r.p.m.
turbo alternators, the front columns by 1,500 r.p.m.
salient pole alternators. The left column in each
group shows the symmetrical short circuit current
which occurs at the main busbars, while the right
column in each group shows the symmetrical short
circuit current at the end of the 300 ft. cable.
Let us compare conditions as indicated by columns
a, i.e. the short circuit currents occurring at the main
busbars. As may be seen from Fig. 6 the differences in
heights of the left back and left front columns, i.e.
the short circuit currents of both types of alternators,
are considerable in example 1 and example 2 (29 %),
in example 3 the difference is smaller (24%) and in
example 4 yet smaller (12%). A comparison of the
left back columns of each group shows that the
short circuit current at the main busbars decreases
from example 1 to example 4. If the currents of
examples la and 2a are considered to be 100%, the
current of 3a is 66.7% and of 4a is 33.2%. (As
indicated on the scale on the right). Comparing the
left front columns of each group one observes that
the short circuit current at the main busbars, if la, and
2a, is 100% becomes less in cases 3a and 4a, where
the percentages are 73.7% and 41.4%. Therefore
the application of transformers proves to have a
greater effect on the short circuit conditions with
3,000 r.p.m. turbo alternators than with 1,500 r.p.m.
salient pole alternators.
Next the conditions as indicated by columns b,
i.e. the symmetrical circuit currents occurring at the
end of the 300 ft. cable are compared. As may be
observed, both the front and back columns show the
following:
The HT cable in case lb, has only a small damping
effect namely from 100% to 95% resp. 97%. The LT
cable however, damps the current quite considerably,
most of all in our comparison in case 2b, from 100%
to 24.2% resp. 31.7%. In example 3b, it is from
100% to 33.6% resp. 39.5% and in example 4b,
only from 100% to 52.1% resp. 55%. In this comparison,
the values of the left columns are considered
as being 100%. The values of b, are smallest in
167
example 4, although they are not much less than those
of examples 2 and 3. On comparison of the extreme
values, namely the ones of examples 2a and 4b, it
may be observed that the short circuit current of the
back columns is reduced from 100% (69.4 KA) to
17.3% (12 KA), while that of the front columns is
reduced from. 100% (49.1 KA) to 22.6% (11.2 KA).
The conclusion of this comparison for the LT
equipment is, that the main busbars present the primary
problem with regard to short circuit conditions
as even small lengths of cables reduce the short
circuit currents by 45 % to 75 %. However, this must
be considered with great care as the permissible
short circuit current values of motor control centres
and sub-distributions are relatively low (as will be
shown in a further part of this paper). Furthermore,
if a second cable is connected in parallel with the
existing one, the short circuit conditions immediately
become very unfavourable. Another conclusion is that
the layout of the electrical equipment before the LT
main distribution has a considerable influence on
the resulting short circuit current. By suitable measures
it is in fact possible to reduce the short circuit current
by 60-70% before the main LT distribution.
In Table II, the maximum permissible symmetrical
short circuit values for switchgear of a certain make
are given:
Cast iron clad switchgear of the make under discussion
can withstand 12 KA. Sheet steel clad switchgear
can withstand maximum symmetrical current
values up to 40 KA, depending on the construction
of the busbars and the copper cross sections. With
sectionalized panels, they can be made to withstand
up to 50 KA.
As may be seen, the permissible current values are
relatively limited, particularly with sub-distributions.
The maximum value for main distributions is 50
KA, but it is of course possible to manufacture LT
switchboards for a greater maximum symmetrical
short circuit current. This however is only possible
by using expensive technical measures, which do not

168
appear to be justified for the purpose. BSS 936/1940
recommends that the maximum short circuit current
in any low voltage system should be limited to 44
r.m.s. symmetrical KA.
When utilizing switchgear as specified above, 60
amp circuit breakers can be used in plant, where the
maximum short circuit current does not surpass 5 KA.
Up to a value of 15 KA, circuit breakers of 100 amps
to 400 amps may be installed, while 600 amp breakers
are required for plant where a maximum of 30 KA
may occur. Where a maximum symmetrical short
circuit current of 45-60 KA may be expected, only
circuit breakers of 1,000-6,000 amps can be utilized.
The only solution for protection of equipment
where the maximum symmetrical short circuit current
surpasses 60 KA, is to install fuses, which prevent
the short circuit current rising to its full value. This
however, is only possible for circuit breakers up to
1,000 amps, being limited by LT fuses. Modern
HRC fuses have melting periods of less than 5 milliseconds,
in other words they respond before maximum
current is attained, as the peak value of the first half
cycle is reached only after 10 milliseconds. Therefore
the thermal and mechanical effects which would
otherwise arise are not present. In all cases, contactors
must be backed up by fuses as they are not
designed to withstand short-circuit stresses. As such
fuses can be installed where the short circuit value
surpasses 100 r.m.s. symmetrical KA, which means
that the peak asymmetrical current value would be
greater than 250 KA, they can be considered as
capable to deal with any short-circuit conditions.
As there are no HRC fuses for currents greater than
1,200 amps available, the maximum symmetrical
current of plant where circuit breakers larger than
1,200 amps are to be installed must not surpass the
mentioned 40-60 KA level.
The limiting factor therefore is less the switching
equipment than the distribution switchgear, primarily
the sub-distributions. In these distribution boards
the circuit breakers are seldom greater than 600
amps, which can be effectively protected by fuses.
Where the difficulty arises is the main distribution
switchgear, where larger circuit breakers are required
and back up protection by means of fuses is not
possible.
In case 4b, of Fig. 6, the short circuit current at
the end of the 300 ft cable is 11.2 KA, resp. 12 KA
which would still be permissible for normal cast iron
clad distributions. In cases 2 and 3, where 14.3 KA
to 16.8 KA results, this type of distribution must be
of a mechanically strengthened design. For the main
distribution switchboard the short circuit current
values of 20.3 KA resp. 23 KA, involved in case 4,
present no problem which also applies for case 3,
where a salient pole alternator supplies the power and
36.2 KA results in the event of a short circuit. However,
in example 3 where a turbo alternator supplies
power and the short circuit current is 46.2 KA,
proper measures must be observed. The same would
apply for the 1,500 r.p.m. salient pole alternator in
example 2, where 49.1 KA is very close to the limit
Proceedlins of the South African Sugar Technologists' Association—March 1966
of circuit breakers required. In all cases, sectionalized
sv/itchgear with a heavy rupturing capacity would be
essential.
Conditions of example 2 with 69.4 KA, where a
turbo-alternator is the power source, is however
completely unsuitable, as special switchgear and
switching equipment would be necessary to handle
the resulting high short circuit current.
With due consideration to the problems of short
circuit currents as discussed, it may be concluded that
5,000 KVA can be considered a limit for 400 Volt
s. pole alternators. As already mentioned in an earlier
part of this paper, the rated current of a 5,000 KVA
400 Volt alternator (7,200 amps) cannot be handled
by the largest available LT circuit breaker. Therefore
4,000 KVA can be considered a maximum size for
saiient pole alternators. In conclusion it results that
salient pole alternators are limited to 4,000 KVA by
the rated current and turbo-alternators to 3,500
KVA by the rupturing current. It must be noted,
however, that the aforementioned calculation is based
on a tension of 400 Volts. In the event of 500 Volts
being used, the limit of salient pole alternators is
5,000 KVA, the rated current being 5,770 amps, i.e.
still below 6,000 amps, where a circuit breaker to
handle this current is available. The limit for turboalternators
of 500 Volts would in turn be 4,700 KVA,
the short circuit current being the limiting factor.
If, however, more than one alternator feeds the
system the total KVA must not surpass 5,000 KVA
with salient pole alternators and 3,500 KVA with
turbo-alternators (with a 500 Volt system, 6,600 KVA
for salient pole alternators and 4,700 KVA for turboalternators).
If power is fed into the system from
other sources, this must also be taken into the calculation
as it reduces the possible generation KVA
considerably, depending on the rupturing capacity.

Proceedings of The South African Sugar Technologists' Association—March 1966 169
As shown, the limitations are quite low and the
risk of surpassing them should under no circumstances
be taken. The only technically sound solution,
therefore, is to generate power at high tension. It is
general practice on the continent to use 6,000 Volts
and the following examples and calculations are
based on this voltage. However, the resulting conditions
practically apply for 6.6 KV which would be
preferably chosen in this country. If a higher voltage,
say 11 KV, is used the possibility of supplying large
motors directly is reduced to quite an extent. For
insulation reasons the minimum limit for 11 KV
motors is about 400 KW, while it is about 150 KW
for 6.6 KV motors.
In Fig. 7, three different examples are given to
indicate the minimum sizes of HT cables required to
withstand the short circuit conditions which may be
expected.
In all three examples the voltage of the busbars
from which a cable is fed is 6 KV. As may be observed,
the time constant of the circuit breakers (the period
taken from initial switching operation to the moment
of contact opening), is decisive for the size of the
cables, as a prolongation from 100 to 200 milliseconds
means that the next larger cable size is required. In
consequence a doubling up of the generation capacity
and thereby of the short circuit current calls for a
larger cable size, as indicated in example 2. If a
further supply source feeds into the system, as shown
in example 3, it influences the required cable size to a
considerable extent. As the impedance voltage of the
transformer is only 6 %, the additional supply source
boosts the rupturing capacity to twice the value of
example 2. Consequently the cable required, must
also be twice the size of the cable of example 2.
Fig. 7 indicates the power which the various cables
used in examples 1-3 could carry if subjected only to
normal continuous operation without being subjected
to short circuit stresses. At the same time it clearly
shows the extent of overdimensioning which is
required to ensure short circuit stability. For instance,
in order to transmit a load of 350 KVA, as shown
in Fig. 5, a cable with across section of 6mm 2 would
suffice to handle the current, but a far larger size must
be chosen for reasons of short circuit stability. The
cross section would range between. 10 mm 2 and 50
mm 2 , depending on the impedance involved, i.e.
TABLE III
Comparison of prices of HT and LT motors of different ratings
2-8 times the size actually required if chosen to carry
the normal current only. The comparison of prices
for PVC cables shows that a maximum of 3 times the
price would be required, in the example given, if the
size 10 mm 2 is considered as costing 100%. Similiar
aspects would of course also apply for LT cables but
as the currents involved are considerably greater the
minimum cross sections required for stability reasons
are usually given anyway, when chosen to cope with
the normal current.
It appears from the foregoing that HT motors
with a rating as low as 150 kW are not a cheaper
solution than LT motors. This question is investigated
in Fig. 8.
FIGURE 8
Comparison of LT and HT motors. Schematic Diagrams
The motor is situated 300 ft. away from the switchboard.
The circuit breakers and cables between the
alternators and the HT busbars in this comparison
can be neglected as they are required in both examples.
For the HT part therefore, only the price of the
circuit breakers, with overload protection, and of 300
feet of cable need be taken into account, while for the
LT part the additional prices for circuit breakers on
the primary and the secondary sides of the transformer
and the portion of the transformer costs, corresponding
to the motor power, will go into the calculation.
In Table III, a comparison between a 180 kW motor
with a rating of 230 KVA (20% of the transformer
rating) and two motors of 240 kW and 500 kW is
given. The 240 kW motor is rated 310 KVA (25%
of the transformer rating) while the 500 kW motor
is rated 610 KVA (50% of the transformer rating).

170 Proceedings of Hie South African Sugar Technologists' Association--March 1966
The result may be seen in the first line of Table III.
In each example the LT solution is considered as being
100%. While the HT solution is 11 % more expensive
with the 180 kW motor, the HT solution is cheaper
by 5% with the 240 kW motor. The HT solution
is considerably cheaper with the 500 kW motor,
namely by 35%. Both HT and LT solutions are about
equal in price with a motor of 200 kW and the larger
the motor is from there on, the greater is the difference
in price to the advantage of HT.
The cost of components are broken up and their
percentages of total prices are given. The prices of
the transformers are about 37% of the total costs in
the LT solution. This is the reason for the LT solution
becoming more and more unfavourable with the
increase in size of the motor. The portion of the LT
cables is between 12% and 17% and must not be
overlooked, while the portion of the HT cables to
the HT motors is only 2 % to 4% and therefore nearly
negligible. This calculation clearly shows that it is
still economical to use HT as long as the motors are
not smaller than 200 kW.
This paper has dealt with fundamental considerations
of the electrical supply system of a sugar
factory. It has in fact only touched the many problems
involved in designing an entire electrical system so as
to give the most practical and economical layout.
Summary
General aspects of designing the electrical equipment
of a modern sugar factory to arrive at most
economical and practical solutions are discussed.
Emphasis is placed on the importance of power
involved and short circuit conditions occurring. The
influence of the physical layout of the factory with
regard to the electrical system is mentioned.
The dimensioning of alternators, switchgear, cables
and motors is discussed in detail.
References
Scherer, W. (1963). Elektrische Energieversorgung von Zuckerfabriken.
Zucker, 16. Jahrgang Aug. 1963, Heft 15 und 16.
Electrical Engineers Reference Book (1955). Section 7/45.

Proceedings of The South African Sugar Technologists' Association—March 1966
PHOSPHORIC ACID AS AN AID TO CLARIFICATION,
AND OBSERVATIONS ON LIMING TECHNIQUES AND
MUD VOLUMES
The contents of this paper are based on experimental
and practical observations collected over the past
years.
(1) An Investigation into the Application of Laboratory
Clarification Tests and a Comparison of the Different
Liming Techniques
Following along the lines of work done by C. W.
Davis of the C.S.R. Co., Australia, a series of experiments
were carried out to determine the effects of
different liming techniques using a laboratory settling
apparatus.
This apparatus consisted of two tubes of 1.5 in.
I.D. by 36 in. suspended in individual water jackets
through which hot water was circulated. The maximum
temperature of the circulating water was 80° C,
which caused initial convection currents when the hot
juice was poured into the settling tubes.
The criteria used in the settling tests were settling
rate and clarity, i.e., the percentage transmission
through a 1 cm. cell at a wavelength of 0.615 microns.
Sampling of mixed juice was done at the scales and
the juice was found to yield a higher clarity on standing.
Davis feels that this is due to enzymatic break
down of Starch and the release of bound Phosphate.
The maximum time lag allowed was six minutes from
the scales to processing in the Laboratory.
The Test Procedure for processing of the Juice was
as follows:
(i) Juice was sampled at the scales and used within
6 mins.
(ii) Approximately 375 mls was treated to varying
lime sequences.
(iii) The lime was added in the vortex of a stirrer
and addition controlled by pH meter.
(iv) The sample was boiled for 3 mins. to ensure
complete degasification.
(v) The boiling slurry was poured into a preheated
settling tube and allowed to settle. The time of
settling was recorded and the final mud vol.
percentage calculated.
(vi) After a certain time (depending on whether a
coagulent was used or not) the juice was
sampled and tested for clarity and impurities
present.
Davis' observation on the mixing of lime and juice
was found to be true and the addition of lime to the
surface yielded juice clarities which averaged 4, whilst
when added in the vortex of a stirrer this improved to
an average of 11.
By G. G. CARTER
The comparison between tubes and factory using
the same process techniques, showed that the clarity
obtained in the tubes was always slightly inferior due
to the thermal difference between settling tube and
jacket.
The tubes yielded a clarity of 9, whereas the comparable
Factory performance was on average 12, at
the time when comparison was made.
Once the initial procedure had been established a
series of tests were begun to establish what liming
technique gave the best results as far as settling rate
and clarity were concerned.
In order to obtain the bulk of data required to
substantiate whether one liming technique was better
than another, use was made of coagulants to cut the
settling time from 3 hrs. to 30 mins. Having established
the best method the coagulant was omitted
when final comparisons were made to ensure no bias.
From over 100 tests conducted the best sequence of
liming was found to be heating to 160° F., allowing
171
The most important part of the procedure was to
establish whether the settling apparatus gave comparable
results to the factory or not, and whether there
was a correlation between the two laboratory settling
tubes. Initially this latter point was markedly different
until the tubes were fed in parallel rather than in
series by the circulating pump.
The test revealed that there was no difference
between the settling rates in the two tubes. See Table I.
Table I

172 Proceedings of The South African Sugar Technologists' Association—March 1966
to stand for 10 mins. to emulate the starch tanks then
liming to 7.6 pH and heating to boiling. The clarity
of juice so obtained was superior to any other technique
used. In settling rate it was not the fastest, the
addition of lime after heating (such as we do here at
Tongaat) having the benefit of reforming the flocs and
resulted in the most rapid settling of the tried liming
techniques.
This reformation of flocs and the consequent increase
it has on settling rate does tend to give evidence
to the theory of floc break up.
The difference in clarity between a straight heat
lime heat sequence and the method here at Tongaat
lay in the haze which the latter juice had and the
darkness of the juice it produced. The CaO content
of mill juices averaged 414 ppm., whereas those in
the laboratory were only 337 ppm. The author feels
that this could in part explain the dark juice colour
since an excess of lime (in our case not properly
reacted) can cause mellanoid darkening. The average
clarity's for clarified juices of different techniques are
given in Table II.
Table II
The addition of Phosphoric acid was tried and improved
the clarity of the Juice so that clarities of 28
were common, whilst figures as high as 41 were
recorded. This opened a new line of investigation on
the use of Phosphoric acid as a clarifying agent.
Table III
(2) The Significance of Mixed and Clarified Juice
Phosphate Contents in Clarification
At Tongaat the Phosphate contents of the Mixed
and Clarified Juices are recorded daily. The analysis
is done on a twenty-four hour composite by the
Official Method. The average monthly results for the
preceding five years are recorded in Table III.
From the table it may be seen that there has been
a gradual decrease in P2O5 in Mixed Juice over the
past four years. This decrease in Phosphate content
is possibly due to the fact that cane is now cut earlier
than in previous years. The lowering of P2O5 level has
resulted in increased difficulty in clarification.
This age-old problem of mud settlement prompted
the investigation of Phosphoric Acid as an aid to
settling — both to improve the settling rate and to
help thicken muds.
From laboratory tests the addition of progressive
increments of Phosphoric Acid to Mixed Juice prior
to clarification showed gains in the rate of mud
settling. However, increases in mud volume above a
limit of 350 ppm. resulted. These facts are shown in
Table IV.
From the results the best P2O5 limit appeared to
be at 330 ppm. of P2O5 in Mixed Juice.
Another point that arose from the tests, was that
the colour and clarity of the acid treated juices was
excellent, and according to Honig 1 the Phosphoric
Acid is beneficial to the removal of silicic acid, iron
salts and nitrogen-containing non-sugars.
It was decided on the results of experimental work
that a plant scale trial should be carried out. This
trial period was given for 42 hours. The dose of H3PO4
added was set at 100 ppm., since this would increase
the P2O5 level in Mixed Juice to the desired 330 ppm.
mark.
Average Monthly Phosphate contents of Mixed and Clarified Juices in ppm.

Proceedings of The South African Sugar Technologists' Association—March 1966
Table IV
173

174
The Factory Trial
On 25th November 1965 at 11.30 a.m. the Phosphoric
Acid treatment was started by adding 100
ppm. of acid to the heated juice prior to secondary
heating. The process conditions at the time were cold
liming to 6.5 pH heating to 160° F. followed by the
addition of Phosphoric Acid and then heating to
boiling before adding the final dose of lime to bring
the pH of clarified Juice to 7.3. This addition ended
at 9 p.m. on 26th November, 1965.
The results of the test are given below together with
all analyses thereon:
Table V
Analysis of Mixed Juice Prior to Test
Table VI
Analysis of Clarified Juice Prior to the Test
Table VII
Analysis of Mixed Juice during the 1st Stage of the Test
Table VIII
Analysis of Clarified Juice during the 1st Stage of the Test.
Proceedings of The South African Sugar Technologists' Association—March 1966
From the results the tests clearly indicated no substantial
improvement in the removal of impurities
known to cause a decrease in filterability. The author
felt that this was due to the fact that the acid addition
was ineffectual in producing a good floc when cold
liming was being done, resulting in excess P2O5 in
the clarified juice, i.e., 69 as against 52.
Thus a run was made for 9 hours on Juice which
was run through the starch tanks and then had 100
ppm. of phosphoric acid added prior to hot liming.
The method found to be so successful in the laboratory.
The phosphate addition was started at 9 a.m. on
1st December 1965 and run until 6 p.m. of the same
day.
The results are as given below in Tables IX and X.
Table IX
Analysis of Mixed Juice during the 2nd Part of the Trial
Table X
Analysis of Clear Juice during the 2nd Part of the Trial
Discussion of Results
From the figures shown in Table IV it may be seen
that the addition of Phosphoric Acid is beneficial to
the mud-settling rate and to the compaction of mud
volume. The results of the factory trial show that there
is an improvement in phosphate removal of 4.5 per
cent, silica removal 30 per cent and calcium oxide
removal 36 per cent. This removal of inorganic matter
no doubt accounted for the better clarity figures
recorded and must result in better sugar quality. It
therefore seems likely that phosphoric acid can prove
an effective aid to the defecation process employed in
Natal. However, it is desirable for more data to be
obtained, preferably on the factory scale, before great
significance can be attached to these results.
(3) Final Mud Volumes and their Significance in
Clarifier Operation
Using the apparatus for settling described in this
paper, daily mud settling tests have been recorded at
Tongaat for the last five years, the significant figure
ESS^^MH

Proceedings of The South African Sugar Technologists' Association—March 1966 175

176 Proceedings of The South African Sugar Technologists' Association—March 1966
at the end of 3 1/2 hrs. being the final mud volume
percentage of original volume.
From the daily mud settling tests the monthly final
Mud Volume percentages have been plotted on Graph
I.
Here it is seen that all the curves with the exception
of 1965 follow the same general trend with a gradual
increase in mud volume over the winter months and
then a decrease as the summer months arrive. This
is in contrast to the general practical difficulties which
are experienced when the rainy months come.
It will be seen that 1965 in contrast to the preceding
years started differently, and then as the season progressed,
became more normal in its characteristics.
This resulted in conditions being such that, whereas
all clarifiers were operating prior to October after this
month, times came when two Bachs were shut down.
What is the purpose of these mud volume figures?
The answer is that from them clarifier capacity can
be calculated using the Kynch construction as illustrated
by C. W. Davis. 2
See Appendices I and II and Graphs I, II, III and
IV where actual cases are taken for conditions at
Tongaat.
It will show the estimated clarifier capacity in September
and November, which bears out the statement
on conditions mentioned in a preceding paragraph.
Estimation of Required Capacity of Clarifier
A common method for predicting area requirements
is based on the mathematical analysis by Kynch.
Using this, Talmage and Fitch have derived the following
procedure for estimating industrial requirements
from laboratory tests:
(a) Plot the height of the settling interface against
time, and draw a horizontal line at a height
(Hu) corresponding to the desired solids concentration
of the underflow. Graph II and III.
(b) Draw the tangent to the curve at the critical
point to interesect the underflow line at time
Tu (in hours). (The critical point is defined as
the point at which the settling interface goes
into compression and the flocs at the surface
receive mechanical support from the neighbouring
solids.)
The unit settling area required is then given by the
relationship:
Area = Tu/Ho (sq. ft./cu. ft. of feed/h.)
where Ho = Initial height, in ft., in the settling
tube.
The determination of the critical point, by observation
of the settling curve, limits the accuracy of this
This is thus a very good method for checking on
whether settler capacity is great enough for existing
conditions.
Discussions and Conclusions
From the information presented in this paper the
following points are made:
(i) It has been found that the method advocated
by Davis for the estimation of clarifier capacity
in a raw sugar factory applies well at Tongaat.
(ii) The liming technique which may be expected
to yield the clarified juice with the highest
clarity is a straight heat lime heat sequence
with proper process lime mixing.
(iii) Addition of phosphoric acid in increments up
to a total P2O5 content of 330 ppm. in mixed
juice increased the mud settling rate, the clarity
of the clarified juice and decreased the final
percentage mud volume.
Acknowledgments
I should like to thank Mr. Boyes for the help he
has given me and wish also to thank the Tongaat
Sugar Company for permission to publish this Report.
References
1. Honig, P., Vol. I, Inorganic Non-Sugars, Chapter 9, 341.
2. Davis, C. W. (1958) Development and Application of a
Laboratory Clarification Test. 18 I.S.S.C.T.
APPENDIX I
method. In this region the slope of the curve changes
rapidly and any slight mislocation can result in major
discrepancies in Tu.
Further expansion of Kynch's mathematical approach
indicated that a plot of settling height against
time on logarithmic co-ordinates should show a discontinuity
at the critical point. Graph IV. The
critical point so defined by this discontinuity was then
employed in the conventional Kynch construction as
shown in Graph II and III.
For Tongaat for the months of June, July, August
and September, the critical point was 70 minutes and
Tu was ± 140 minutes. The Initial Height of the tube
was 60 cms., thus the settling area required was:
Thus for 250 tons per hour plus 20 per cent recycling
of Oliver filtrates this would be :

Proceedings of The South African Sugar Technologists' Association—March 1966
Clarifier capacity required in November.
Here ratio of sq. ft./cu. ft. of feed required is cut to
177
Total area available was thus 9,880 sq. ft., which
shows that under these conditions the plant was at
full capacity theoretically during the four months.
APPENDIX II

178
Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 179

180
Dr. Graham: With the heat lime heat sequence
advocated in the paper, retention of the juice at 160°
for ten minutes would probably result in appreciable
loss of sucrose. If this were modified so that the first
liming took the pH up to say 6.0 before holding in the
retention tank would the clarification be as efficient
as was found with the conditions used in the tests?
Nicholson said in 1959 that the CSR Co. was not
using this method for starch removal only because
at the time they did not have adequate pH control
of mixed juice.
Mr. Carter: I do not have sufficient evidence to
answer that question. I fail to see why it should make
any difference to the clarity of the juice but it will
certainly mean a saving in inversion losses.
Mr. Boyes: Tongaat lost very little sucrose in the
starch removal process because returned filtrate was
added to the mixed juice increasing the pH to 6.
With a retention time of ten minutes the percentage
loss of sucrose entering the removal process was
approximately 0.12%.
Mr. Buchanan: The inclusion of 20% recycled
filtrate in the feed to the clarifier (as in the appendix)
Proceedings of The South African Sugar Technologists' Association—March 1966
is incorrect for sizing purposes since the recycled
filtrate contains only a small amount of solids.
Mr. Angus: On page 1 of the paper it states "the
tubes yielded a clarity of 9 whereas the comparable
factory performance was 12".
The reason for this was probably that in the
laboratory testing was in batch as against the continuous
clarification that applies in factory clarifiers.
I wonder also about the effect of seeding old sludge
on to new sludge that is forming. Is this relied on for
mixed juice clarification? I think a pilot plant should
be used to carry out tests from jars up to factory
clarifiers.
Mr. Carter: I must stress that on average the clarity
in the laboratory tubes was worse than in the factory.
Mr. Boyes: When Tongaat and Umfolozi tried to
clarify filtrate from the Oliver filters separately it
was found that the clarity of the raw juice dropped.
On stopping filtrate clarification and returning the
filtrate to the raw juice the clarity returned to normal.
I think the reason is that the filtrate contains tiny
granular particles of precipitate which assist in the
formation of floc by seeding the precipitate formed.

Proceedings of The South African Sugar Technologists' Association—March 1966
VACUUM PAN CONTROL
Introduction
Consideration has been given to the effect of vacuum
pan circulation on sugar quality — particularly with
reference to filterability. It has been suggested 1 that
the introduction of a mechanical circulator, or stirrer,
would enable the process to be carried out within
closer temperature limits, so that the occlusion of
undesirable substances could be prevented. In considering
the possible installation of a circulator, the
likelihood of the massecuite short-circuiting into the
eye of the impeller presented itself. It was considered
that a movable sleeve, located within the downtake
and rising as the level of the massecuite in the pan
rises, would ensure that the heated massecuite could
be constrained to approach the surface more regularly
and short-circuiting would thus be avoided. A similar
idea was proposed by Waddell. 2 Arrangements were
therefore made in. 1964 to install such a sleeve in an
"A"-massecuite pan at Illovo. The experiment however
proved inconclusive as the conditions under
which sugar was boiled were too erratic, and examination
of the data obtained showed that too many
variables existed for a complete analysis of the results.
It was consequently agreed upon that before continuing
with any further experiments along similar
lines, an examination should be carried out into pan
boiling as normally practised in Natal to obtain more
detailed information concerning techniques, circulation
criteria and other phenomena.
An investigation was accordingly arranged at the
Gledhow factory, where an "A"-massecuite seed pan
was made available. It was originally resolved to
record as many of the variables which occur during
the boiling process as possible, but due to the shortage
of time only a few experimental positions were chosen.
This report therefore outlines the progress made to
date, along with some of the more interesting results
obtained.
Equipment
For the measurement of temperature, various
methods were possible, but it was proposed that as
the temperatures had to be measured at so many
points, as was the original intention, that an electrical
method was the most obvious. All temperature
measurements were therefore recorded by a Yew
twelve point recorder, which was adapted by additional
circuits and stepswitches to be able to select
the range or alter the sensitivity of the instrument as
required. An accuracy of 0.1° C. was decided upon
as being sufficient. Three positions for the measurement
of massecuite temperature were chosen, being,
(1) just beneath the calandria, (2) just above the
calandria, and (3) just below the recommended strike
level. The exact positions can be seen in figures 1 and
2. The temperature measurement was carried out by
platinum resistance thermometers which were sealed
PROGRESS REPORT No. 1
By D. H. JONES AND D. E. WARNE
181
into brass capsules and mounted into their respective
positions in a supporting pipe. The leads from the
thermometers were then passed along this pipe and out
through the pan wall to the recorder.
With the same recorder, the conductivity of the
massecuite was measured by placing an 8 volt AC
current through a pair of electrodes mounted in the
pan. Two pairs of identical electrodes were utilized,
being situated (a) in the conventional position in the
bottom of the pan, and (b) at a point 18 ins. above
the calandria in the same vertical plane. The exact
location of the electrodes can be seen in figure No. 1,
while the positioning of the electrodes with respect to
the steam inlet can be seen in figure No. 2. The
measurement of conductivity was achieved by placing
a resistance in series with the electrodes and rectifying
any voltage drop across this resistance with a metal
rectifier. A suitable proportion of this rectified DC
voltage was then measured by the recorder. This conductivity
circuit was built together with the recorder
and all controls into a neat portable console (see
photograph) all construction and design being carried
out by J. Bruijn and N. Bowes of the staff of the
S.M.R.I.
Vacuum control was effected by a Foxboro absolute
pressure recorder/controller, operating on an 8 in.
diameter Fischer pneumatic control valve. This valve
was positioned in a bypass around the manual control
valve on the existing 14 in. cooling water delivery pipe
to the pan condenser.
The pan itself was a modern two diameter double
bottomed pan, having a conventional calandria with a
central downtake. Figure No. 1 shows the principal
dimensions.
Procedure and Results
The procedure followed during the investigation,
was to commence the recording of temperatures as
the syrup inside the pan was being concentrated. This
enabled the correct range to be chosen and allowed
time for the sensitivity of the instrument to be checked.
Thereafter, the recorder automatically recorded both
temperature and conductivity values every minute.
Further, a complete record of all the operations performed
by the pan boiler was kept relative to time,
so that at the completion of a boiling all alterations
to the boiling procedure could be compared with the
measurements made by the recorder.
At the outset of the investigation, it was resolved
merely to record such measurements as were made
available during the normal course of boiling operations
as practised at Gledhow. It soon became apparent
however that the results obtained would be very
similar for each boiling, and it was therefore agreed
upon to try to control at least one of the variables
which affect the process, namely the absolute pressure.
Consequently an automatic control of the cooling
water to the pan condenser was effected.

182 Proceedings of The South African Sugar Technologists' Association—March 1966
As previously mentioned, the investigation was
begun rather late in the season and is therefore by no
means complete. However, even during the short
period of time available some most interesting results
have been obtained.
(a) Temperature Measurement
By measuring the massecuite temperature at three
different levels, the results produced have shown that
it appears as if the circulation in the pan tested was
not in accordance with the simple theory generally
accepted. Unfortunately temperatures were measured
in one vertical plane only, but even so the effect was
still noticeable. In Graphs 1 and 2 it can be seen that
the temperature recorded beneath the calandria is
very close to that recorded just above the calandria.
It seems, therefore, that either a considerable proportion
of the vapour formed is not liberated to the
vapour space, but is entrained by the viscous medium
and forced under the surface in the downtake, or this
elevation of temperature occurs by short-circuiting of
the hot massecuite to the cold downtake stream. This
short-circuiting effect could be caused by sub-surface
flow or possibly by mixing at the interface of the two
streams. It will also be noticed that the difference in
the temperatures between positions (1) and (2) (see
Figure 1) becomes progressively smaller at the end of
the boiling until they actually cross over, thus indicating
a higher temperature for below the calandria
than for above! This alteration is common to all the
boilings carried out during the investigation, being
more pronounced in some than in others. A more
detailed record of this occurrence was therefore kept
and it was found to occur when the level of the massecuite
was approximately 4 ft. 6 in. above the calandria.
At Gledhow this particular pan is used as a seed pan
for the "A"-massecuites, and as no seed storage tank
was available it was common practise to boil to 6 ft.
above the upper tube plate, in order to have enough
seed for 3 strikes. On examining the manufacturers'
specifications for this pan, it was found that the recommended
strike level was 1,300 cu. ft. or 4 ft. 11 in.
above the calandria. The figures recorded at Gledhow
appear to be in agreement with the conditions observed
by Gillet, 3 who established for low grade boilings that
when the massecuite level reaches a point about 4.5
to 5 ft. above the calandria upper tube plate, the
natural circulation diminished to an objectionably
low level.
In Australia, 4 similar circulation investigations on
calandria pans have also indicated a very small temperature
rise as the massecuite passed through the
calandria. Here a value varying between 1.2° F. and
4.5° F. as the boiling progressed was recorded, with
the thermometers positioned at 60° from the steam
inlet. The particular pan under investigation enjoyed
a heating surface to volume ratio of 1.88 and a total
heating surface of 3,000 sq. ft. The suggestion was
therefore made that there was a large steam side
pressure drop across the tubes of the calandria. This
did not mean that the pan was starved of steam, but
rather that all the steam necessary to maintain a high
rate of boiling could be condensed in less than the
total heating surface. In other words the boiling was
mostly occurring over the steam inlets. Subsequent
experiments in Australia 5 have actually proved this,
where temperature measurements taken in regions
remote from the steam inlet have indicated even
smaller temperature rises and often negative results
were obtained.
At Gledhow the average maximum temperature
rise across the calandria was in the order of 3.0° F.
decreasing to a negative value at the end of the boiling.
The positioning of the thermometers was approximately
30° from the nearest steam inlet. The heating
surface to volume ratio of the pan was 1.83, and the
total heating surface was 2,380 sq. ft. It would therefore
be most interesting to see the influence of the
steam distribution on this pan by re-positioning the
thermometers at 90° from the steam inlet.
(b) Absolute Pressure
The consequence of having perfectly controlled
absolute pressure is obvious from Graphs 1 and 2.
With the control of absolute pressure, and irrespective
of the boiling technique, the boiling process can be
carried out within closer temperature limits and all
temperature fluctuations are therefore kept to a minimum.
Before the investigations were actually begun,
a number of absolute pressure recorder charts were
taken during the normal boiling operations at Gledhow
(see Chart No. 1). From this record can be seen
a large number of fluctuations. Chart No. 2 is another
record of the absolute pressure, again recorded during
a normal boiling by the pan boiler. A comparison of
these two charts indicates that the pan boiler has
improved his control by merely paying more attention
to the boiling. In each case the flow of cooling water
to the pan condenser was not altered. Chart No. 3
on the other hand is an example of what can be
achieved by having automatic control.
(c) Conductivity
Measurement of the conductivity was effected by
two identical pairs of electrodes mounted in the positions
shown in Figures 1 and 2. The fact that these
two pairs of electrodes were mounted in these positions
has in itself produced interesting results. It was seen
during boiling operations that two distinctly separate
readings could be obtained. In one example of this a
deviation could actually be observed visually, where
the recorder operating on the lower pair of electrodes
was indicating that the massecuite was slack and
becoming slacker, whereas at the boiling surface it
could be seen that the massecuite was tight and
becoming tighter. On changing over the recorder to
the upper pair of electrodes, a lower conductivity value
was recorded. This stratification effect was not just
instantaneous, but required some little time before the
influence of the syrup added by the pan boiler could
be seen at the boiling surface.
This variation between two pairs of conductivity
electrodes is in fact a further measure of the pan circulation,
and has been used as such in Java by Honig. 6
(d) General
Although the various results obtained have been
reported on in different categories, it is realised that
they are all influenced by each other. The more obvious

Proceedings of The South African Sugar Technologists' Association—March 1966
Conductivity and Temperature Recorder
results were however not commented upon as this
would take up too much time.
Under the heading General Observations, one engaging
experience noted during the investigation was
the manner in which the weather influenced the boiling
operations. Cold wet days could be expected to give
a considerably lower absolute pressure for boiling,
whereas the temperature of the cooling water on hot
days was not conductive to good absolute pressures.
Any attempt to correlate boiling proceedings with
sugar quality was unfortunately unsuccessful.
183
Conclusions
Wide variations from accepted theory with regard
to circulation in a vacuum pan and allied phenomena
make it obvious that a great deal of work still has
to be done before any definite conclusions can be
reached. During the past season the investigation
carried out at Gledhow forms merely the opening
paragraph as to what must be undertaken in the
future.
The coming season will see more elaborate massecuite
temperature measurements to try to evolve a
definite circulation pattern. Attempts to determine the
most suitable position for conductivity measurements
means further electrodes must be constructed and
placed in various positions within the pan. It was also
obvious from the results recorded that many variables
occur during the process and before any deductions as
to their effect can be made, these variables will have
to be measured and suitable methods made available
for their control.
In conclusion it is evident that the programme for
next season must be extensively modified and enlarged
and can therefore be expected to produce some most
interesting results.
Acknowledgments
Acknowledgments are due to the management of
the Gledhow factory, for allowing the use of a vacuum
pan for this investigation, and also to the staff personnel
for their generous co-operation during the
tests.
References
1. Bruijn, J. The construction of two Laboratory pans.
S.A.S.T.A. 38th Congress, April, 1964.
2. Waddell, C. W. International Sugar Journal, 41 (1939) 65.
3. Gillet, E. C. Low grade sugar crystallization.
4. Technical Report No. 63. Sugar Research Institute, Mackay,
Queensland (June, 1960).
5. General Chemistry Report. Sugar Research Institute, Mackay
Queensland (October, 1962).
6. Honig, P. I.S.S.C.T. 8th Congress, B.W.I. (1953) 792.

184 Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE No. 1
Sectional elevation of Pan showing positions of thermometers
and conductivity electrodes.

Proceedings of The South African Sugar Technologists' Association—March 1966 185
FIGURE No. 2
Sectional Plan of Pan showing positions of thermometers and
conductivity electrodes.

186
Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 187

•w
4 v /
188
Proceedings of The South African Sugar Technologists' Association—March 1966
CHART No. 1
Showing vacuum recorded during boilings before
commencement of investigations.

' * •
Proceedings of The Soutli African Sugar Technologists' Association—March 1966
CHART No. 2
Showing vacuum recorded during investigation without
control of cooling water.
189

190
Proceedings of The South African Sugar Technologists' Association—March 1966
CHART No. 3
Showing vacuum recorded when cooling water is
automatically regulated.

Proceedings of The South African Sugar Technologists' Association - March 1966
Mr. Robinson: We have a "C"-massccuite pan at
Darnall equipped with a mechanical stirrer and independent
vacuum supply. The feed to the pan is
controlled by an automatic valve which is in turn
controlled by the load on the electric motor driving
the stirrer. With this arrangement we have found
that a fairly good control of vacuum can be obtained
just by controlled feeding. The temperature of the
massecuite as indicated by a pan thermometer is very
steady during boiling; unfortunately, as you have
mentioned in your paper, the temperature of the
cooling water does affect the vacuum.
Mr. Warne: I am very pleased to hear that you have
achieved such results but as most of the other factories
do not have independent vacuum systems on their
pans the automatic vacuum control as used by us
would benefit them greatly. With an automatic control
of the cooling water to the condenser it is possible to
produce boilings at exactly the same absolute pressures
regardless of the cooling water temperature, throughout
the entire season.
Mr. Jones: If you look at the second chart I would
like to emphasise the improvement in the pan boiling
by the operator without any assistance whatsoever,
purely because he was taking note of what he was
doing. This confirms what Mr. Deon Hulett said in his
paper "Where should the downtake be"—no matter
what you do unless you can have some sort of control
over the conditions existing in the pan, you are very
much in the hands of the pan boiler. We found quite
a remarkable difference in the attitude of the pan
boiler after 2 or 3 weeks and he certainly boiled a very
much better controlled pan without any of the extra
aids that we provided.
There are many variables involved in this project,
and if you look again at the diagrams of the graphs
nos. 1 and 2—graph no. 2 does not really line up with
the control we had of the vacuum on the pan itself
if that- is related to the other readings we took of
opening of valves and other actions on the part of the
pan boiler. You can correlate some of the variations
even in graph no. 2 with the controls operated and the
way they were operated and I think that points again
to the fact that control of two elements is by no means
sufficient. This is only the first report of what promises
to be a long investigation.
Mr. Renton: If the temperature recorded just below
the calandria is higher than that just above the
calandria surely this shows that the massecuite is being
forced downwards through the tubes,
Mr. Warne: We assume that is what happens. In
Australia, where similar investigations have taken
place, they too have found this negative value and
have also come to the conclusion that the circulation
is less definite at these points with a strong possibility
that the massecuite is being forced downwards through
the tubes.
Mr. Asfee: The temperature increase recorded below
the calandria appears to occur right at the end of the
boiling. In my opinion the circulation in the pan is due
to the fact that you have water being evaporated
which forms a vapour bubble and forces itself through
the massecuite to the boiling surface. At the end of a
boiling your concentration of massecuite is almost
complete and there is consequently very little water
left to evaporate, therefore circulation slows right
down and must rely on convection currents and it is
due to these convection currents that the figures which
were recorded are all over the place.
Mr. Warne: Graph No. 2 on page 7 records a boiling
which had an automatic control of the vacuum
and you will see that this so-called switch over of
temperatures occurred after about three hours of
boiling time leaving a further hour of boiling to be
completed. At this stage the massecuite in this pan
was cut over completely to two further pans for use as a
footing and as such it was not concentrated above the
normally accepted boiling concentration. It would
appear therefore that this switch over of temperatures
above and below the calandria is more likely due to
the hydrostatic head than to a concentration effect.
Mr. Hulett: With regard to the massecuite being
hotter beneath the calandria than above when the
pan is getting full, I notice that all the temperature
measurements were carried out in the tubes close to
the downtake and maybe that indicates that the downtake
is too small and with the pan getting fuller the
effect would be for the massecuite to come down
the downtake and also the surrounding tubes.
Mr. Grieves: Was the controller used a proportional
plus reset model?
Mr. Warne: The controller was a proportional
plus reset model. The proportional band was set at
approximately 5% whilst the reset time was approximately
half a minute. We must not forget however,
that the control valve was on a bypass which was not
ideal under the existing conditions. However, it
appeared to suffice.
191

192
Proceedings of The South African Sugar Technologists' Association —March 1966
IMPROVEMENTS IN RAW SUGAR QUALITY
Introduction
A few years ago there was little doubt that the
refining quality of Natal raw sugars ranked among the
worst in the world. Exported raws, generally representing
the cream of the sugar produced in. Natal,
were the subject of numerous complaints from overseas
refiners.
These storms of criticism precipitated a welcome
reaction among Natal millers, who began to realise
the "refining quality" was a worthwhile attainment
and not just the pet aversion of the local Refinery.
This awareness prompted various actions which have
brought about considerable advances in the quality
of sugar produced.
This paper deals with the improvement in quality
of sugars sent to the Rossburgh Refinery. Since
production for our Refinery very often represents those
sugars which are not good enough for anywhere else,
it follows that overall improvement in the quality of
sugars produced in Natal will be manifested in exaggerated
form in raw sugars sent to the Refinery.
Thus, although this paper indicates a very real
advancement in the quality of raw sugars received at the
Refinery, the overall improvement is unlikely to have
been as great.
Bases for Comparison
The refining quality of a raw sugar is not easily
defined in strict analytical terms. Obviously the
polarisation, and therefore the quantity of impurity
in the sugar is a criterion of quality. More important
to the refiner, however, is the quantity of impurity
present in the crystal, though it must be remembered
that impurities removed by a simple affination process
are not lost to the refiner but may well cause a loading
on the recovery station, especially in a refinery geared
for higher polarising sugars.
It is not only the quantity but the nature of impurities
which affect raw sugar quality. Certain
impurities notably starch, have a far more severe
effect on refining quality than, for instance, reducing
sugars.
Grain size is another aspect of raw sugar quality of
importance to the refiner. Raw sugar which has a
small, conglomerated or mixed grain will place an
unnecessary load on the affination station. The
optimum is a grain size between 0.60 and 0.80 mm.
containing as little fine grain (defined as grain passing
a Tyler 28 sieve) and conglomerate, as possible.
If raw sugar is to be stored, safety factor is of
considerable importance.
In recent years filterability has become widely
accepted as a major criterion of refining quality.
Since the filterability of a sugar is dependent to a
large extent on the quantity and nature of impurities
in the crystal, little definition is lost if refining quality
is described in terms of filterability alone, or preferably
filterability and grain size.
by R. P. JENNINGS
The Improvement in Filterability
Filterability determinations have been made on all
raw sugar processed at the Refinery since February,
1963. Omitting the results for that month, the last
month of raw sugar production during the 1962-63
season, the following trends emerge for each of the
five major contributing mills and for all local raws
received. The figures for the 1965-66 season represent
sugars received up to the end of December 1965
only:
Table I
If this year's production from Tongaat is excluded
from the average of all raws received, the filterability
figure for 1965 66 becomes 39 per cent.
The improvement may be viewed differently by
grouping sugars according to their filterability ratings.
The tons of sugar produced falling into each group
was used to determine the percentage distribution,
on a filterability group basis, of sugars processed in
each season.
Table II
Filterability Grouping
If the disappointing sugars from Tongaat are
omitted from the 1965 66 evaluation, the following
distribution emerges:
>40% 40-30% 30-20%

Proceedings of The South African Sugar Technologists' Association — March 1966
Renishaw, Umfolosi, Doornkop and Gledhow have
contributed sugar during the latter half of 1965.
The fluctuations in supply affect the overall picture
in. two ways. Firstly, the recent offerings of good
filtering sugars from Melville, Sezela and Doornkop
may have brought about a bias in favour of an apparent
overall improvement for this season. This bias
is more than balanced however by the other effect
which has resulted from the all too familiar pattern
of good quality sugars from the refinery's main
suppliers being sidetracked for export.
Despite influencing factors the tables indicate
that, apart from Tongaat, which has recently been
supplying a high proportion of B sugar, all contributors
have improved the filterability of their sugars
since 1963/64.
The Improvement in Grain Size
The improvement in the specific grain size of raw
sugars received over the past three seasons may not
seem as dramatic as the increase in filterability. It is
likely however that grain size results are more susceptible
to variations in supply, for some of the
"part time" suppliers have contributed sugars of
very poor grain size. The specific grain size and percentage
of grain passing a Tyler 28 sieve for each of
the major contributing mills and for all Natal sugars
received are given in Tables III and IV.
Table III
Specific Grain Size (mm)
Table IV
Percentage of Fine Grain
The depressing influence on this year's contributions
from Tongaat can again be seen. With Tongaat
analyses omitted the average grain size for 1965/66
is 0.61 mm.; with 33.5 per cent of fine grain.
Reduction in Crystal Impurity Content
The variations in crystal ash and. crystal starch
content of sugar received are shown in Tables V
and VI. Another major impurity, gums, has been
determined in the crystal since December, 1964,
too late to allow a meaningful analysis of trends.
An evaluation of variations in the concentration of
other impurities such as reducing sugars, silica,
phosphate and wax, has not been attempted.
Table V
Sulphated Ash in Crystal (ppm.)
Omitting Tongaat results the average crystal ash
content for 1965/66 becomes 1000 ppm.
Empangeni
Felixton
Amatikulu
Darnall
Tongaat
All raws received
Table VI
Starch in Crystal (ppm.)
1963/64
770
630
560
800
770
730
1964/65
730
540
550
530
390
520
193
1965/66
590
500
420
470
520
490
Table VII summarises the percentage reduction
in the concentration of starch and ash in the crystal
over three seasons.
Table VII
The reduction in the concentration of impurities
within the crystal over three seasons has been most
marked. It is outside the scope of this paper to comment
on the various process changes within the Industry
which may have brought about this reduction.
It should be mentioned, however, that the Refinery
has received during 1965 considerable consignments of
B sugars from factories who produce A sugar for
export. The crystal impurity content of a B sugar is
likely to be significantly higher than for an A sugar
from the same factory, whatever clarification and
boiling process is employed.
The Seasonal Effect
Sugars produced at different times of the year may
vary considerably as far as refining quality is concerned.
Using as an example filterability, graph 1 illustrates
the number of tons of raw sugar received every month
for the past three seasons which filtered above and
below an arbitary datum of 30 per cent filterability.

194
Proceedings of The South African Sugar Technologists' Association - March 1966

Proceedings of The South African Sugar Technologists' Association — March 1966
Table VIII summarises the percentage of sugars
received each month which, fell into each filterability
grouping.
Table VIII
The following points regarding Graph 1 and Table 8
should be noted:
(a) The floods of June and July 1963 following on
drought and cane fires, seriously affected the
quality of cane delivered to some mills with the
result that the quality of sugars produced by
these mills slumped dramatically.
(b) Once again the fluctuations in supply must be
taken into account. A good example occurred
in September 1965 when Felixton, who had
been supplying sugar with a filterability well in
excess of 40 per cent, had their production
diverted from the Refinery to export. The
removal of this considerable tonnage of good
filtering sugar has resulted in a misleading
deterioration in the filterability pattern for
that month.
For discussion see page 204
(c) During the three months June, July, August
1965, Tongaat supplied 34 per cent (40,200
tons) of all Natal raws processed at the Refinery.
Of this total 28,000 tons filtered below 20 per
cent filterability. Within the same period
Felixton (1,600 tons) and Empangeni (3,000
tons) were the only other suppliers of sub-20
per cent filterability sugar. It can be seen that
the influence of Tongaat sugars or the filterability
patterns for these three months was very
considerable.
Conclusions
The quality of raw sugars processed at Hulett's
Refinery has shown considerable improvement over
the past three seasons. Filterability has increased,
impurities within the crystal have decreased and grain
size has improved though the sugars contain, on an
average, about 10 per cent too much fine grain.
The improved quality of sugars sent to the Refinery
is probably indicative of overall improvement in Natal
sugars though the overall change in quality is likely
to be less obvious.
It would appear, from a study of variations within
each season, that the Refinery receives a greater
percentage of good filterability sugars during the
period July to December, with the peak months
August to October.
However, because of the many factors which combine
to influence the supply of raw sugar to the Refinery,
it is not possible to assess whether the intraseasonal
fluctuation in quality noted for refinery
sugar extends to Natal raw sugar as a whole.
Analytical Methods Used
Filterability was determined by the method devised
by the Colonial Sugar Refining Company (2).
The Spectrophotometry method, based on the
formation of the iodide complex (1) was used to determine
starch.
Sulphated ash was determined (3) on sugar affined
in the laboratory.
Grain size was calculated after sieving sugar washed
according to the ICUMSA method (3).
Summary
The improvement in refining quality of raw sugars
sent to Hulett's Refinery over Fhe past three seasons is
assessed in terms of filterability, grain size and impurities
within the crystal.
Acknowledgments
The author wishes to thank the staff of the Laboratories
of Hulett's Refineries, and in particular Mri.
Paul de Froberville for the analytical results expressed
in this paper. Thanks are also given to the Directors
of Hulett's S.A. Refineries for permission to publish
the results.
References
1. Alexander, J. B. Some Notes on Starch in the Sugar Industry.
Proc. S.A.S.T.A. Vol. 28, 1954, p. 100.
2. ICUMSA Proceeding, 12th Session (Subject 10) Washington,
1958.
3. South African Sugar Technologists' Association Laboratory
Manual for S.A. Sugar Factories, fifth edition, 1962. Published
by S.A.S.T.A.
195

196
THE QUALITY OF IMPORTED RAW SUGARS
Introduction
During the 1965/66 season the Refinery has processed
more than 117,500 tons of imported raw
sugars. Of this total 38.7 per cent came from Mauritius,
28.7 per cent from Brazil, 18.9 per cent from
San Domingo and 13.7 per cent from Thailand. The
analyses of these sugars is the subject of this paper.
Sugars Imported in Previous Seasons
10,600 tons of raw sugar from Cuba was processed
at the Refinery in 1961. No imported sugars were
refined during the 1962/63 and 1963/64 seasons but
in 1964/65 a total of 49,350 tons was imported from
Taiwan, Indonesia and San Domingo.
Analysis of Imported Sugars
Table I compares the analysis of Cuban imported
sugars with the average analysis of Natal raw sugar
for the 1961/62 season.
While the analyses of the two consignments from
San Domingo were very similar in most respects, the
larger, 12,860 tons, possessed a 24 per cent filtera¬
bility compared with 41 per cent for the smaller
consignment of 11,090 tons.
Proceedings of The South African Sugar Technologists' Association March 1966
by R. P. JENNINGS and RITA VAN KEPPEL
Table II
Table I
No figures are available for the filterability of
Cuban raws, but it was the experience of the Refinery
that Cuban sugars were far easier to process than their
Natal counterparts.
The comparative analyses of imported sugars and
Natal sugars for the 1964/65 season are shown in
Table II.
Table III summarises the analyses ofsugars imported
during 1965. The average of Natal raws up to the
end of November is given for comparative purposes:

Proceedings of The South African Sugar Technologists' Association — March 1966
The 33,792 tons of sugar imported from Brazil
included 11,193 tons of white sugar, extremely high
in moisture and reducing sugars either as a result of
deterioration or "doctoring" with an invert solution
to reduce the polarisation. The average pol of the
sugar was 98.55".
Comparison of Imported and Local Raw Sugars
It is apparent that there are considerable differences
between Natal raws and those produced in Brazil,
Thailand, Taiwan, San Domingo, Indonesia and
Cuba. Sugars from these countries are characterised
by:
1. Low polarisation.
2. High invert/ash ratio.
3. Low starch — exception Taiwan.
4. Good filterability.
While it may appear from analytical differences
that sugars imported from these countries are more
desirable, from the refiners point of view, than local
sugars — two points should be noted. Firstly, many
of the imported sugars are difficult to handle because
of stickiness, this was especially noticeable with
Indonesian and Thailand sugars. Secondly, difficulties
are encountered in handling large quantities of low
pol sugar in a refinery geared for higher polarising
Natal raws. The excess load placed on the Recovery
House was, at times, critical.
Sugar produced in Natal and in Mauritius are
similar in many respects. Polarisation and reducing/
sugar ash ratio are practically the same, while both
filter less satisfactorily than other imported sugars.
The concentration of silica, phosphate, wax, gums and
starch in the crystal of Mauritian and Natal sugars is
above the average.
The quality of Mauritian sugar varied considerably
among the nine consignments received. Filterability
ranged from 37 per cent for a consignment received
in May to 11 per cent for sugars processed in August.
Table III
197
While these differences may parallel the variations in
quality of raw sugars produced by different Natal
mills, it is pertinent to mention that the grain size
of every consignment from Mauritius fell within the
range 0.62 mm to 0.74 mm.
Analytical Methods
Analysis were carried out on composites representing
the sugar processed from each source over periods
of ten days. Pol, reducing sugars, moisture and ash
in raw sugar were determined daily.
Insolubles were determined by centrifuging in
graduated tubes. (3).
Gums in both raw sugar and crystal were precipitated
from solutions by acidified alcohol. (6).
Starch was determined spectrophotometrically by
the iodide complex method. (1).
Filterability determinations were made using the
Colonial Sugar Refining Company method at 20° C.
(5).
Phosphate (2) and silica (2, 4) were determined
colorimetrically.
Wax was extracted by the method suggested by
Alexander (2).
The methods recommended by the South African
Sugar Technologists Association (7) were followed
for all other analyses.
Summary
Analytical data are given for imported sugars
processed at Hulett's Refinery during 1965. Reference
in made to sugars imported in previous years. Essential
differences between imported sugars and Natal sugars
are noted.
Acknowledgments
The authors express their thanks to the staffs of
the laboratories of Hulett's S.A. Refineries and to the
Directors for permission to publish the results used
in this paper.

198 Proceedings of The South African Sugar Technologists' Association — March 1966
1. Alexander, J. B. Some Notes on Starch in the Sugar Industry.
Proc. S.A.S.T.A. 28, 1954, p. 100.
2. Alexander, J. B. Some Observations on the Filterability of
Natal Raw Sugars. Proc. S.A.S.T.A. 31, 1957, p. 70, 71.
3. Alexander, J. B., and Douwes-Dekker, K.. The Presence and
Determination of Insoluble Impurities in Raw Sugar.
Quart. Bui. S.M.R.I. July, 1959, p. 20
4. Alexander, J. B. and Parrish, J. R. Silica in Raw and Clarified
Cane Juices. S.A. Sugar Journal, 37, September, 1953,
p. 573.
References:
For discussion see page 204
5. I.C.U.M.S.A. Proceedings 12th Session (Subject 10) Washington,
1958.
6. Jennings, R. P. Some Notes on the Determination of Gums
in Sugar Products with Special Reference to their Distribution
in Hulsar Process. Proc. S.A.S.T.A. Vol. 38, 1964,
p. 87.
7. South African Sugar Technologists Association Laboratory
Manual for South African Sugar Factories, fifth edition,
1962.

Proceedings of The South African Sugar Technologists' Association—March 1966 199
A MODIFIED METHOD FOR DETERMINING FILTERABILITY
Introduction
During the 1963-64 season two different tests were
used within the Natal Sugar Industry to determine
the filterability of raw sugar. These tests were:
1. That using the bomb filtration apparatus of the
Johns Manville Corporation,
and
2. That developed by the Colonial Sugar Refining
Company of Sydney.
Comparison of the results obtained by these two
methods on affined Natal raw sugars indicated appreciable
differences on occasions, while correlation
of results with the actual behaviour of the sugars in
a carbonatation refinery proved difficult.
The Modified C.S.R. Test
It was decided to develop a filterability test incorporating
the best features of the two existing tests,
and to compare results obtained by the new method
with those obtained by the bomb method and the
C.S.R. test at 20" C. To overcome sampling problems
it was decided that results should be grouped on a
monthly basis. Details of the modified C.S.R. method
appear in the appendix.
The main advantages of the new method appear to
be:—
(a) The test is conducted at approximately factory
working temperature.
(b) The time taken is shorter than for the bomb
test, about the same as for the C.S.R. method
at 20° C.
(c) The quantity of sample used is less than for the
bomb test.
(d) The apparatus lends itself to the direct determination
of the filtration rate of refinery
liquors.
The chief disadvantage, compared with the bomb
method, is the absence of stirring while the test is in
progress. Since the duration of the test is far shorter
than for the bomb method this shortcoming is proportionately
less serious.
The necessity for maintaining the apparatus at
80° C, which might seem a disadvantage when compared
with the standard C.S.R. method, is easily
overcome by the use of an efficient constant temperature
circulating water bath and by ensuring that the
connecting pipes and the apparatus itself are well
lagged. It has been our experience that it is easier to
maintain temperature with the modified apparatus
than with the C.S.R. apparatus at 20° C.
Results obtained with the new Method:
Table I and the attached graph illustrate the
monthly variations in the filterability of affined melt
by R. P. JENNINGS
sugar during the period August 1964 to December
1965. Results are given for analyses by each of the
three methods.
The usefulness of a filterability test for raw sugar
depends upon the accuracy with which the test may
be used to predict the actual behaviour in the refinery
of liquor produced from that sugar.
Attempts were made during the 1963-64 season to
find an expression for the ease of filtration of sugar
liquors under actual factory conditions in a carbonation
refinery. Factors considered were, inter alia, tons
of solids filtered in unit time, number of filter cycles
per unit time, quantity of filter aid used and the length
of life of the filter cloth. When all these factors were
taken into account it was found impossible to derive
a usable expression for "Refinery Filter Performance".
Approaching the problem from a different angle in
the 1964-65 season it was decided to determine daily
the rate of filtration of factory carbonated liquors
through the modified C.S.R. apparatus. No filter aid
was used for the determination. Research showed
that a comparison of the flow rate of the Liquor at
natural brix with the flow rate of sucrose at the same
brix yielded a figure which differed very little from the
"filterability" figure obtained by comparison at 60°
brix. It was thus decided that there should be no
dilution of the carbonated liquor before testing. The
figure for the filterability of carbonated liquor in
Table 1 and in the graph is thus the comparison of
flow rates of the liquor and of a sucrose solution at
the natural brix of the liquor.
Table I
Filterability of Melt Sugar and Carbonated Liquor

200 Proceedings of The South African Sugar Technologists' Association - March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966 201
Note: Due to a fault in the pressure system of the
modified CSR apparatus, filterability figures for melt
sugar and carbonated liquor for July 1965 were not
determined.
Certain factors have to be taken into account when
attempting a correlation betweeen the filterability of
melt sugar and the corresponding carbonated liquor:
1. The melt sugar tested was a composited sample
made up of 90 per cent affined raw sugar and
10 per cent recovery sugars. This represents an
approximation of the actual ratio of recovery
sugars to total melt in practice but may vary
considerably, especially when low pol sugars are
being processed.
2. Melt sugars were composited on a daily basis
sub-sampled, and composited as a monthly
sample on which the filterability was determined.
3. The carbonated liquor tested was a composite
of catch samples taken hourly over 24 hours.
Daily figures were averaged arithmetically to
give the monthly filterability of the liquor.
4. The compositions of neither the carbonated
liquor nor the melt sugar samples were dependent
upon tons of sugar melted. Errors may have
occured through accepting arithmetic instead of
weighted averages.
5. The comparisons have been made at a time
when the refinery has been processing a very
"mixed bag" of raw sugars. Table II shows the
percentages of Natal and Imported raws for
each month of the comprison.
Table II
6. During April the refinery reprocessed a considerable
quantity of contaminated refined sugar.
7. The factory carbonation process is by no means
standard, and the amount of "gassing" is varied as the
quality of the incoming sugar dictates.
With these factors considered there appears to be
definite correlation between the filterability of melt
sugar determined using the C.S.R. apparatus — both
at 20° C and 80° C and the filterability of carbonated
liquor derived from that sugar. The correlation is
closest between October 1964 and March 1965, when
the refinery was processing raws obtained almost
exclusively from Natal. The anomalous result obtained
in September 1964 occurred during a time when the
refinery was processing sugars from San Domingo
and Indonesia. The latter sugars, though yielding a
modest laboratory filterability figure (± 50 per cent,
C.S.R, 20° C), behaved in practice like sugars of far
better filtering characteristics. It would appear that,
for this sugar, the filterability tests used for the affined
sample were unable to predict factory performances
with any degree of accuracy, while tests on the
carbonated liquor showed the situation in a truer
light.
There is a fair correlation between the filterability
results obtained on the melt sugar by the three different
test methods. The correlation between the two
sets of results obtained with the C.S.R. apparatus
is better on the whole than correlation with figures
obtained using the bomb method. There is little to
choose between the C.S.R. method and modified
C.S.R. method as far as correlation with carbonated
liquor filterability is concerned. While both appear
to predict carbonated liquor filterability more
accurately than the bomb test, this may result from
the method used for determining carbonated liquor
filterability. Had the bomb apparatus been used the
situation might well have been reversed.
Summary and Conclusions
The C.S.R. method at 20° C, and the modification
of the C.S.R. method at 80° C may be used to indicate
trends in the performance of the refinery filter
station. The modified apparatus may be used to
determine the filtration rate of carbonated liquors.
It would appear that correlation of data between
laboratory filterability and factory filter performance
is closest when sugars of one origin are being refined,
though this may well be due to changes in the carbonation
process in the factory.
Acknowledgments
The author wishes to thank the staffs of the laboratories
of Huletts S.A. Refineries and the Sugar
Milling Research Institute for the analytical results
used in this paper, and to thank the Directors of
Huletts S.A. Refineries for permission to publish
these results.
References
Alexander, J. B. and Graham, W. S., 1964, unpublished Report.
Archibald, R. D., Variation of Filterability with Brix, Proc.
S.A.S.T.A., Vol. 30, 1965, p. 50-55.

202
Preparation of Sample
As with the Bomb and C.S.R. methods, the raw
sugar sample is affiliated prior to testing, according
to the following technique.
1,000 ml saturated aqueous refined sugar solution
at room temperature are added to 1,200 g. of the sugar
to be analysed in a one gallon Mason jar, the lid of
which is provided with a rubber gasket. Thorough
mingling is effected by rotating at 30 r.p.m. for 30
minutes after which the magma is centrifuged in a
laboratory centrifugal. Washing of the sugar in the
centrifuge is carried out using 50 ml. cold water
delivered in a fine jet from a specially constructed
wash bottle. Centrifuge for six minutes at 3,000 r.p.m.
in an eight inch diameter basket. The sugar is then
spread out in a thin layer on a sheet of paper to air
dry before proceeding with the analysis.
Reagents
Standard Buffer solution which should be prepared
as follows:
1. Prepare about one litre of 50 per cent W W
glycerol made from B.P. Glycerol.
2. Take two beakers:
(a) Dissolve 15 g. A.R. calcium acetate in
sufficient of the 50 per cent glycerol solution,
with heating if necessary.
(b) Dissolve 400g. triethanolamine in sufficient
of the 50 per cent glycerol solution. The
triethanolamine should be commercial water
white or pale yellow grade.
3. Transfer the contents of the two beakers to a
clean, dry litre flask, rinsing the beakers thoroughly
with the 50 per cent glycerol solution.
Cool, if necessary.
4. Make up to one litre with the 50 per cent glycerol
solution, mix well, and allow to stand
overnight. Add a little supercel, and filter in the
test filter or Buchner funnel. Store in a stoppered,
clear glass bottle which has been cleaned,
and rinsed with some of the filtered solution.
Filter Aid
Laboratory Standard Filter Cel supplied by the
Johns Manville Corporation, California, U.S.A.
Filter Septa
Filter cloth conforming to the South African
Bureau of Standards Specification 512-1956 Type
D157. the essentials of which are:
Proceedings of The South African Sugar Technologists' Association—March 1966
APPENDIX I
DETAILS OF THE MODIFIED COLONIAL SUGAR REFINING METHOD FOR THE
DETERMINATION OF RAW SUGAR FILTERABILITY
The fabric composition is exclusively cotton.
The cloth is cut into circles 5.5 cm. in diameter.
Apparatus (See Photograph)
The apparatus is a modification of that developed
by the Colonial Sugar Refining Company of Sydney,
details of which may be found in I.C.U.M.S.A.
Proceedings, 12th Session (Subject 10), Washington
1958.
The filtration apparatus consists essentially of
a cylindrical brass tube 1.7/8 inches in diameter by
9.7,8 inches long. The filter cloth is supported by a
filter disc which is firmly held in position in the base
end of the tube. The top of the tube is closed by a
screw on lid. Pressure may be applied to the contents
of the tube through an air line passing through the
lid.
As much of the length of the tube as possible (not
less than 7 1/2 inches) is enclosed by a metal jacket
through which water at 80° C is circulated from a
constant temperature water bath. The apparatus and

Proceedings of The South African Sugar Technologists' Association—March 1966 203
the pipes connecting apparatus and water bath are
lagged with asbestos. The apparatus is attached to a
rigid stand by means of a clamp, designed in such a
way that the apparatus may be tipped for cleaning
purposes without complete removal from the stand.
To obtain values for the filtration rate of sucrose
it is necessary to connect together, end to end, two
jacketed tubes, attaching the filter disc to the base of
the lower tube. The tubes are easily connected by
means of a short brass connecting piece giving a
pressure tight seal on both tubes.
Procedure
300 g affiliated sugar are dissolved in 200 g water
and the pH of the solution adjusted to 9.0 using
standard buffer solution (±4 ml.). The beaker is
covered with a watch glass and heated to just over
80° C in a water bath. Maintaining the solution at a
little over 80° C an amount of Standard Filter Cell
equivalent to 0.35 per cent on brix (1.050 g for 300 g
sugar) is added to the beaker and dispersed in the
solution by stirring for 30 seconds with a mechanical
stirrer. The solution is transferred to the previously
assembled apparatus where the test is carried out at
80° C. The lid of the apparatus is screwed on after
checking the temperature of the solution.
Air pressure of 50 p.s.i.g. is built up as quickly
as possible in the apparatus and a stop watch started
simultaneously with the application of pressure.
Filtrate collected during the first two minutes of the
filtration is discarded. The remainder of the filtrate
is collected in a graduated cylinder, noting the volumes
in the cylinder at the end of the sixth, seventh and
eighth minutes from the commencement of filtration.
The pressure is maintained at 50 p.s.i.g. for the duration
of the test. The following table illustrates the
method of calculation:
With the fast filtering sugars it was found more
practical to use filter discs of 1 1/2 inches to an inch in
diameter in place of the 1 3/4 inch I.D. disc used in the
C.S.R. method. In each case the volume filtered is
compared with volume of a sucrose solution filtered
under the same conditions, using the same sizes
disc.
Note
1. With fast filtering sugars, especially refined sugar,
it is necessary to double the quantities of solution
used.
2. Care must be taken to avoid a substantial increase
in the brix of the solution due to evaporation. It
may prove necessary to reduce the original brix,
before heating, to 59° to allow for evaporation.
3. Periodic checks should be made on the apparatus
using refined sugar solutions. The filtration rate
should be such that the volumes of filtrate collected
should differ, on the average, by not more
than 10 ml. from those shown in columns 3 of the
example given above. (This is assuming that the
highest quality refined sugar is used in every check.)
THE FACTORY FILTER PERFORMANCE OF RAW SUGARS
Introduction
Provided that the many variables in the refining
process prior to filtration are kept constant, the filter
performance of a sugar may be gauged by the number
of tons of solids handled by unit filter:
Filter Performance =
Tons solids filtered in unit time
Number of filters started in unit time.
For comparison purposes, allowing for a constant
ratio of re-melted recovery sugars to total melt, the
equation becomes:
Filter Performance =
Tons sugar melted in unit time
Number of filters started in unit time.
APPENDIX 2
Practical Considerations
During the tests on filter performance of various
sugars at the refinery, the fullest co-operation of the
Process department was obtained to ensure the following
:
1. Sugar from one source only was processed during
the period of the test.
2. A constant melt rate of affinated sugar.
3. A constant ratio of recovery sugars to total melt.
4. A constant supply of liquor to the Carbonatation
station.
5. No variation in the procedure of carbonatation.
(Quantity of lime added, gassing rate, composition
of gas, pH of liquor at various stages.)
6. A standard technique in the filter station using
cloths of approximately the same age for each
test.

204 Proceedings of the South African Sugar Technologists' Association—March 1966
Samples of affinated sugar, carbonatation supply
liquor and carbonatated liquor were taken at regular
intervals and composited to give a sample representative
of the period of the test. The filterability of each
sample and the gum and starch content of the affinated
sugar were determined.
Note:
(1) Tons melt per filter is based on tons of test sugar melted
and does not take into account re-melt sugar.
(2) The test run with Darnall sugar was curtailed because of
the poor filtration of the carbonatated liquor. Despite the
use of filter aid, the flow of liquor to the Carbonatation
station had to be reduced, upsetting the controls of the
test.
Discussion
Though the accumulation of data concerning the
performance of raw sugars in the filter station of the
refinery is still at a very early stage, some very interesting
features have already come to light.
There appears to be a decided absence of correlation
between affinated sugar filterability and "filter
performance". This is most marked for Umzimkulu
sugar.
Though the laboratory filterability of the sugars
from Umzimkulu indicated that its performance in
the refinery would be about equal to that of the
Empangeni and Felixton sugars (test 2 and 3), in
practice the performance was far superior. The figures
of 120 and 105 tons per unit filter are three times the
values for the earlier tests.
At the other end of the scale, laboratory figures were
unable to predict the very poor filtration of the Darnall
sugar in test one.
It would appear that more significance can be
attached to the filterability of carbonatated liquor. The
carbonatated liquor filterability figures of tests four
and five are both well above average, while Darnall
sugar tested produced a carbonatated liquor filtering
only 14 per cent.
It seems certain that the excellent filtering characteristics
of the sugar from Umzimkulu are connected
with the below average gum content and the very low
starch content of this sugar. Less easy to explain is
the relatively low filterability figure obtained for this
sugar in the laboratory. A parallel situation already
mentioned in the earlier part of this paper, was noted
for sugar from Indonesia. It is interesting to note that
these sugars also contained very little starch. These
results tend to indicate that the laboratory filtration
test cannot be used with any confidence to compare
the potential filter performance of sugars of high and
low starch content. With such sugars it would be
better to follow the lead of at least one overseas
refinery and predict filter performance on the strength
of starch content alone.
Investigations into the factory filter performance of
raw sugars will be continued next season. It is hoped
that, when a lot more data has been accumulated, the
upper and lower confidence limits of the present
laboratory filterability tests may be determined.
Mr. Carter: Mr. Jennings mentioned that "low
starch" sugars filtered better in practice than seemed
likely from laboratory results. Because of the nature
of the Rabe process is it not possible that the explanation
lies in the poor removal of other impurities,
e.g. phosphate, during clarification?
Mr. Jennings: I think it more likely that the reason
for the discrepancy lies in the mechanism of the laboratory
filterability test, which employs an added filter
aid, compared with factory carbonatation using a
produced calcium carbonate as filter aid. It would
appear likely that starch has considerable influence
on the formation of the calcium carbonate precipitate.
Mr. Carter: We should then direct our research
towards a laboratory test incorporating carbonatation.
Dr. Graham: It seems apparent that the filterability
tests used at present are misleading, especially for
sugars low in starch. I agree that the development of
a new test incorporating laboratory carbonatation is
important. Work started in this direction at the
S.M.R.I, was discontinued even though a carbonatation
tank was built. It is especially important to
develop this test now since the possibility of the establishment
of a bonus system for fast filtering sugars is
being considered by the South African Sugar Millers.
Using present tests Umzimkulu sugar would be
penalised under such a scheme.
Research towards improvement in filterability
has centred on establishing what impurities impede
filtration and the removal of these impurities. It has
been established that starch is a major contributor
to poor filtration and various starch removal methods
have been tried. The procedure devised at Umzimkulu
is cheap to operate, requires little capital equipment
and has been operating successfully for almost
a complete season. It is my opinion that the sugars
produced by this method are of such a high quality
that, were the process used throughout Natal, the
filterability problem would cease to exist.
Dr. Roux has pointed out to us that a delay in the
application of new ideas can cause a major loss in the
advantages to be gained. I hope his remarks will not
be prophetic of this industry's attitude in respect of the
solution to our filterability problem.

Proceedings of The South African Sugar Technologists' Association—March 1966 205
Mr. Alexander (in the chair): I agree wholeheartedly
with all that Dr. Graham has said and would
like to congratulate all associated with the experiment
at Umzimkulu, and especially Mr. Rabe, on the
success of the project and the excellence of the sugar
produced.
Dr. Graham: Why was the figure used for filter
performance in the refinery test runs, tons melt per
filter and not tons melt per filter per hour?
Mr. Young: The figure which best illustrates factory
filter performance is tons melt divided by the number
of filters started during the time in question. In the
refinery the term "filter starts" is used to indicate
the interval between putting filters on stream. If a
filter station processes 70 tons per hour using 10
filters each filter handles 7 tons per hour. With 1/2 hour
starts each filter is left on stream for 5 hours and
processes 35 tons. If 5 filters are on stream the result
is also 35 tons per filter since though the quantity
processed per filter per hour is double, the length of
the filter cycle is halved.
Dr. Graham: Mr. Jennings mentioned a "below
average gum content" for Umzimkulu sugar. It was
the experience of the S.M.R.I. that the Rabe process
removed no starch-free gums.
Mr. Jennings: It appeared from our analyses that
there was some removal of starch-free gums, though
nothing like the removal of starch.
Mr. Buchanan: Was any record kept in the refinery
of the performance of Taiwan sugar, which has a
high starch content ?
Mr. Alexander: The refinery did not carry out tests
on Taiwan sugar since the sugar was not processed
separately but mixed with Natal raws. A recent paper
reporting research in Japan disclosed a disappointing
correlation between starch and laboratory filterability
determined using the CSR apparatus, but a very
good correlation, with a correlation coefficient of
0.95, between starch and the filterability of laboratory
carbonatated liquors.
Mr. Jennings: It may appear from the paper on
filterability that the laboratory tests used at present
are not very meaningful. This is not the view of the
refinery which considers the laboratory tests very useful
for normal Natal sugar, containing average starch
content of 350-500 p.p.m. on the affinated sample.
For sugars of low starch content the test is not satisfactory
but if any Natal mill can send sugar to the
refinery consistently containing less than 200 p.p.m.
starch they are welcome to throw away their filterability
apparatus!
Dr. Matic: I must congratulate Mr. Jennings in
particular and the refinery in general on their research
and on the applicability of the data presented. I hope
that other industrial laboratories will follow Hulsar's
example and report results of investigations which are
of general interest to the industry.

206
Proceedings of The South African Sugar Technologists' Association—March 1966
SOME EFFECTS OF BORAX ON THE POLARISATION
OF SUGAR SOLUTIONS
Introduction
The influence of boric acid on the specific rotation
of carbohydrates was studied by Boeseken 1 who used
the phenomenon to determine the structure of
hexoses.
Hernandez 4 investigated the possibility of using
borax to nullify the polarisation of common monosaccharides
in sugar products in order to obtain a
direct reading of sucrose content. Studies of the
effect of borax on the polarisation of solutions of pure
sugars and mixtures of pure sugars proved encouraging.
Hernandez extended his investigations to impure
sugar solutions, comparing results for direct pol, borax
pol and Clerget sucrose. He concluded that the
addition of 25 ml. of 2 per cent solution of disodium
borate to 6.5 g. of a sugar product, diluted to 100 ml.
resulted in a pol value very close to true sucrose.
Hernandez's proposals were investigated by, inter
alia, the Sugar Milling Research Institute'-, who considered
that no practical benefit could be gained by
adopting the technique in South African sugar
laboratories.
During investigations of the specific rotation of
gums, Jennings 6 found that the addition of 20 ml.
5 per cent borax solution depressed the pol of pure
sucrose by more than 2 per cent. Since other investigators
had found similar discrepancies with the
original experimental results of Hernandez it was
decided that the whole question of using borax to
obtain close approximations of sucrose content by
direct polarisation should be re-examined.
Aspects Studied
At the outset the effect of borax on the rotation of
solutions of pure sugars, and mixtures of pure sugars
was investigated. Attention was paid to the effects
on the rotation of (a) varying the concentration of
borax added, (b) the pH of the solution and (c) the
presence of impurities simulating the inorganic ash
found in sugar products.
The effects of the addition of different concentrations
of borax on the polarisation of impure sugar
solutions was studied, special emphasis being given to
devising a method for the rapid determination of the
approximate sucrose content of invert syrups. The
dependence of the final polarisation figure on the
time interval between the addition of borax and polarisation
was investigated.
Methods Used
(a) Pure Sugar Solutions
A. R. Dextrose having a specific rotation between
52.5° and 53.0°, and dried levulose having a specific
by D. ADAM and R. P. JENNINGS
rotation not less than —81° and an ash content less
than 0.5 per cent were used for all determinations
involving pure monosaccharides. Highest quality
sucrose, specially prepared, and polarising not less
than 99.95° was used for all investigations with
sucrose.
Aliquots of prepared solutions of pure sugars were
transferred to 100 ml. flasks, the required quantity
of borax added and the mixture made to volume,
shaken, and allowed to stand for one hour before
polarising in 200 mm. tubes in a standard saccharimeter.
If the pH of the solution was to be adjusted
the sodium hydroxide or carbonate was added before
making to volume and pH determined after standing
for one hour and polarising. A similar procedure was
adopted during the investigations of the influence of
inorganic matter.
(b) Molasses
The method used for investigations on Refinery
molasses was as follows:
32.5 g. molasses weighed into 250 ml. flask and
made to volume.
50 ml. aliquots pipetted into 100 ml. flasks.
Required amount of borax added before making
to volume, clarifying with 2g. dry lead acetate and
polarising, after leaving the filtrate to stand for one
hour in a covered vessel, in 200 mm. tubes.
Sucrose on Refinery molasses was determined using
the Jackson and Gillis Method IV, clarifying with
6 g. lead acetate and 2 g. potassium oxalate per 32.5 g.
molasses and using the Walker method for inversion 7 .
The treatment of Mill molasses samples varied only
in that 3 g. lead acetate were added to clarify borax
pols and 10 g. lead acetate with 2 g. potassium oxalate
per 32.5 g. molasses for sucrose determinations.
(c) Invert Syrups
The following method was used to investigate the
effect of standing time on the polarisation of invert
syrups treated with borax.
6.5 g. syrup weighed into 100 ml. flask. Required
amount of borax added and the flask immediately
made to the mark and shaken. A 200 mm. tube was
filled immediately and sealed to prevent evaporation,
polarisations being recorded every 2 minutes
from the time of the addition of the borax. No
clarification was necessary with any of the invert
syrups tested. A similar method, allowing a 60
minute interval before polarising was used for other
tests with invert syrups.
Results
A. The action of borax on solutions of pure sugars.
1. A re-examination of the investigations of
Hernandez.

Proceedings of The South African Sugar Technologists' Association — March 1966
2. Variation of the concentration of Borax.
3. Variation of the pH of the solution.
(a) Aliquots of a dextrose solution, each containing
2 g. of dextrose, treated with 50 ml. 4 per cent
borax and N/1 Sodium Hydroxide to vary pH between
8.0 and 12.0.
Table III
Table I
Table II
(b) 0.25 g. dextrose + 0.25 g. levulose +
25 ml. 2 per cent borax in 100 ml. pH adjusted using
N/1 sodium hydroxide.
Table V
Table IV
207
4. The influence of inorganic impurities.
A solution was prepared containing salts in a
ratio suggested by Deerr 3 to correspond closely
to the inorganic content of sugar products. The
aqueous solution contained the following salts,
diluted to 500 ml.:
K2CO3
NaCl
CaCO4
KC1
2.1 g.
0.3 g.
2.1 g.
1.15 g.
Ca3PO4
MgCO3
K2SO4
CaSO4
0.4 g.
0.9 g.
1.76 g.
0.25 g.

208
5. The effect of time upon the polarisation of
sucrose/borax solutions. A. R. Sucrose from two
different sources was used.
Table VI
B. The action of borax on impure sugar products.
2. Variation of the borax concentration.
(a) Imported raw sugars.
Table VIII
Proceedings of The South African Sugar Technologists' Association - March 1966
(b) Refinery Molasses.
Table IX
(c) Mill Molasses
Table X
1. The Hernandez technique applied to Refinery
Products.
Table VII

Proceedings of The South African Sugar Technologists' Association—March 1966
(d) Invert Syrups. Table XI
3. The effect of time.
(a) Table XII illustrates progressive pol readings
at two minute intervals for solutions containing 6.5 g.
(b) Aliquots containing 6.5 g. invert syrup
adjusted to various pH values between 4.2 and 7.8,
50 ml. 4 per cent borax added to each aliquot.
Table XIII
Discussion and Conclusions
Tichla and Friml 9 have concluded that three moles
of borax are required to nullify the polarisation of one
mole of dextrose, while 2 moles of borax reduce the
pol of one mole of levulose to a minimum. However,
Swann, McNabb and Hazel 8 suggest the formation
of a 1 : 1 complex of borate with the enol form of
levulose, while Boeseken 1 and. Hernandez 4 favour a
combination in the ratio of 2 moles of the hexose to
one mole of borate.
In the original experiments of Hernandez the mole
ratio of borax to both dextrose and levulose was
129 : 277. This ratio would appear to be too low to
allow for complete nullification of the pol of the sugars.
It can be seen from table II that if 50 ml. 4 per cent
borax are added to 0.5 g. of dextrose the reduction of
the polarisation of the monosaccharide is complete.
The ratio in this case was 2 borax to 1 dextrose.
However a 2 : 1 ratio of borax to levulose will not
reduce the polarisation beyond -0.8°.
Table XII
209
invert syrup, with 25 ml. 4 per cent borax in 100 ml.
(solution 1) and 50 ml. 4 per cent borax in 100 ml.
(solution 2).
Without a knowledge of the respective rates of
reaction of borax on levulose and/or dextrose it is
not possible to calculate the expected polarisation of
a mixture of the two hexoses when treated with borax
in various concentrations. Clearly, however, the
mechanics of the nullification of the polarisation of
the hexoses by borax is complex.
A major discrepancy between the results obtained
by Hernandez and the findings expressed in this paper
is the depression of the polarisation of sucrose (Tables
I, II, VI). While Hernandez mentions the depression
of the sucrose pol in his text, and even goes so far as
to offer an explanation for the phenomenon, his
published results indicate no depression. One is
left wondering whether a "sucrose" which yielded a
polarisation of 49° in semi-normal solution (Table I)
was a satisfactory starting material for his investigations.
Tichla and Friml have noted that "if more than
0.5 g. of borax is added to 100 ml. of a 0.25-1.0 N
solution, the sucrose polarisation will be reduced.
From Table I it can be seen that even the addition
of 0.5 g. of borax to semi-normal solution reduces the
sucrose pol by 0.5 per cent. The results in Table VI
indicate that the addition of larger amounts of borax
can reduce the pol of the sucrose by almost 3 per cent.
This phenomenon must always be taken into account
when proposing techniques for direct sucrose determination
using borax. Whether the pol of sucrose is
reduced if reducing sugars are present in sufficient
quantity to utilise all the existing borax is not clear.
It seems certain however, that the addition of borax
to sucrose/reducing sugar solutions in a ratio larger

210
Proceedings of The South African Sugar Technologists' Association—March 1966

212 Proceedings of The South African Sugar Technologists' Association—March 1966
than required to neutralise the reducing sugars will
cause a depression of the sucrose pol (Table VIII).
Table III illustrates the effect of pH on the reduction
of dextrose polarisation by borax. It is not clear however
whether the reduction of the pol at higher pH
values is due to enhancement of the effects of the
borax on the dextrose, or to other factors.
The effect of inorganic impurities on the polarisation
of sugar/borax solutions is shown in Table V
The changes in polarisation are probably due to the
influence certain salts have on the pol of monosaccharides
5 , and not to any additive effect by the
mineral salts on the nullifying action of the borax.
While the addition of borax in the ratios suggested
by Hernandez has proved of little value under South
African conditions 2 (Table VII), the use of larger
concentrations of borax has yielded encouraging
results. With both refinery molasses and mill molasses
(Tables IX and X) the addition of 50 ml. 4 per cent
or 5 per cent borax yields a polarisation figure far
closer to true sucrose than a direct polarisation. It
would appear that a very reasonable approximation
of sucrose in molasses can be made rapidly by using
50 ml. of 4 per cent borax.
With invert syrups, having very high reducing sugar
contents, the agreement between Clerget sucrose and
a direct pol; using 50 ml. 4 per cent borax is even more
striking (Table XI). With the large quantities of
reducing sugars present in the syrup it is difficult to
reconcile this agreement with the theories of Tichla
and Friml, for the mole ratio in, for example, sample
nine of Table XI is of the order of 5 borax to 18
reducing sugars. A further conclusion of these authors
that "the nullpoint occurs when the weight of borax
added equals that of the invert sugars" is also not
borne out by this sample for which 2 g. of borax
were added to nullify the polarisation of 3.3 g. reducing
sugars.
It is obvious that the reaction between borax and
the reducing sugars is most complex. Table XIII
indicates that the polarisation of a sucrose/reducing
sugar solution in the presence of borax is a function of
time, while the rate of reaction is affected to some
extent by the addition of alkali. Graph 1 has been
constructed to illustrate the change in the polarisation
of the reducing sugars with time, taking P , the final
pol of the solution, as the pol due to sucrose alone.
Graph 2 is a plot of time against the logarithm of
the pol of the reducing sugars.
It is reasonable to expect that at least three separate
reaction mechanisms combine in this complex
chemical relationship: (1) The reaction between
dextrose and borax, (2) the reaction between levulose
and borax, (3) the mutarotation of the monosaccharides.
To these can be added the possible reduction
in the polarisation of sucrose though it can be seen
from Table VI that the reduced pol of a sucrose/borax
solution is unaffected by time.
While it would be foolhardy at this stage to attempt
to decipher the mechanisms of the reactions which take
place when borax is added to an invert syrup solution
it is not unreasonable to suggest a method for the
rapid determination of sucrose in invert syrups. In
all cases studied the pol of the solution had reached
stability after 40 minutes. The suggested method for
direct sucrose determination in invert syrups allows a
60 minute interval between the addition of borax
and. polarisation.
Rapid determination of Sucrose in Invert Syrups
6.5 g. of the test syrup are weighed into a 100 ml.
flask 50 ml. 4 per cent borax is added, the contents
of the flask mixed by swirling and set aside for 60
minutes. After this time the solution is made to
volume, mixed and polarised in a 200 mm. tube in a
standard saccharimeter. If clarification of the solution
is necessary a minimum of dry lead acetate
should be added after the solution has been made to
volume. The pol reading x 4 will give the percentage
sucrose in the sample.
Summary
Observations of the effects of borax on the polarisation
of solutions of pure sugars have proved inconsistent
with the findings of J. A. Lopez Hernandez
(I.S.J., 65, 1963, 46-48, 72-73, 107-109). The use of
borax in greater concentrations than recommended
by Hernandez, to determine the sucrose content of
impure solutions, has proved promising. Following
a brief study of the mechanism of the reaction between
borax and the reducing sugars, a method is proposed
for the rapid determination of sucrose in invert
syrups.
Acknowledgments
The authors wish to thank the directors of Hulett's
South African Refineries, Limited, for permission to
publish the results which appear in this paper.
References

Proceedings of The South African Sugar Technologists' Association—March 1966
Mr. Alexander (in the chair): Although this paper
was presented by two authors I would like especially
to congratulate Mr. Dave Adam since he must have
created a record by producing his first paper at the age
of 65. Mr. Adam is a pensioner and has given up a large
amount of his time to carry out this research.
Dr. Matic: I was about to congratulate the Refinery
once again on producing a paper both of
scientific merit and of practical significance to the
industry as a whole. Now I find that the work, was
done by an individual researcher working in his spare
time—an even more creditable effort!
Mr. Alexander: In reply to Dr. Matic I would like
to mention that Mr. Adam has connections with
213
the Refinery—he in fact worked at Hulsar for
47 years!
While it is obvious from the paper that we at the
Refinery do not understand the workings of the borax/
reducing sugar reaction, we feel that the phenomenon
has considerable practical possibilities and we appeal
to other laboratories to test the method, especially
as far as molasses is concerned. The advantages of a
quick method for determining sucrose in molasses
are obvious.
Dr. Graham: The S.M.R.I. conducted experiments
using borax concentrations greater than those proposed
by Hernandez but did not reach the high
concentration levels used by Mr. Adam.

214 Proceedings ofThe South African Sugar Technologists' Association—March 1966
REFINED SUGAR CONDITIONING AND STORAGE
Summary
In view of the wide variation in South African
atmospheric conditions and, in particular, its humidity,
refined sugar when transported from the humid
climate of Durban to the dry climate of Bloemfontein
and the Reef, is exposed to conditions which invite
"caking". This is the co-cementation of crystals at
their various points of contact due to the loss of
moisture or the migration of moisture between crystals.
1
The term "conditioning" is now used to describe
the treatment applied in its various forms to refined
sugar in order to reduce handling, packing and
storage problems.
In recent years there has been a change in the
recognised practice of cooling sugar after drying. This
change follows the work of Powers 2 which clearly
shows that there exists an excess of "bound" moisture
within the crystal. This moisture migrates to the
surface over a period of several days and its migration
rate is dependant upon temperature. A satisfactory
condition of Refined Sugar in storage is dependant
upon three requirements:
1. The exhaustion of this "bound" moisture.
2. The stability of the surrounding atmosphere, and
3. A necessary ventilation in times of excessive
ambient temperatures.
Factors Known to Affect the Condition of Sugar
1. Conglomerates.
2. Grain Size and Distribution.
3. Compression.
4. Moisture Content.
5. Temperature.
1. Conglomerates are caused by the adhesion of
two or more crystals normally created during one or
more of the boiling, curing, and drying processes.
The higher proportion of syrup film held around
the crystal interface requires additional washing, drying
and more attention in conditioning. Conglomerated
crystals not only impede the flow of sugar in high
speed packing equipment, but also affect the bulk
density of the sugar
Mechanical circulation is considered necessary to
reduce the formation of conglomerates. In conjunction
with automatic control, circulators are claimed 3 to
reduce steam consumption by 33 per cent and increase
pan boiling capacity by the same proportion.
2. Grain Size is normally controlled in pan boiling
to certain limits of "Mean Aperture". If the sugar is
completely screened into different fractions as is the
practice in some Canadian and American Refineries,
By A. M. HOWES
there would be little tendency for crystal separation
in silos or bins. However, it is necessary to prevent
separation if the sugar is not screened. There exists
a greater tendency for small grain to consolidate due
to an increase in both the surface area and the number
of points of contact between crystals; for example,
the tendency of icing sugar to cake is significant.
3. Compression of sugar occurs in silos or bins
where the bulk density might increase between 5 and
6 per cent above the standard laboratory test figures. 4
Pressure not only in bulk sugar vessels, but also in
bagged sugar stacked to appreciable heights will tend
to fuse together the fine crystals in particular; more
so in the presence of migrating moisture vapour.
4. Moisture Content of refined sugar is normally
0.04 per cent, while sucrose purity exceeds 99.93 per
cent. A moisture content of 0.08 per cent, that is,
double the normal figure would provide a sugar with
a definite dampness. This minute quantity of moisture
signifies the susceptibility of refined sugar to changes
in humidity.
From the time wet sugar enters the granulators for
drying there exist three stages of moisture removal.
The first is the free moisture around the crystal surface
which is removed by evaporation in the granulators.
The second stage is the moisture remaining on the
crystal in the form of a saturated film of syrup. The
latter moisture may be removed either by vaporisation
through elevated temperatures or, if by cooling,
a state of supersaturation is reached where further
crystallisation occurs with the release of more moisture.
Now the slow release of this additional moisture at
the lower temperature of sugar and surrounding air
will surely create a problem where "cooling before
packing" is advocated.
The third stage of moisture removal is that moisture
described by Powers as "bound" moisture and
by others as "inherent" moisture. It is located within
the sugar crystal and cannot be removed in the
granulators but migrates to the crystal surface during
a period of 7 to 10 days. It has been shown that more
than 50 per cent of "bound" moisture is liberated
within 48 hours. 5 Graph I.
The removal of this moisture is now normally
achieved by aeration at the base of the silo or bin with
warm dry air at a pressure of approximately 5 lb. per
square inch.
5. Temperature must be regarded as an essential
controlling factor in the removal of all stages of
moisture. Furthermore, a constant sugar-bin temperature
would assist in maintaining a steady discharge
rate and a steady bulk density, both necessary factors
for a high speed packing unit.

Proceedings of The South African Sugar Technologists' Association—March 1966 215
Storage
In February 1965 a considerable quantity of refined
sugar suffered damage from dampness while in storage,
possibly as a result of extreme weather conditions.
It is an accepted procedure to seal stores containing
packed refined sugar without conditioning the air.
If we examine the reasons for sealing the store we
must conclude that the intention is to prevent changes
in temperature and humidity prevailing within the
store.
Some observations made at the time in the Refinery
stores were:
1. The dampness was observed on February 17th
1965. (Graphs II and III).
2. The block stack affected was on the west side of
the store which is exposed to the afternoon sun.
3. The block stack was built to the maximum height
of the roof trusses.
4. The roof and walls arc constructed of asbestos
on steel frames and. the roof has a shallow pitch
with very little ventilation.
According to information received from the Louis
Botha Meteorological office the ambient temperature
at 2 p.m. on 17th February rose to 32.4° C, with a
relative humidity of 65 per cent. This would provide
an atmosphere with a dewpoint temperature of 27° C.
and a humidity of over 0.020 lb. of water per lb. of dry
air. Here we have exceeded a condition of equilibrium
between saturation temperature and the hygroscopic
property of sugar above which moisture is absorbed
by the sugar from the atmosphere. This condition has
been described as the Equilibrium Relative Humidity
of Sugar.
There is little doubt that the temperature of the air
above the sugar in the Refinery store exceeded that
recorded by Louis Botha on 17th February, due to
the absorption of heat and the insulation provided by
the asbestos roof.
Conclusion
In the knowledge of the many observations made
previously and with our observations over the past
year, the following conclusions are drawn:
1. Elevated temperatures with a minimum of warm
dry aeration are both necessary for optimum
conditioning of refined sugar after discharge
from the granulators.
2. Sealed storage is to be commended but with
provision for forced exhaust ventilation when
temperatures reach the critical stage where the
hygroscopic property of sugar comes into play.
3. The choice of construction material for the
storage roof is significant when considering heat
deflection. Aluminium is recommended with
either timber or precast concrete framework.
Asbestos absorbs and retains a considerable
amount of heat while steel framework is a focal
point for moisture condensation.
4. Good housekeeping is necessary to keep pockets
free of spilled sugar. Exposed sugar will be the
first to absorb moisture which in turn will be
transmitted through the paper pocket.
Acknowledgments
The author wishes to acknowledge the contributions
made to this paper by members of the Laboratory
and Refinery Staff. Also to record our appreciation
to Messrs. J. Lorden and E. Ducel of C. G. Smith &
Company Limited, and Mr. A. J. J. Smith of the
Louis Botha Meteorological Office for their discussions
and information supplied.
References
1. Alexander, J.B. (1965) Unpublished Reports.
2. Powers, H. E. C. "Sucrose Crystal Studies" I.S.J. Vol. 58,
September 1956.
3. Neilson, A. P. and Blankenbach "Investigations into Sugar
Boiling". S.I.T. Proceedings, 1964.
4. Baldt, G. H.and Compton, E. F. "Density of Refined Sugar
in Bins and Silos." S.I.T. Proceedings, 1960.
5. Schwer, F. W. "Some Fundamentals in Drying Granulated
Sugar". S.I.T. Proceedings, 1964.

216
Proceedings of The South African Sugar Technologists' Association

Proceedings of The South African Sugar Technologists' Association—March 1966
217

218 Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966
Mr. Buchanan: Bearing in mind the general drying
rate curve which shows a constant drying rate period
followed by two distinct falling rate periods, as would
correspond to control by moisture diffussion rates
successively from syrup film to air through the syrup
layer and through the crystal lattice, it would appear
to me that temperature is not necessarily the essential
controlling factor in all stages of drying. If the outer
syrup layer exists then air velocity and humidity
219
should also be an essential controlling factor in the
mosture migration in sucrose.
Mr. Alexander: Mr. Buchanan's reasoning is quite
correct but in the case of refined sugar storage the
syrup layer is thin compared with crystal size and
therefore most of the migrating moisture is in the
crystal lattice. For this reason the effect of temperature
in controlling moisture migration is much larger than
the other effects.

220 Proceedings of The South African Sugar Technologists' Association—March 1966
WEATHER REPORT FOR THE YEAR 1st JUNE 1965 TO
General Scope of Report
This report records the weather experienced along
the South African sugar-belt during the year ending
31st May, 1966, and compares it with data accumulated
in the past. As in previous years, the report
will deal primarily with the rainfall recorded by 54
measuring stations scattered throughout the canegrowing
areas from Port Shepstone in the south to
Pongola in the north. Other climatic data quoted,
such as evaporation rates and soil and air temperatures,
refer specifically to Mount Edgecombe where
these readings were taken. These figures will, however,
reflect broadly the conditions prevailing in the rest
of the area.
Rainfall during the year under review will be discussed
in some detail. In addition, the rainfall experienced
during the year June, 1964 to May, 1965
will be referred to, since the crop being harvested this
season will have been influenced by the weather
during both years.
Tabulated Data
Table I gives the annual rainfall recorded at each
of the 54 measuring stations for the past 5 years.
Table II indicates the mean monthly rainfall during
the past year for each of the magisterial districts
covered by this survey, as well as for each of the
3 main sub-divisions.
In Table III can be seen the calculated mean rainfall
for the past 42 years, as well as the monthly percentage
distribution. Also given are the actual mean
monthly rainfall figures for all recording stations,
plus the corresponding evaporation figures for the
Experiment Station. The evaporation figures are
recorded from an open water surface in a square
"Symons" tank.
Table IV gives the rainfall distribution for 2 years
according to growing periods for the magisterial
districts and for the main sub-divisions.
Table V gives the monthly rainfall for the 54
centres for the past 4 years, and also the rainfall
deficiency, if any, per month.
Table VI is a list of the maximum, minimum, and
mean screen temperatures as recorded at the Experiment
Station during the past year, plus the comparative
mean figures over the past 38 years.
Table VII lists the mean monthly earth temperatures
at Mount Edgecombe over the past year, as well as
the figures for the past 31 years for comparison.
31st MAY, 1966
By K. E. F. ALEXANDER
TABLE 1
Rainfall for 54 Centres

Proceedings of The South African Sugar Technologists' Association—March 1966
Comments on Rainfall
During the past year the South African sugar
industry has suffered a repetition, on a smaller scale,
of the disastrous summer drought of 1964/65. It is
true that the drought of 1965/66 was less severe, and
less prolonged than the previous one. Nevertheless
an appreciable reduction in cane-growth has occurred
this season as the result of moisture shortage during
February, March and April.
Total rainfall for the year under review was 39.17
inches, or slightly higher than the computed mean
annual figure of 38.23 for the past 42 years. Above
average rains fell during the period June to October
inclusive. November and December were drier than
usual, but good rains fell in January, 1966. February
followed with below average rainfall, and March
(normally our wettest month) had only 0.68 inches,
or 13% of the 42-year average for this month. Cane
fields were still dry in April, but welcome showers
during May relieved the position considerably.
Monthly details
The following is a more detailed month by month
report for the past year. The rainfall for June, 1965
was a very satisfactory 4.29 inches, or nearly three
times the average. Cane greened up well, but little
growth occurred due to a cold spell. Frost damage
was reported from many areas. During July the sugar
belt had slightly above average rainfall. Nevertheless
crops were reported to be dry in many parts.
Early spring rains fell at the end of August when
most centres reported about 3 inches of rain during
the last 4 or 5 days of the month. September followed
with reasonable showers, and good growth was
anticipated with the onset of warmer weather. Although
further good rains fell in October, air and
soil temperatures were still low. This lack of heat
continued into November and retarded growth
despite fairly good rainfall that month. December's
cool, dry conditions also did not permit maximum
cane growth.
The month of January saw the South African cane
crop off to a fine start for 1966. Good steady showers
during the month averaged 6.65 inches for the 54
recording centres throughout the area. Satisfactory
soil moisture conditions, combined with sharply
higher soil and air temperatures, got the cane growing
really well for the first time this season. This rapid
growth continued until mid-February, when lack of
moisture and cooler weather slowed things down.
The worst affected fields were those in the shallower
sandy parts of the North Coast.
It was not only the Ides which boded ill for the
sugar industry in March. During the entire month
only 0.68 inches of rain fell. This is the lowest March
rainfall figure recorded for the sugar belt, and contrasts
with the 22.52 inches of rain which fell during
March, 1925. Cloudless sunny days lifted the evaporation
figure by 30% for the month. At the end of March
conditions were serious, with brown patches of dead
cane apparent in places. Odd showers during April
did little to alleviate the drought. Soil moisture was
at a low ebb, and good soaking rains were needed to
bolster the crop against the traditionally dry winter
period.
The May rainfall was very beneficial to the North
and South Coasts, but was well below expectations
for Zululand. Thus by the 31st May, 1966, most of
the cane crop was green and in fine fettle. Some areas
around Empangeni and Mtubatuba, however, were
still badly affected by drought.
Two-year Summary
The following paragraph is a brief review of weather
conditions experienced over the past two years. In
June and July, 1964, frost occurred at many points
in the sugar belt. Dry conditions prevailed until
October, when excellent soaking rains fell. The worst
summer drought ever recorded in the industry took
place from November right through until the end of
May, 1965, when widespread rains brought relief
to most centres. In June frost was again reported from
many areas. Rainfall in early winter was satisfactory
but soils dried out progressively until the end of August
when good rains fell. The next three months were
moist but cool. December tended to be dry and cool.
January, 1966, provided optimum growing conditions
which followed through into mid-February. From
this stage onwards, however, the soils became increasingly
drier and the industry suffered a short
but severe drought. Good rains fell over most of the
sugar belt during May. Thus by May 31st 1966, South
African cane fields were (with the exception of those
in Northern Zululand) quite moist, and the crops
carried were green and healthy.
Temperatures
The mean screen temperature for the year under
review was 67.6° F. at the Experiment Station. This
was 1.1° F. cooler than the 38 years' mean. With the
exception of September, January and March, all
months from June, 1965 to May 1966 were below
normal in regard to air temperature. The average
soil temperatures were also consistently lower than
in previous years. The grass minimum temperature,
however, did not once fall below freezing point.
Evaporation
This year the evaporation from a free water surface
was above normal by 3.82 inches. Nevertheless, the
rainfall distribution was such that a fairly normal
rainfall deficiency resulted (see table V). Unfortunately,
the bulk of this deficiency occurred during the vital
growing months of December, February, March and
April. The deficit for March is almost certainly the
highest recorded for one month in the South African
cane belt.
Hours of Sunshine
221
During the year, Mount Edgecombe has had
2406.4 hours of sunshine, representing 1.4% more
than the 39-year average. July. October and December
were sunnier than usual, with March being well above

222
Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE II
Rainfall in Inches by Districts for Months of June, 1965, to May, 1966 inclusive
TABLE III
Rainfall and Evaporation Data

Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE LV
Rainfall in Inches by Districts for the Two-year Period June, 1964 to May, 1966 inclusive
TABLE V
Rainfall and Evaporation in Inches for the Past Four Years
223

224 Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE VI
The following are the Screen Temperatures by Months in Degrees Fahrenheit at the Experiment
Station for the Year June, 1965 to May, 1966, compared with the Means for the Period 1928 to 1966
TABLE VII
The following table gives the mean monthly earth temperatures

Proceedings of The South African Sugar Technologists' Association—March 1966
average. September, January and February were more
cloudy than usual, with November being exceptionally
overcast. The other four months had average hours
of sunshine.
Wind
The anemometer in the meteorology site at the
Experiment Station recorded 38,475 miles of air as
having passed the site during the past year. This
represents an average wind speed of 4.4 m.p.h. over
the entire period. Based on figures for only two years,
the wind pattern ranges from 3.2 m.p.h. for the
month of July up to 5.8 m.p.h. for November and
December.
Conclusions
225
Adverse growing conditions have been experienced
by the South African sugar industry during the past
two years. The winter of 1964 brought frost damage
to many areas. Early spring was dry. Good rains
fell in October, but from then until the end of May
a very grave drought retarded cane growth tremendously.
The winter of 1965 saw another, but less
severe, bout of frost affect some cane fields. Spring
and early summer were moist and cool. The first
six or seven weeks of 1966 provided ideal cane growing
weather. This deteriorated into a short severe
autumn drought which was relieved only in May.
Parts of Zululand however, were still dry at the end
of that month.

226 Proceedings of The South African Sugar Technologists' Association- March 1966
AUTOMATION OF A ROUTINE LABORATORY
Introduction
With the recent expansion of the sugar industry in
South Africa a substantial increase in the number of
soil and leaf samples submitted to the Fertiliser
Advisory Service of the South African Sugar Association
was predicted. To maintain the efficiency of this
laboratory it was necessary to revise many of the
existing analytical techniques. Procedures used for
normal quantitative analysis of soils and leaves involve
preparation of the sample, extraction of elements
from the sample, filtering and dilution of the
by R. T. BISHOP and. T. D. ALGIE
extract and determination of the concentrations of
elements present. Some suggestions to facilitate each
of these operations are presented below.
1. Grinding of Soil Samples
FIGURE 1. Plan of a simple soil grinder.
A simple soil grinder developed at the Soils Testing
Laboratory at Raleigh by Dr. P. D. Reid (Director,
Soils Testing Division, North Carolina, Department
of Agriculture) can be easily constructed. The plans
of a modified version of this grinder being tested at
Mount Edgecombe are presented in Figure 1.

Proceedings of The South African Sugar Technologists' Association—March 1966 227
The two case-hardened steel rollers which rotate
in opposite directions are spring loaded at 2 mm.
apart and are driven by a 0.33 h.p. electric motor.
The rollers are swept clean by the two brushes mounted
below and parallel to them. After passing through
the rollers the soil is funnelled into a receiving vessel.
Sieving of the soil is a separate operation. To eliminate
cleaning of the bench after grinding and sieving of
each sample, these operations are carried out over a
wire frame mounted above a funnel and dust bin.
By making the ratio of the distances from the ends
of the balance arm to the knife edge (A) 10:1, a
weight ratio of the same order is possible. In the
extraction of exchangeable cations (normally exactly
lOg. of soil are weighed out and leached with exactly
100 ml. of N ammonium acetate) a spoonful
of soil which weighs approximately lOg. is added to
the tared container (B). The addition causes the
balance arm supporting the soil to swing down and
contact is made between the metal disc (C), which is
insulated from the rest of the balance arm, and the
micro-switch (D). This contact completes an electrical
circuit which opens the solenoid valve (E). Ammonium
acetate is injected into the tared beaker (F)
2. Weighing
Normal quantitative chemical analyses involve the
addition of a known volume of solution to a known
weight of material. Provided the ratio of solution to
material is kept constant it is not necessary to know
the exact weights and volumes used. A "ratio-balance"
employing this principle was observed at Oosterbeek
(Laboratory of Soil and Crop testing, Oosterbeek
Holland) and an instrument designed at the Experiment
Station of the South African Sugar Association
to achieve the same end is illustrated in Plate 1.
PLATE 1. Balance used at Mount Edgecombe for providing a solution to soil ratio of 10:1.
until the weight ratio of solution to soil reaches 10:1.
At this point the balance arm swings upward, breaking
the electrical circuit and closing the solenoid valve.
3. Handling of Glassware
Where a number of samples are to be subjected to
the same treatment, the speed of each operation e.g.
pouring, filtering, washing etc. can be greatly increased
by fixing the required glassware permanently
in 18 gauge aluminium racks. In this way ten or more
samples can be handled simultaneously. The tray
sizes adopted at the Experiment Station are 26 in.
long by 3 in. wide (height measurements will vary

228
with the glassware used). Two such racks, one used
for the extraction of acid soluble P from soils and the
other for receiving the filtered solution, are included
in Plates 2 and 4 respectively.
4. Extraction
To determine the concentrations of elements in
biological material it is generally necessary to bring
Proceedings of The South African Sugar Technologists' Association—March 1966
them into solution. This procedure normally requires
shaking of some kind and two shakers which are
operating satisfactorily are shown in Plates 2 and 3.
(a) End over end shaker
PLATE 2. End over end shaker capable of holding 110 vessels.
The aluminium rack (A) is held in position
by the stopper (B). This stopper consists of a
strip of 2 in. thick foam plastic covered by
rubber which is tacked to a half-inch thick
wooden board. Bolts (1 in. x 4 BA) fixed to a
hinge on the end of the board are slipped
through holes in the bracket (C). The other
end of the stopper is screwed down (D) effectively
sealing the mouths of all 10 containers.
The stoppers are removed and washed between
extractions.
(b) Vertical shaker
The vertical shaker used at Mount Edgecombe
was designed by R. A. G. Rawson
Rothampstead Expermental Station, Harpenden,
England, formerly of S.A.S.A. Experiment
Station) and makes use of the block of a
motor car engine in which pistons 1 and 4
fired simultaneously. (See Plate 3).
The engine block (A) is mounted on concrete
supports and the piston rods of cylinders 1 and
4 (B) have been extended upwards above the
The shaker presented in Plate 2 can hold 110
vessels and the wooden drum is 27 in. long.
Other measurements are given in Figure 2.
level of the bench. The frame of the shaker (C)
is secured to these rods. A flywheel is fixed
to the front of the crankshaft and a 0.33 h.p.
electric motor (D) provides the necessary drive.
5. Removing Aliquots
The possibility of using a vacuum-compressed air
system for removing equal aliquots simultaneously
from a number of different solutions was suggested
by Dr. Reid. A piece of equipment utilising this idea
has been constructed at Mount Edgecombe and is in
daily operation. (See Plate 4).
When the system is subjected to a vacuum with
the tips of the pipettes in solution a constant suction
is maintained when incoming air is drawn down the
glass tube (A), which is immersed in the container of
water (B), to compensate for the air removed. By
raising or lowering the depth of this tube in the water
the degree of vacuum in the whole system can be
altered. To remove exactly the same aliquots in all ten
pipettes it is essential that the distances between their
tips and graduation marks be the same.

Proceedings of The South African Sugar Technologists' Association - March 1966 229
FIGURE 2. Front view of end over end shaker.
PLATE 3. Vertical shaker modified from the block of a petrol engine.
s l \ •»

230
Proceedings of The South African Sugar Technologists' Association—March 1966
PLATE 4. Apparatus used for removing 25 ml. aliquots simultaneously from 10 solutions.
The steps involved in the operation of the apparatus
are:
(a) The rack of solutions to be sampled (C) is
lined up under the pipettes. The pipettes are
lowered manually into the solutions and the
high vacuum switch is turned on.
(b) When the pipettes are approximately 75 per
cent full (after about 2 seconds) the high
vacuum switch is released and the low vacuum
switch is depressed.
(c) When the levels of solutions in the pipettes
are maintained at the predetermined heights
the pipettes are lifted out of the solutions.
The counter weight (D) permits the pipettes
to be raised easily and smoothly.
(d) A rack of receiving vessels is inserted under the
pipettes and the compressed air switch (which
isolates the water cylinder from the rest of the
system and exposes the pipettes to a compressed
air supply) is depressed thus ejecting the solutions
into the receiving vessels.
(e) The receiving vessels are removed and the
water supply is turned on releasing a supply of
deionised water (E) into each pipette.
The circuit diagram of the electrical system of the
apparatus is given in Figure 3.
6. Measurement
Two techniques commonly used in the laboratory
for measuring concentrations of elements are colorimetry
and flame photometry.
Tubes for the colorimeter are expensive and the
general procedure is to use only one tube which must
be rinsed and then refilled before the colour intensity
of each solution can be measured. Although modern
instruments permit the filling and emptying of the
colorimeter tube to be carried out automatically,
most laboratories still have the older type. By using
thin walled glass specimen tubes in place of the colorimeter
tubes results of an accuracy acceptable for
fertiliser advisory work have been obtained. Colour
development is carried out directly in these tubes
which are then inserted into the colorimeter.
Filter type flame photometers are commonly used
for K, Ca and Na determinations. For the determination
of concentrations of each of these elements in
the same solution a different light filter must be inserted
into the instrument. This means that for K,
Ca and Na determinations each solution must be
atomised three times. A modification of this instrument
which permits the three elements to be determined
during a single atomisation is illustrated in
Plate 5.
The unit consisting of focusing lens, filter holder
and selenium cell found in the conventional instrument
(A) has been reproduced on the opposite side
of the flame (B). The new guides for the filter have
been extended (C) to accommodate two filters one
mounted above the other. The Ca filter is placed above
the K filter, and to increase the sensitivity of the
instrument to the former element, the reflector (D)
is fixed opposite to it. The Na filter is inserted in the
normal filter holder.
When operating the instrument the Ca filter and
reflector are pushed down into line with the selenium

Proreedings of The South African Sugar Technologists' Association—March 1966 231
cell. Knob (E) is set on position 1 and the blank i.
set on O using knob (F). The top standard is set on
100 with knob (G). The Ca filter is pulled upwards
and the K filter automatically lines up with the cells
Knob (E) is set on position 2 and the top standard
is set on 100 using knob (H) and the reading of the
blank is noted. Knob (E) is set on position 3 and
knob (J) is used to set the top Na standard on 100.
The blank reading is again noted. The unknown
solutions are then atomised and by altering the setting
of knob (E) the concentrations of Ca, K and Na are
reflected on the common galvanometer. The filters
for Ca and K must be changed each time and the
blank readings for K and Na subtracted from the
observed results.
PLATE 5 Modified filter type flame photometer used for
determining K, Ca and Na concentrations during a single atomisation.
A circuit diagram of this instrument is presented in
Figure 4.
Summary
Descriptions, together with diagrams and/or plates,
are given of apparatus suitable for (1) grinding soils,
(2) measuring out solution to solid ratios of 10:1,
(3) handling 10 glass containers simultaneously,
(4) shaking either end over end or vertically, (5) removing
equal aliquots simultaneously from 10
different solutions and (6) measuring the concentrations
of K, Ca and Na during a single atomisation
of the solution into a filter type flame photometer.

232
Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 3. Circuit diagram of the electrical system of the apparatus used for removing equal aliquots.

Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 4. Circuit diagram of modified filter type flame photometer.
233

234 Proceedings of The South African Sugar Technologists' Association-March 1966
DETERMINATION OF COPPER AND ZINC IN SUGAR­
CANE LEAVES BY ATOMIC ABSORPTION
Introduction
Since the introduction by Walsh 12 in 1955 of atomic
absorption spectrophotometry as an analytical method,
it has found application in many fields as a means of
rapid analysis of certain metallic elements at very low
concentrations. The analysis of agricultural materials,
i.e., leaves, soil extracts and fertilizers is one sphere
in which atomic absorption has been applied by several
workers 1 , 2 , 4 , B for the analysis of Na, Mg, Ca, K
as well as Cu and Zn. This method has a number of
advantages over the usual colorimetric and flame
photometric determinations of these elements; (a) it
is relatively free from interference by other elements
in the sample, (b) it has a higher sensitivity than the
flame photometric methods and (c) it is much faster
than the usual colorimetric methods.
Atomic absorption spectrophotometry is based on
the principle that free atoms of an element can absorb
the characteristic radiation, known as resonance radiation,
of that specific element. This phenomenon is
well known in producing the Frauenhofer lines in the
solar spectrum. The amount of light absorbed by the
atoms of the element is proportional to the number
of atoms through which the light beam is passed, and
therefore proportional to the concentration of the
element in the sample, provided that the sample is
atomised at a constant rate.
The atomic absorption spectrophotometer consists
essentially of a light source producing the resonance
radiation of the element, a flame for atomizing the
sample, a means of isolating the resonance radiation
from unwanted background and non-resonance radiation,
and a photo-electric detector to measure the
intensity of the light after it has passed through the
flame. The operation of the instrument is closely
analogous to a spectrophotometer in which the absorption
cell has been replaced by the flame, and the
light source usually emitting continuous radiation by
a lamp emitting the resonance radiation of the element.
A hollow cathode lamp operating at a low current is
used as a light source for all but the alkali elements
where discharge lamps can be used. The current
through the lamp has to be kept constant in order to
avoid changes in intensity and sensitivity when analysing
a series of standards or samples. It is obvious
that a long absorbing path would increase the sensitivity,
that is, the percentage of incident light absorbed
per part per million of the element in the sample. The
burner head is therefore designed to provide a long
thin flame as opposed to the small confined flame
required in flame emission spectrophotometry.
As the number of atoms in the flame is not only
dependent on the concentration of the element in the
sample but also on the rate at which the sample is
sprayed into the flame, the rate of sample uptake
must also be closely controlled to prevent changes in
By P. DU PREEZ
sensitivity. This can be accomplished by stabilizing
the flow of air aspirating the sample into the flame.
Spectrometers for isolating the resonance radiation of
the elements have degrees of sophistication varying
from simple filter separation to double beam spectrophotometers
in which the light beam from the hollow
cathode lamp is split into two separate beams, one of
which is used as a continuous monitor of the intensity
of the hollow cathode lamp, and the other one passed
through the atom cloud in the flame.
As the atomic absorption method, like flame photometric
and colorimetric methods, does not provide an
absolute measurement of the concentration of an
element in a sample, the use of standard solutions for
comparison is necessary, and it is therefore essential that
no changes in sensitivity occur while determining a
series of standards and unknown samples. Variations
in air flow, air pressure and viscosity of the sample
alter the rate of sample uptake which influences the
density of the cloud of atoms in the flame. Variations
in fuel flow to the flame change the temperature and
character of the flame which may in turn affect the
degree of breakdown of the sample. Fluctuations of
the current through the hollow cathode lamp affect
the amount of absorbable light passing through the
flame, altering the percentage of incident light absorbed
for a specific concentration. All these variables
except the emission from the hollow cathode lamp
are fairly easy to control if the necessary care is taken
while a series of determinations is being carried out.
The atomic absorption spectrophotometer used for
this work was the Perkin Elmer model 303, with a
burner regulator providing control of air and fuel flow
and standard burner using an air-acetylene mixture.
The instrument was obtained in order to facilitate the
routine analysis of Zn and Cu in plant tissue for
fertilizer advisory purposes.
Currently leaf samples are routinely analysed for
Na, K, Mg, Ca, P and N after being digested with
sulphuric acid in the presence of selenium as a catalyst.
Zinc and Cu are determined colorimetrically with
diphenylthiocarbazone (dithizone). This employs a
separate sub-sample digested with a mixture of nitric
and sulphuric acids, as the selenium used in the
digestion for the macro-elements also forms a diphenylthiocarbazone
complex interfering with the Cu and
Zn determinations. For both methods of digestion
2.5 gm. of leaf sample are taken and made up to
100 ml. after digestion, thus diluting the constituents
in the leaf 40 times. With this dilution the extracts
prepared with sulphuric acid and selenium, and those
prepared with nitric and sulphuric acids contain 20
per cent and 8 per cent v/v sulphuric acid, respectively.
The requirements for a procedure to replace the
colorimetric method for Cu and Zn would be (i) an
accuracy at least equal to or better than the colori­

Proceedings of The South African Sugar Technologists' Association—March 1966
metric method and (ii) a higher speed and greater
simplicity of analysis plus a minimum of sample preparation.
It would also be advantageous if the determination
could be carried out on the same digest as
the macro-elements and if both Cu and Zn could be
determined on the same solution.
Preliminary Investigation
From the work of Bradfield and Spincer 4 , it seemed
hopeful that both Zn and Cu could be determined
directly on the selenium-sulphuric acid digest, thus
eliminating a separate digestion for the trace elements.
Work carried out in establishing the optimum operating
conditions given in Table I showed that adequate
sensitivity and accuracy could be obtained for Zn.
confirming results reported by David 5 and Bradfield 3 ,
TABLE 1
Standard Operating Conditions
The method suggested by Wilson 13 , where the detection
limit of a method is expressed as a multiple of
the standard error of the background, was used to
determine the detection limits for Cu and Zn. The
standard error of the background was found to be
0.02 ppm. for Zn and 0.014 ppm. for Cu. This gives
limits of detection for Zn and Cu respectively of 0.06
ppm. and 0.04 ppm. in solution, and 2.4 ppm. and
1.6 ppm. in the undiluted leaf material. As the critical
limit for the Cu content in the leaf is taken for fertilizer
advisory purposes as 3 ppm., the sensitivity for Cu
is inadequate to ensure accurate determinations at the
concentrations near this limit. Bradfield and Spincer 4
obtained satisfactory results because the Cu content
in their samples varied from 12 to 16 ppm. in the
original plant material, which is appreciably higher
than the usual concentrations of 2 to 10 ppm. Cu
occurring in sugarcane leaves.
Determination of Cu and Zn after Concentration
by Organic Solvent Extraction
The difficulty experienced in the determination of
low concentrations of trace elements has been overcome
by Allan 1 , and Strasheim, Eve and Fourie 9 by
concentration of the trace elements through their
extraction with an organic solvent, and also by
Strelow 11 , who combined organic solvent extraction
with a cation exchange separation. As the atomic absorption
method is free from interference by Fe as
opposed to some emission spectrometric methods, it is
unnecessary to include an ion exchange separation of
Fe from the other trace elements. Therefore only the
possibility of the simultaneous extraction of Cu and
Zn with an organic solvent was investigated.
Ammonium pyrrolidine-dithiocarbomate was suggested
as a complexing agent by Malissa and Schoffman
6 , who pointed out that both Zn and Cu formed
a complex with this reagent at pH levels from 2 to 9,
while selenium formed only a weak complex with it.
Similar reagents were used for the concentration of
trace elements by several workers 1 , 8 , 10 , 11 and it was
hoped that it would be possible by correct choice of
pH to eliminate interference of the complexing agent
by selenium.
A quantity of ammonium pyrrolidine-dithiocarbomate
was prepared by the method of Malissa and
Schoffman 6 from redistilled and purified reagents.
Unfortunately the excess of selenium in the leaf samples
digested with selenium and sulphuric acid was so
great that it was impossible to eliminate the interference
of selenium by choice of pH. Other methods
of removing the selenium from solution by oxidation,
or precipitation as elemental selenium, also proved
unsuccessful, and the separate digestion with nitric
and sulphuric acids has therefore to be used.
The effect of pH on the extraction is shown for
three different concentrations of Zn and Cu in Figure
1, and demonstrates that the extraction of Cu varies
very little with pH over the range pH 3.5 to pH 5.8,
but that a sharp decrease in the efficiency of the Zn
extraction occurs at pH 4.7. This means that the
extraction should be carried out at a pH between 5.5
and 5.8 which is contrary to the report of Allan 1 , who
obtained 100 per cent extraction of Zn from pH 2.5
to 5.0.
Procedure
Reagents
1. A saturated solution of sodium acetate purified by
extraction with dithizone and CCl4.
2. A 1 per cent aqueous solution of ammonium pyrrolidine-dithiocarbomate.
3. Methyl isobutyl ketone purified if necessary by
extracting with 6N HC1.
Method
A leaf sample of 2.5 gm. is weighed out and digested
with 5 ml. H2SO4 + 15 ml. HNO3 until, when white
fumes of H2SO4 are evolved, a clear solution is obtained.
Additional portions of HNO3 are added if
necessary to obtain a clear solution. After cooling, the
leaf extract is made up to 100 ml. A 15 ml. aliquot
235

236 Proceedings of The South African Sugar Technologists' Association—March 1966
is then transferred to a 50 ml. separating funnel after
which 15 ml. buffer solution, 1 ml. 1 per cent ammonium
pyrrolidine-dithiocarbomate, and 4ml. methyl
isobutyl ketone are added. The separating funnel is
shaken mechanically for three minutes and the organic
phase run into a thimble. A series of standards containing
0.1 to 1.0 ppm. Zn and 0.1 to 0.4 ppm. Cu
in 1.1 N H2SO4 is made up and extracted in the same
way. The solutions and standards are aspirated with
the instrument settings given in Table I, the zero of
the instrument being set with pure methyl isobutyl
ketone. The percentage absorption for each sample
and standard is measured, and the Cu and Zn in the
unknown samples are calculated from the working
curve obtained from the standards.
The coefficient of variation for duplicate readings,
taken over 20 extractions of the same sample was 13
per cent for a concentration of 0.4 ppm. Zu, and 10
Figure 1. Effect of pH on extraction.
per cent for Cu at a level of 0.1 ppm. A comparison
between the results obtained for five leaf samples with
the colorimetric method and the atomic absorption
method following the extraction procedure is given in
Table II.
TABLE II
Comparison between Results obtained with the Dithizone
Colorimetric and Atomic Absorption Spectrophotometric
Method after Concentration

Proceedings of The South African Sugar Technologists' Association—March 1966
Direct Determination of Cu and Zn
on the Leaf Extract
A direct determination of Cu and Zn on leaf samples
digested with nitric and sulphuric acids and diluted
1:10 instead of 1:40 was undertaken, but difficulty
was experienced due to blocking of the flame slit by
the high concentration of solids or sulphuric acid in
solution. When samples were digested with a mixture
of nitric and perchloric acids and diluted 1:10, the
burner slit no longer became blocked, showing that
the previous blocking was not due solely to the high
solids content of the samples, but was aggravated by
the sulphuric acid. As the digestion of plant material
with nitric and perchloric acids may be dangerous,
this method is unsuitable for routine use, and it is
therefore necessary to add a small quantity of sulphuric
acid to the sample before ashing according to
the method of Piper 7 . It was found that if the sulphuric
acid used in the digestion was limited to
approximately 8 per cent in the final solution, blocking
of the burner slit was lessened to such an extent that
it was no longer troublesome. No interference with
the absorption of Cu and Zn by Na, K, Ca, Mg or
phosphate at the levels occurring in the leaf extracts
was found, thus confirming the results obtained by
Allan 1 and David 5 .
The effect of different acids and acid concentrations
on the readings obtained for Zn and Cu was investigated.
It was found that up to 10 per cent hydrochloric
acid and 8 per cent nitric acid had no effect on the
readings, while sulphuric acid had a marked depressing
effect at the higher concentrations of Cu and Zn as
shown in Figures 2 and 3. At lower concentrations an
enhancement of the Zn readings was observed, most
probably due to contamination of the sulphuric acid
with Zn. Satisfactory results were obtained when 8
per cent sulphuric acid was added to the standards,
leaf samples being analysed according to the following
procedure.
Fig. 2. Effect of different acid concentrations on Cu absorption.
237

238
Procedure
A sub-sample of 2.5 gm. is taken from the ground
oven dried leaf sample and digested with H2SO4:
HCIO4: HN03 in the ratio 2 : 3 : 20 ml. With proper
control of the heating rate it is unnecessary to add
more HNO3 to obtain a clear extract. After adding
5 ml. H2O the extract is transferred without filtering
to a 25 ml. volumetric flask and made up to volume.
The solutions are left overnight to allow the insoluble
residue of silica to settle out completely. When aspirating,
the sample tube of the burner is put directly
into the volumetric flask without disturbing the sediment.
This reduces handling of the sample considerably
thereby saving time and lessening the chances of
contamination. Standards containing 0.4 to 4 ppm.
Zn and 0.04 to 1.5 ppm. Cu in 8 per cent H2SO4 are
prepared, and samples and standards are aspirated
Proceedings of The South African Sugar Technologists' Association — March 1966
with the operating conditions shown in Table I. The
zero of the instrument is set with distilled water and
duplicate readings of per cent absorption for each
sample are obtained by running through the series of
samples and standards twice. In this way any drift in
the instrument can be detected.
Results
Comparative figures for the results obtained from
different digestions of six leaf samples by a method
of addition not described, and direct determination,
both with the atomic absorption spectrophotometer,
and the dithizone colorimetric method are shown in
Table III. The recovery of different amounts of Zn and
Cu added to three leaf samples is given in Table IV,
while Table V shows the reproducibility of the method
as determined from a number of readings on each of
21 digestions of a single leaf sample.
Figure. 3. Effect of different acid concentrations on Zn absorption.

Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE III
Comparison of Results obtained with the Direct Determination on Leaf Extracts
by Atomic Absorption and Dithizone
TABLE IV
Recovery by Atomic Absorption of Zinc and Copper added to Leaf Samples
TABLE V
Reproducibility of Method
This method has proved to be extremely rapid and
it is possible to determine up to 30 samples per hour
for a single element, which is appreciably more than
can be handled when using an extraction method.
Discussion and Conclusion
From Table III it is evident that the accuracy of the
atomic absorption method is not as high as expected,
the results tending to be slightly lower than those
obtained by the dithizone method. This is confirmed
by the results in Table IV for the recovery of added
amounts of Zn and Cu, which also indicate a bias
towards lower values. This inaccuracy is possibly
caused by the high solids content of the plant samples
which might increase the viscosity of the sample, and
therefore decrease the rate of sample uptake. It was
not possible to determine the effect of high concentrations
of solids in the sample, due to Cu and Zn
contamination in the reagents used to prepare synthetic
plant samples.
239

240
The results in Table III obtained with the method of
addition, are much higher than those obtained by the
dithizone method, and the direct determination with
atomic absorption. This is due to a slight nonlinearity
of the working curve. This nonlinearity is more pronounced
for Zn than for Cu (see Figures 2 and 3) and
explains why the results for Cu compared to those for
Zn are closer to the results obtained by the other two
methods.
As the extraction procedure with ammonium pyrrolidine-dithiocarbamate
is more laborious, and less
reproducible than the direct determination, the latter
method is to be preferred unless the concentrations
of Zn and Cu are very much lower than those usually
occurring in sugarcane. The direct determination has
the advantage over both the colorimetric and extraction
methods, in that it is unnecessary to use specially
purified reagents, as the standards contain the same
quantity of sulphuric acid as the samples. The risk of
contamination of the sample is also considerably
lessened by reduced handling of the sample.
It is felt that the advantages of the direct determination
of Cu and Zn by atomic absorption out-weigh
the somewhat lower accuracy of this procedure compared
to the dithizone colorimetric method, and that
it is sufficiently accurate to find application in the
routine determination of Cu and Zn in sugarcane
leaves.
Summary
Two methods for the analysis of Cu and Zn in
sugarcane leaves by atomic-absorption spectrophotometry
are described, and compared with the diphenylthiocarbazone
(dithizone) colorimetric method. In the
first procedure the Cu and Zn are concentrated by
extraction as an ammonium pyrrolidine-dithiocarbamate
complex in an organic solvent, while in the
second the Cu and Zn are determined directly on the
leaf digest.
The direct determination is much faster and more
reproducible than the extraction procedure, and is to
be preferred except when very low concentrations of
Cu and Zn are encountered.
Although the accuracy of the direct determination
by atomic-absorption is somewhat less than can be
obtained by the dithizone method it is quite accurate
enough for all practical purposes.
References
1. Allan, J. E. (1961). The determination of zinc in agricultural
materials by atomic-absorption spectrophotometry.
Analyst 86, 530-534.
2. Allan, J. E. (1958). Atomic-absorption spectrophotometry
with special reference to the determination of magnesium.
Analyst 83, 466-471.
3. Bradfield, E. G. (1964). Leaf analysis as a guide to the
nutrition of fruit crops IV — Scheme for the rapid determination
of copper, iron, manganese and zinc in plant
material. Journal of the Science of Food and Agriculture
15, 469-473.
4. Bradfield, E. G. and Spincer, D. (1965). Leaf analysis as
a guide to the nutrition of fruit crops VI — Determination
of magnesium, zinc and copper by atomic-absorption
Proceedings of The South African Sugar Technologists' Association — March 1966
spectroscopy. Journal of the Science of Food and Agriculture
16, 33-38.
5. David, D. J. (1958). Determination of Zinc and other
elements in plants by atomic-absorption spectroscopy.
Analyst 83, 655-661.
6. Malissa, H. and Schoffman, E. (1955). Uber die verwendung
von substituierten Dithiocarbamaten in der Mikroanalyse
III. Mikrochim. Acta 1955, 187-202.
7. Piper, C. S. (1944). Soil and plant analysis. Hassel Press,
Adelaide, Australia.
8. Stetter, A. and Exler, H. (1955). Naturwissenschaften 42,
45.
9. Strasheim, A., Eve, D. J. and Fourie, R. M. (1959). A
spectrographic method for the analysis of plant material
using sodium pyrrolidine-dithiocarbamate for the concentration
of the trace elements. J.S.Afr.Chem.Inst. 12, 75-80.
10. Strasheim, A., and van der Walt, H. R. (1962). Die konsentrasie
en skeiding van spoorelemente in plantmatsriaal
en die ontleding daarvan volgens die horisontale boogmetode.
J.S.Afr.Chem.Inst. 15, 1-10.
11. Strelow, F. W. E. (1961). Separation of trace elements in
plant materials from ferric iron by ion exchange chromatography.
Anal.Chem. 33, 994-997.
12. Walsh, A. (1955). The application of atomic absorption
spectra to chemical analysis. Spectrochim.Acta. 7, 108.
13. Wilson, A. L. (1961). The precision and limit of detection
of analytical methods. Analyst 86, 72-74.
Dr. Sumner: Were the hollow cathode lamps used
of the sealed type?
Mr. du Preez: The C.S.I.R. advised us to use
locally manufactured lamps that could be regenerated
but we are having so much trouble with them that
we are considering the use of sealed non-regeneratable
lamps.
Dr. Matic: The atomic absorption method suffers
from the disadvantage that the warming up of the
lamps takes a long time. Has polarography been
considered as a method for the simultaneous determination
of copper and zinc? The half-wave
potentials of these two metals are well separated
and the sensitivity is probably adequate, particularly
if a cathode ray polarograph is used. Some additional
metals could also be determined—for example a
method for the simultaneous determination of copper,
lead and zinc has been developed by the Transvaal
and Orange Free State Chamber of Mines Laboratory.
Mr. du Preez: The sample throughput of the atomic
absorption method is so high that even allowing an
hour for the instrument to warm up it is still possible
to handle a hundred samples in a morning.
The polarographic method was not attempted for
the determination of copper and zinc. I had the impression,
however, that this method is slow and time
consuming.
Dr. Matic: With a cathode ray polarograph the
method can be very rapid.

Proceedings of The South African Sugar Technologists' Association — March 1966
NITRIFICATION IN RELATION TO pH-ITS IMPORTANCE
IN FERTILIZER NITROGEN UTILIZATION BY CANE IN
SOME SUGAR BELT SOILS*
Introduction
The importance to the cane plant of an adequate
supply of inorganic nitrogen, during the first few
months of growth, either in the ammonium (NH4-N)
or nitrate (NO3-N) form, has been clearly demonstrated
by recent work at the Experiment Station, both
in the field 2 and the greenhouse 9 .
It is well known, however, that leaching losses of
fertilizer nitrogen applied at the onset of growth, are
able to deprive the plant of much of its N supply
before this can be fully utilized, especially on sandy
soils. These losses can be reduced where the applied
N is retained by the soil in the NH4-N form for a
reasonable period before being wholly converted to
NO3-N which is subject to more pronounced leaching.
The conversion of NH4-N to NO3-N in the soil
by bacteria of the genera Nitrosomonas and Nitrobacter
is referred to as nitrification. Providing that
conditions such as soil moisture, temperature, and
oxygen supply are optimum for nitrate production,
the other most important factor governing the rate of
nitrification is soil pH. Robinson 7 has pointed out
that if pH is not limiting and NH4-N is readily available,
then it will be converted to NO3-N quite
rapidly. Where pH is too low for optimum growth of
nitrifying bacteria, the population will be slow to
develop and NO3-N will only accumulate gradually.
By R. A. WOOD
TABLE I
Variable Nitrification during Mineralisation of Soil Nitrogen
(incubated at 35° C. and 60% W.H.C.)
241
Since the Sugar Belt contains a wide range of soils
of variable pH, their rates of nitrification must influence
the utilization by cane of added fertilizer N.
particularly on sandy soils. With this in mind various
aspects of nitrification in relation to pH for some of
the main soil groups were investigated.
Delayed Nitrification During Soil Nitrogen
Mineralisation
It had been previously demonstrated 8 from a series
of N mineralisation studies on a range of cane soils,
that while ammonification usually proceeded rapidly
some of them showed a marked delay in nitrification.
Comparative figures giving the amount of NH4-N
and NO3-N present before and after an eight week
period of incubation are presented in Table I.
From this data it is clear that in the soils of low
pH a partial or virtually complete delay in nitrification
had occurred, being most noticeable on the highly
acid Lytton and Cartref sands.
The Effect of Liming on Delayed Nitrification
To confirm that soil acidity was primarily responsible
for the delay in nitrification just described, the
following experiment was carried out.
Procedure. Duplicate air dry samples of six soil
groups, of which four had previously shown delayed
* This work will form part of a thesis to be submitted to the University of Natal, Soil Science Department, for the Ph.D. degree.

242
nitrification, were treated with limestone (CaCO3) at
the following rates; nil, 2, 4 and 8 tons per acre. The
soils were moistened to 60 per cent water holding
capacity and incubated for a period of four weeks at
35° C, after which the amounts of NH4-N and
NO3-N present were determined.
Results. The data obtained for the various soil
groups is shown in Table II. This clearly indicates that
once the pH has been raised sufficiently by the limestone
to allow the nitrifying bacteria to proliferate, the
changeover to NO3-N rapidly occurs, and that in
some cases liming apparently gives rise to increased
production of nitrate during mineralisation.
The Effect of Ammonium Nitrogen Carriers
on Nitrification in Sandy Soils
Just as liming affects the rate of nitrification in acid
soils by changing their pH, it can be expected that the
Proceedings of The South African Sugar Technologists' Association — March 1966
TABLE II
addition of various types of ammonium fertilizers will
also influence pH and nitrification rate. This could be
particularly important in sandy soils subject to marked
leaching losses especially of NO3-N (see Table III).
Soil pH and the fertilizers themselves might well be
responsible for differential utilization of fertilizer N by
cane on these soils, so affecting eventual yield.
Some evidence for the existence of this differential
utilization was recently obtained from a nitrogen
carrier trial, on a Cartref sand (pH 5.7), which was
designed to test the efficiencies of four ammonium
fertilizers among themselves and against a control with
no nitrogen. The yields obtained from the plant cane
crop harvested at 19 months are given in Table IV
The Effect of Liming (CaCO3) on Nitrification
(ppm. N* present after four weeks incubation at 35" C. and 60% W.H.C.)
* Mean of duplicate determinations.
TABLE III
Relative Leaching Losses of NH4-N and NO3-N from Two Sandy Soils
(2 in. water applied to pots one week after N fertilization)
* Mean of duplicate pots.

Proceedings of The South African Sugar Technologists' Association — March 1966
TABLE IV
Efficiency of Ammonium Fertilizers on a Cartref Sand
(120 lb. N/acre)
*T.C.A. - tons cane per acre. †T.S.A. tons sucrose per acre
From these results it is apparent that while all the
N carriers significantly outyielded control, both for
tons cane per acre (T.C.A.), and tons sucrose per acre
(T.S.A.), the ammonium fertilizer treatments differed
significantly amongst themselves. In an attempt to try
and explain these differences a laboratory experiment
was designed to follow the nitrification pattern occurring
in sandy soils of variable pH, after addition to
them of the ammonium nitrogen carriers used in the
above field trial.
Procedure. 800 g. air dry samples of three series of
sands widespread in the Sugar Belt, namely Lytton
(pH 4.60), Cartref (pH 4.80) and Clansthal (pH 6.15),
were subject to the following treatments:
(1) Control — no fertilizer.
(2) 80 mg. N as (NH4)2SO4 — 100 ppm.
(3) 80 mg. N as urea — 100 ppm.
(4) 80 mg. N as NH4NO3— 100 ppm.
(5)80 mg. N as L.A.N. (limestone NH4NO3)—
100 ppm.
(6)80 mg. N as NH4NO3—100 ppm. + 4 tons
CaCO3 per acre.
The soils were moistened to 30 per cent water
holding capacity with solutions of the above fertilizers
and incubated at 30" C. After periods of 0, 7, 14, 21
and 28 days duplicate samples of all treatments were
withdrawn and pH, NH4-N and NO3-N determinations
carried out.
Results. Changes in pH, and amounts of NH4-N
and NO3-N over four weeks following the various
fertilizer additions are presented in Figure 1. The
main effects of the different treatments are now briefly
detailed.
Control. While nitrification was partially delayed
during the first two weeks in the Lytton sand little
delay was noted in the other two control soils.
(NH4)2SO4. Apparently the acidifying action of this
fertilizer coupled with the highly acid Lytton sand
effectively prevented nitrification from occurring even
after four weeks, while on the Cartref sand slow conversion
to NO3-N commenced after the first week.
The slightly acid Clansthal sand started nitrifying
immediately, only 10 ppm. NH4-N remaining after
28 days.
Urea. The initial rise in pH due to the hydrolysis of
the urea was obviously responsible for providing a
more favourable environment for nitrification in both
the acid sands. As a result over 30 ppm. and 60 ppm.
NH4-N respectively had been nitrified in the Lytton
and Cartref soils after four weeks. Nitrification was
rapid in the Clansthal sand being complete within the
first two weeks of incubation.
NH4NO3. Addition of this fertilizer also had an
acidifying effect and in consequence patterns of NH4-N
changeover in the three soils, were closely similar to
those obtained when (NH4)2SO4 was applied.
L.A.N. The presence of limestone with the NH4N03,
raised the pH sufficiently on both the strongly acid
sands to bring about an increase in their rate of nitrification,
compared to that when NH4N03 only was
used. Despite this, conversion to N03-N in the
Lytton sand was still very slow.
NH4NO3 + 4 tons CaCO3 per acre. The effect of
this treatment on the pH of all soils is clearly seen
and nitrification was greatly accelerated being complete
within 21, 14 and 7 days for the Lytton, Cartref
and Clansthal sands respectively.
The pH data reveals that after an initial reduction
in acidity following the application of certain of the
treatments, acidity gradually increased as nitrification
proceeded except where the equivalent of 4 tons of
limestone was added. This is due to the release of
hydrogen during the process so that these weakly
buffered sands eventually tend to be more acid than
they were before ammonium fertilizer was applied.
Discussion and Conclusions
Results confirming those cited above have been
obtained by workers elsewhere. Munk 6 found that
nitrification was considerably better in an acid soil
(pH 4.40), with L.A.N. or urea than with (NH4)2SO4,
and Dijkshoorn 5 showed in pot experiments with
(NH4)2SO4, that nitrification was complete within 21
days at pH 7.0, considerably retarded at pH 5.3, and
inhibited at pH 4.3. Ayres and Humbert 1 studied
nitrification patterns of two ammonium fertilizers in
Hawaiian soils in relation to pH, while Botha 3 showed
that urea nitrified more readily than (NH4)2SO4 when
applied to three soils of widely differing pH and texture.
Broadbent et al. 4 suggested that inhibition of nitrification
might result from the presence of high concentrations
of free ammonia in the soil at low pH exerting
selective inhibition on nitrate forming bacteria. This
may partly explain the complete absence of nitrification
in the highly acid Lytton sand when (NH4)2SO4
and NH4NO3 were added. At low NH4-N concentrations
(see control treatment) a population of nitrifiers
is gradually able to develop.
243

244 Proceedings of The South African Sugar Technologists' Association — March 1966

Proceedings of The South African Sugar Technologists' Association — March 1966
It would appear that in soils of pH 5.2 or below
considerable delays in nitrification are likely to occur.
Apart from the pH effect of the soil as such, the particular
ammonium nitrogen carrier applied can apparently
influence to a greater or lesser extent the rate
of nitrification in the soil. This may well be beneficial
from the aspect of N utilization by the cane plant,
particularly in the case of acid sandy soils, and provides
an explanation for the yield differences between
treatments on the nitrogen carrier trial previously
mentioned.
The NH4-N fraction from both the (NH4)aSO4
and urea was probably available to the developing
cane for a longer period than that from NH4NO3 or
L.A.N., as double the amount was present initially.
The NO3-N fraction of the two latter fertilizers
would have been prone to immediate leaching following
rain and could easily have been lost before much
of it was utilized, thus leaving the plant with only half
the original quantity of N applied. Enhanced nitrification
in the case of L.A.N. due to the presence of
limestone would increase the possibility of even
greater leaching losses from this fertilizer than from
NH4NO3 alone.
Therefore laboratory and field data so far obtained
indicates that the application of L.A.N, or NH4NO3
to acid sandy soils may not always be completely
effective, and that under these conditions the use of
urea, or (NH4)2SO4 on slightly acid sands, would
seem to be preferable. If the plant is able to fully
utilize NO3-N immediately it is applied, and subsequent
leaching losses are slight, then yield reductions
from these N carriers would probably be insignificant.
Also the possibility that part of the leached NO3-N
may be taken up at a later stage of development by
cane roots penetrating to depth on the sands should
not be discounted. Seasonal effects and stage of plant
development therefore play an extremely important
part in deciding whether or not maximum use will be
made of the fertilizer N applied.
Summary
During N mineralisation studies on several Sugar
Belt soils, delays in nitrification were noted where pH
was 5.2 or below, being particularly marked in the
Lytton and Cartref series of sands.
Limed at the rate of 2, 4 and 8 tons per acre, and
incubated for four weeks, soils previously showing
delayed nitrification, nitrified rapidly, the decrease in
acidity due to liming providing an environment favourable
to the development of nitrifying bacteria.
Patterns of nitrification were examined during a
four-week period of incubation, after several ammonium
nitrogen carriers had been added to three
sandy soils of variable pH. Apart from the soil pH
effect as such, it was shown that a specific carrier can
markedly influence the rate of nitrification in any
particular soil, which in turn may accelerate or retard
susceptibility to leaching and subsequently affect plant
utilization of applied N.
These results are discussed with regard to the manner
in which they affect choice of N fertilizers for cane
in acid sandy soils.
References
1. Ayres, A. S., and Humbert, R. P. (1956). Nitrification of
aqua ammonia and ammonium sulphate in Hawaiian soils
in relation to pH. Rep.Assoc.Hawaiian Sugar Tech.: 22-25.
2. Bishop, R. T. (1965). Mineral nutrient studies in sugar cane
Proc.Annual Conf.S.Afr.Sugar Tech.Ass. 39: 128-133.
3. Botha, A. D. P. (1960). Nitrification of urea and ammonium
sulphate in three types of soil. S.Afr.J.Agric.Sci. 3:651-652.
4. Broadbent, F. E., Tyler, K. B., Hill, G. N., et al. (1957).
Nitrification of ammoniacal fertilizers in some California
soils. Hilgardia 27: 247-267.
5. Dijkshoom, W. (1960). Effect of the rate of nitrification of
ammonium fertilizer on nitrate accumulation and cationanion
balance in perennial ryegrass. Inst.biol.scheik.Onderz.
Landb-Gewassen Meded.Jaarb. 102-128: 123-134.
6. Munk, H. (1958). The nitrification of ammonium salts in
acid soils. Landw. Forsch. 11: 150-156.
7. Robinson, J. B. (1963). Nitrification in a New Zealand
grassland soil. Plant and Soil 19: 173-183.
8. Wood, R. A. (1964). Assessing the potential of Sugar Belt
soils to supply nitrogen for plant cane. Proc. Annual Cong.
S.Afr.Sugar Techn.Ass. 38: 176-179.
9. Wood, R. A. (1966). The influence of trash on nitrogen
mineralisation-immobilization relationships in Sugar Belt
soils. Proc. Annual Cong. S. Afr. Sugar Tech. Ass. 40: (in
press).
Mr. Wardle: Does not Mr. Wood think there is a
case for split applications of nitrogen on sand?
Mr. R. A. Wood: The data we have is not such as to
give a clear answer to this. All we know definitely
is that cane requires most of its nitrogen during the
early stages of growth.
Professor Orchard: One should be careful not to
generalise from the results of a single experiment.
The pattern of response to nitrates in the Cartref
soil shown in Table IV need not necessarily be the
same for other soils.
Mr. R. A. Wood: I agree, and the results of the
Cartref trial have not been recommended for general
guidance. I was careful to point out that application
of nitrates may not always be completely effective in
acid sandy soils.
Mr. du Toit: I agree with Professor Orchard that
we must not generalise on the results from one
experiment but we should certainly not ignore them.
Mr. Wyatt: In the Cartref experiment the interesting
point is that the application of nitrogen was split—
60 lb. in March 1963 and another 60 lb. the following
spring.
Mr. Odendaal: I note that the Cartref sand used
in the nitrogen carrier trial had a pH of 5.7, while
that used in the Laboratory experiment was only
4.8. Therefore nitrification patterns occurring in the
245

246
field would be different to those found in the laboratory.
Mr. R. A. Wood: This is true, in fact the chances
of more rapid nitrification taking place in the field
trial and possibly enhancing leaching losses would
seem likely.
Mr. Sherrard: Is there any danger of loss of nitrogen
if urea is applied on top of the soil in dry weather
as at present ?
Mr. R. A. Wood: If it became wet, from dew for
instance, there could be considerable loss due to
volatilisation of ammonia, in warm sunny conditions.
Proceedings of The South African Sugar Technologists' Association — March 1966
Mr. Gilfillan: Surely in the case Mr. Wyatt mentioned
nitrification is not so important as it was a split
application.
Mr. Wyatt: The experiment is being repeated; not
only on the same soil but also in others.
Mr. King: Does cane take up nitrogen in the
ammonium form?
Mr. R. A. Wood: Yes we have been able to demonstrate
this by successfully growing cane in soils to
which an organic compound called N-Serve has been
added, that is capable of inhibiting nitrification for
several weeks.

Proceedings of The South African Sugar Technologists' Association—March 1966 247
ABNORMAL NITROGEN REQUIREMENTS OF SUGAR
CANE ON THE MONTMORILLONITIC BLACK CLAY
SOIL OF THE KAFUE FLATS IN ZAMBIA
Introduction
Large scale cane growing trials on the Kafue Flats,
Zambia, were initiated by the Tate & Lyle Group in
September, 1963, in collaboration with the Rhodesian
Selection Trust Group on the 700 acre Kafue Pilot
Polder.
Soil Characteristics
The soils of the Kafue Flood Plain, which covers
an area of 1.3 million acres, are some of the heaviest
tropical black clays in the world, averaging 70 per
cent clay, of which more than half consists of Mont¬
morillonite. This high percentage of Montmorillonitic
clay causes the soil to have very marked swelling,
shrinkage and cracking properties. On wetting, the
soil swells to a compact, structureless, impervious
mass, and on drying out shrinkage causes extensive
cracking. In an established crop this cracking may
ramify through the top three feet of soil. The somewhat
specialised methods of crop management evolved
on these lands are governed by these physical properties.
Infiltration of irrigation water depends almost
entirely on cracking. The cracks allow lateral movement
of water for considerable distances and up to
sixty feet has been recorded. Below about four feet,
the subsoil is permanently wet and virtually impermeable,
so that losses through deep percolation are
negligible. The severe drainage problem brought about
by the characteristics of this soil when wet are overcome
by growing crops either on cambered beds or
on high ridges. Other relevant properties of the soil
are that it is not readily leached and the pH is neutral.
Free lime and gypsum are often found.
Fertiliser Requirement
Of the major nutrients, the main response is obtained
from Nitrogen and exceptionally heavy applications
are needed. Phosphates also give a positive yield improvement.
No response is obtained from Potash.
It would appear that mineralisation of organic
Nitrogen is severely restricted, and this prevents crops
from obtaining Nitrogen from any organic matter in
the soil, so that Nitrogen has to be applied as straight
chemical fertiliser.
Cane Trials
By 1963 it had already been demonstrated in a
small trial plot on the Polder that very high yields of
cane could be obtained on these soils, the major
agronomic problems to be contended with being:
(a) The design of afield layout combining a system
of cheap surface irrigation with the necessity
for good drainage, especially during the summer
rains.
By D. S. HUGHAN AND D. R. C. BOOTH
(b) The selection of varieties best suited to the
conditions.
(c) The control of weeds during the summer rains
when it is usually difficult, if not impossible,
to keep the lands clean by conventional hand
weeding methods.
(d) The overcoming of the excessively high Nitrogen
requirements.
System of Cane Irrigation
Two basic systems of irrigation are being used:
(i) Furrow Irrigation — over graded lands at 1:600,
with line lengths of 700 feet.
(ii) Flood irrigation of level basins — basin areas
varying from 6.5 to 13.5 acres, with line lengths
ranging from 400 feet to 950 feet.
Under both systems cane is grown on high and wellformed
ridges resembling the Louisiana Bank System.
The basin irrigation method is the easiest to manage
and requires only 2 labourers to handle 10 cusecs of
water.
Nitrogen Requirements
Indications from Foliar Diagnosis and Fertiliser Trials
Kerkhoven (1963) found that the nitrogen requirements
for cotton, maize and rice on these soils were
very high — in the region of 200-240 lb. of actual N.
per acre, and for sugarcane 400-500 lb. N. For this
reason all fertiliser trials laid down in 1963 and 1964
included treatments with Nitrogen levels up to 600
lb. N per acre.
The first trial laid down in October, 1963, was
harvested in October, 1964.
Design: Randomised block, 10 treatments, 5 replicates.
Plot size: 1/66 acre.
Planted: 7th October, 1963.
Variety: Co.419.
Harvested: 26th October, 1964.
Age at
Harvest: 12 1/2 months.
Treatments
4 Levels of Nitrogen
0 lb. N per acre = 0 lb. Sulphate of Ammonia
100 lb. N per acre = 476 lb. Sulphate of Ammonia
300 lb. N per acre = 1,428 lb. Sulphate of Ammonia
600 lb. N per acre = 2,857 lb. Sulphate of Ammonia
3 Levels of Phosphate
0 lb. P2O5 per acre = 0 lb. Double Superphosphate.

248
80 lb. P2O5 per acre = 210 lb. Double Superphosphate.
160 lb. P2O5 per acre = 421 lb. Double Superphosphate.
No Potash was applied as previous experiments had
shown that no response had been obtained from
Potash applications.
All the Phosphate was applied at planting.
The Nitrogen treatment of 100 lb. N per acre was
applied at planting. The other Nitrogen treatments
received two equal dressings, half at planting and the
remainder at 3 months of age.
Leaf Analysis
Leaf samples were taken on the 7th April when the
cane was at 6 1/4 months from planting. Analysis of the
leaf lamina of the middle third of 20 leaves was carried
out and the content of the major nutrients was expressed
as a percentage of the dry weight. The analytical
results followed remarkably closely to the
Nitrogen application.
For comparative purposes the minimal optimum
levels of nutrients in cane leaves at 5 months of age
are shown for Jamaica. These levels are considered
by R. F. Innes to be as follows:
N —1.80%
P2O5 —0.42%
K2O —1.40%
If nutrient levels are below these standards it is
considered that insufficient nutrient has been available
to the plant and that an economic increase in yield
would have been obtained from additional fertiliser.
The standards which would apply to Zambia have
unfortunately not been determined.
The results of the trial, together with leaf analysis
figures are laid out in Table I.
Proceedings of The South African Sugar Technologists' Association—March 1966
Table I
(a) Leaf Analyses
L.S.D. = 3.34 T.C.A. (at P = .05)
Comments
(i) Nitrogen content of leaves at the N-0 and
N-100 levels showed a marked deficiency.
N-300 treatments were low and only at the
N-600 level was the optimum attained.
(ii) Phosphate contents were good at all levels
of fertilisation, though with indication of
some suppression at the higher rates of N.
(iii) Potash contents were adequate throughout.
(b) Yields
Yield data, in tons cane per acre, are summarised
in Table II:
Table II
Yield in Tons Cane per acre
These results show highly significant increases in
yield from N-0 up to N-600. Response to P2O5 shows
a significant increase in yield of 5.0 T.C.A. from P-0
to P-80, followed by a very slight depression from
P-80 to P-160.
(c) Juice Quality
Effects of treatments on juice quality, expressed in
terms of tons cane/tons sugar ratio, are summarised
in Table III.

Proceedings of The South African Sugar Technologists' Association—March 1966 249
Table III
Tons Cane/Tons Sugar Ratio
These results show little variation in. TC/TS
ratios at the lower levels of Nitrogen from N-0 to
N-300, followed by a sharp deterioration of 0.70
TC/TS from N-300 to N-600. Increasing Phosphate
has shown a trend of a steadily improving TC/TS
ratio from P-0 to P-160.
Conclusions
Owing to the serious effect on juice quality at 600
lb. Nitrogen, it appeared that there would be a case
for reducing the level in plant cane to between 300
and 450 lb. The effect of Phosphate on juice quality
was interesting, and indicated that at any level of
Nitrogen from 300 to 600 lb. it would be worthwhile
to apply at least 80 lb. of Phosphate per acre.
Additional Experimental Results
1. Ratoon Fertiliser Trial (Co.331), 1964 crop
Results of a ratoon cane fertiliser trial harvested in
1964, although not statistically significant, showed up
two further important trends. Treatments in this
experiment included Nitrogen at 300, 450 and 600 lb.
with single and split application at each level, all plots
receiving a standard dressing of 120 lb. P2O5 and 60
lb. K2O.
At the higher levels of Nitrogen there was a notable
increase in yield coupled with a sharp deterioration
in juice quality when split application of fertiliser was
made. The first application was made on the 26th
October, 1963, and the second on the 15th February,
1964, which is normally considered very late. It is felt
that there will be a real benefit in applying Nitrogen
in two stages, provided that the second and final
application is made within three months of planting
or harvesting, and before January.
Table IV
Yield in Tons Cane per Acre
2. Ratoon Fertiliser Trial (Co.419), 1965 crop
The 10 by 5 randomised block N.P. trial referred
to previously (see Tables I, II and III) was continued
in the 1st ratoon crop, with identical treatments. The
results of this trial when harvested in 1965, at 10
months of age, are summarised in Table IV and
Table V.
These results show very marked increases in yield
with increasing Nitrogen, with some flattening off of
the response curve from N-300 to N-600. Yields
increase linearly with increasing Phosphate from 0 lb.
to 160 lb. of P2O5 per acre.
The results of the ratoon experiment, when compared
with the plant yield of the same trial, show a
general reduction in yields. This is probably attributable
to:
(i) Inferior drainage due to flattening of ridges,
which were not reformed (as is now a routine
practice) in the ratoon crop.
(ii) The 1st ratoon crop was reaped at 10 months
only — 2 1/2 months younger than the plant crop.
Table V
Tons Cane/Tons Sugar Ratio
Increasing Nitrogen levels have caused a progressive
deterioration in juice quality from N-0 up to N-300.
The mean TC/TS ratio at N-600 is the same as that
at N-300.
Phosphate has in this case had little effect on juice
quality.
3. Plant Cane Fertiliser Trials (Co.419), 1965 crop
Results of an 8 by 5 randomised block N.P. trial
laid down in Co.419 plant cane in 1964, when harvested
at 12 1/2 months of age in 1965, are summarised
in Table VI and Table VII.
Table VI
Yield in Tons Cane per Acre

250
Cane yields show a progressive increase from N-100
to N-600, the most notable response being an increase
of 17.94 T.C.A. from N-100 to N-300. Increasing
Phosphate from P2O5-80 to P2O5-120 has not brought
about any significant increase in yield.
Table VII
Tons Cane/Tons Sugar Ratio
These figures show the usual trend of deteriorating
juice quality with increasing levels of Nitrogen, with
the exception of an apparent slight improvement from
N-300 to N-450. Increasing Phosphate has again had
little effect on juice quality.
Conclusions
The results of the trials harvested in 1965 confirm
that the optimum Nitrogen level under the present
system of cultivation should be in the order of 300 lb.
actual N per acre. At higher rates the yield response
curve generally begins to flatten out and TC/TS
ratios deteriorate. Increasing Phosphate above 80 lb.
P2O3 per acre has not brought about any significant
increases in yield or juice quality.
Drainage Depth, Interaction upon Nitrogen
Requirement
During the course of the 1964,65 growing season
an important physical factor influencing Nitrogen uptake
came to light. This was a very critical "drainage
depth" (this being defined as the depth of temporary
water tables caused by flood irrigation and during
heavy storms). Jelley (1964) found in other crops that
not only were yields affected by a relationship between
growth and depth of temporary water tables, but that
there also exists a strong interaction between Nitrogen
and drainage. For efficient use of Nitrogen at high
yield levels, a drainage depth of 8 in. to 12 in. was
indicated. This effect was clearly demonstrated in cane.
Even before the onset of rains, cane on ridges where
irrigation water had risen close to the crest, or cane
which had been completely submerged over areas of
local depression showed stunted growth, poor tillering
and severe Nitrogen deficiency symptoms in spite of
receiving 525 lb. N. In general, the higher the ridges
are above the temporary water level, the better the
growth has been. For example, a 7.25 acre basin which
yielded an overall average of 72.82 tons cane per acre
when harvested as plants in 1965, at 13 months of
age, gave 91.99 tons cane per acre over a measured
better drained local area of 2.40 acres.
Proceedings of The South African Sugar Technologists' Association—March 1966
(a) 1963:64 Crop
Field Scale Fertiliser Applications,
Costs and Cane Yields
The basic fertiliser dressing applied to the first land
planted to large scale cane trials in September, 1963,
was as follows:
Nitrogen — 100 lb. N per acre (as 476 lb. Sulphate
of Ammonia).
Phosphate — 160 lb. P2O5 per acre (as 421 lb. Double
Superphosphate).
Potash — Nil.
By January, 1964, severe Nitrogen deficiency symptoms
became apparent. The cane was a very pale
green and tillering was poor. Leaf samples taken in
this Co.419 plant cane at 3 1/2 months of age gave the
following dry weight percentage analysis:
In view of this, and Kerkhoven's findings referred
to in Section 1, it was decided to apply a large top
dressing of Nitrogen to overcome the severe deficiency
symptoms and to bring the level into line with his
recommendations. The crop was top dressed in February
with 350 lb. N per acre (using Sulphate of
Ammonia at 1,667 lb. per acre); this brought the total
nutrient application up to 450-160-0. Response to the
additional Nitrogen was rapid and. a spectacular
greening up of the yellow cane canopy was observed.
This late planted field went on to yield an average
of 61 tons cane per acre when harvested at 12 months
of age.
(b) 1964/65 Crop
In July, 1964, when further cane planting on the
Polder was commenced, no cane had yet been harvested
on these soils and. it was necessary to arrive at
a reasonable level of Nitrogen application based on
the limited information available. On the strength of
leaf analysis results from the fertiliser trial shown in
Table 1 it was decided that the standard field dressing
should provisionally be:
(i) For plants:
Nitrogen — 525 lb. N per acre (as 2,500 lb. Sulphate
of Ammonia, half at planting,
half at 3-4 months).
Phosphate— 95 lb. P2O5 per acre (as 250 lb.
Double Superphosphate, at planting).
Potash — Nil.
Cost: £31 10s. per acre.
Plant cane receiving the above treatment went on
to yield 73 tons cane per acre at 13 months of age.

Proceedings of The South African Sugar Technologists' Association—March 1966
(ii) For ratoons
Nitrogen —525 lb. N per acre (as 2,500 lb.
Sulphate of Ammonia, half after
cutting, half at 3 months).
Phosphate — 76 lb. P.,05 per acre (as 200 lb.
Double Superphosphate, after
cutting).
Potash — Nil.
Cost: £30 15s. per acre.
Ratoon cane receiving the above treatment yielded
62 tons cane per acre at only 10 months of age.
The standard dressing already established for upland
red soils was 111-76-0 (200 lb. Sulphate of
Ammonia + 200 lb. Double Superphosphate at planting
or after harvest, and 150 lb. Urea at 3 months)
costing £6 15s. per acre, and the high cost of Nitrogen
fertilisation on the Polder soils was a matter of considerable
concern.
(c) 1965/66 Crop
Results of the fertiliser trials harvested in late 1964
and during the 1965 cropping season, indicated that a
reduction in the Nitrogen application could be made
to 300 lb. N per acre and a reduction in phosphate
application to about 80 lb. P^Or, per acre. The standard
field dressing for the 1966 crop (now as 1st and 2nd
ratoons) was therefore fixed at:
Nitrogen —315 lb. N per acre (as 1,500 lb.
Sulphate of Ammonia, half after
cutting, half at 3 months).
Phosphate — 76 lb. P205 per acre (as 200 lb.
Double Superphosphate, after
cutting).
Potash — Nil.
Cost: £19 14s. per acre.
The cost of fertilising at the above rates represents
a reduction of £11 per acre below the 1964 ratoon
cane cost, but is still about triple that incurred on
nearby upland red soils — however, measured yields
from well drained sections of fields, where drainage
depths are adequate, have shown a yield potential of
over 90 tons cane per acre in 12 months. If, by improved
land preparation and management, such yields
can be achieved on a commercial scale, then a fertiliser
requirement of 300-80-0 would be economically
justified, especially taking into account the very low
cost of water application under the basin system of
irrigation — within an empoldered estate — close to
the perennial Kafue River.
Summary
The Montmorillonitic soils of the Kafue Flats are
described, followed by notes on the agronomic aspects
of cane growing on these lands, with special reference
to the excessively high Nitrogen requirements indicated
by early trials carried out at the Kafue Pilot
Polder.
The system of cane cultivation, and irrigation as
practised on the Kafue Pilot Polder is briefly described.
Foliar analyses made during the 1963/64 growing
season indicated a Nitrogen requirement of between
300 and 600 lb. actual N per acre. Fertiliser trials
harvested in 1964 and 1965 showed that the optimum
Nitrogen level could be reduced to 300 lb. N per acre.
The interaction between Nitrogen and drainage
depth is discussed.
The standard dressing already established for upland
red soils was 111-76-0, costing £6 15s. per acre.
Following on the results of fertiliser trials harvested
in 1964 and 1965 the standard, field application on the
Black Clay soils has been 315-76-0, using 1,500 lb.
Sulphate of Ammonia (half after cutting, half at 3
months) and 200 lb. Double Superphosphate per acre,
costing £19 14s. per acre — still almost triple that
incurred on the upland red soils — however, given
good drainage, yields of over 90 tons cane per acre
in 12 months can be achieved on these lands under
a very cheap system of irrigation.
Acknowledgments
We wish to record our appreciation of the assistance
given by the Management and staff of the Kafue Pilot
Polder and the help subsequently given by the Government
of Zambia, who took over the Polder as the
Kafue Irrigation Research Station in July, 1965.
The authors thank the Directors of the Ndola Sugar
Company Limited for permission to publish this paper
and acknowledge the help given by M. E. Johnson,
the Company's Research Assistant, who carried out
most of the field work on these trials during 1964
and 1965.
References
Hughan, D. S. and Booth, D. R. C. Kafue Pilot Polder Cane
Growing Trials. Annual Report for 1963/64 (unpublished).
Jelley, R. M. The Physical Properties and Field Management
of Soils of the Kafue Pilot Polder and Kafue Flats, 1964
(private publication).
Kerkhoven, G. J. Crop Potential of Margalitic Soil in the Kafue
Flats. Rhod. J. Agric. Res. Vol. 1, No. 2, July 1963, pp. 71-79.
Kerkhoven, G. J. and Jelley, R. M. Kafue Pilot Polder. World
Crops, June, 1964 (reprint).
Mr. R. A. Wood: It would appear that nitrogen
losses are the main reason for the abnormal requirements
on this soil. Large quantities of sulphate of
ammonia are being applied to a neutral soil in which
nitrification will be rapid. Once this occurs on a
poorly drained soil in which the water table is close
to the surface, nitrogen losses, resulting from denitrification
under these anaerobic conditions, will be very
severe. The main problem therefore is one of trying
to reduce these losses, which could amount to more
than 50 per cent of the nitrogen applied, by improved
drainage and judicious timing of fertilizer applications.
Mr. Booth: It is felt that one reason for high fertilizer
nitrogen requirements is possibly lack of mineralisation
of organics.
251

252 Proceedings of The South African Sugar Technologists' Association—March 1966
Mr. Wood: Anaerobic conditions might cause
denitrification.
Mr. du Toit: How and when is nitrogen applied?
Mr. Booth: Nitrogen is applied as sulphate of
ammonia, one half at planting (placed in the furrow)
or immediately after harvesting, and one half as a top
dressing at three months.
Mr. du Toit: I have never before heard of such
fantastic nitrogen applications and at least 80 or 90
per cent must be lost so I think that the time and
method of application should be reviewed.
Mr. Booth: Trials incorporating different nitrogen
carriers and using single and split applications of
fertiliser are being carried out. So far these have not
given any significant results, though in the case of
split applications there is a trend towards higher yields
coupled with deteriorating juice quality unless the
second application is made within three months of
planting or harvesting.
Mr. Gunn: How was sucrose calculated?
Mr. Booth: Sucrose was calculated from Brix and
Pol readings taken on 20 cane samples crushed in the
laboratory mill. Tons cane/tons sugar ratios were
calculated from sucrose % juice using assumed Java
Ratio, mill extraction and "Boiling House Recovery"
figures.
Mr. Gunn: Did you worry about juice purity?
Mr. Booth: Yes, very much so.

Proceedings of The South African Sugar Technologists' Association — March 1966 253
THE INFLUENCE OF TRASH ON NITROGEN MINERAL­
ISATION-IMMOBILIZATION RELATIONSHIPS IN SUGAR
BELT SOILS*
1. LABORATORY INVESTIGATION OF NITROGEN
TURNOVER
Introduction
Biological immobilization and mineralisation of
nitrogen in the soil can be regarded as opposing processes.
The former is the transformation by microbial
action of inorganic or mineral N into an organic form
and thereby rendered unavailable to the plant, while
the latter is the conversion of the immobilized N to
the plant available inorganic forms by microbial decomposition.
Winsor 14 points out that from the biological
nature of these two processes it is obvious that
both must occur simultaneously, since the formation
of organic nitrogenous compounds by micro-organisms
is an essential feature of their growth, these compounds
in turn being decomposed on the death of the
organisms concerned. The cycling of products in the
above manner is referred to as nitrogen turnover and
takes place continuously in all soils.
Many workers have shown that the effect of adding
carbonaceous materials such as straw 2 , 5 , 6 , 7 , 8 , 9 , sawdust
1 , 6 and other plant residues 10 , 11 , 12 to the soil, is
to lower the inorganic N content, the C/N ratio of
the added material largely governing the degree by
which it is reduced. Waksman 13 in 1924, showed that
organic materials containing N in excess of 2.0—2.5
per cent, decomposed with immediate release of
NH4-N, whereas residues of lower N content generally
caused assimilation of any inorganic N present by
soil organisms. Jansson 8 and others 2 , 4 have shown
By R. A. WOOD
TABLE I
Soil Analytical Data — Nitrogen Turnover Experiment
C.S. - Coarse sand. F.S. - Fine sand. S - Silt. C - Clay.
that nitrogen tied-up in this way is eventually released
as mineral N at a characteristic turning point, i.e.
when nett immobilization changes into nett mineralisation.
In view of these findings it was decided to investigate
the rate of nitrogen turnover in Sugar Belt soils, many
of which are heavily trashed, in order to obtain some
assessment of the importance of this process. Incorporation
of trash into the topsoil, generally of high
C/N ratio, which inevitably occurs under field conditions,
may lock up considerable quantities of available
N, often at a time when plant requirement is
greatest. Under these conditions additional fertilizer
N would probably be necessary to counteract this
effect and assist in the turnover process. An experiment
was therefore undertaken to establish immobilization
rates and subsequent release of nitrogen,
following the addition of trash and fertilizer N to a
representative group of Sugar Belt Soils.
Procedure
800 g. air dry samples of seven soils series, analytical
data for which is presented in Table I, were
subject to the following treatments:
1. Control — no fertilizer or trash.
2. 80 mg. N as (NH4)2S04 — 100 ppm. (200 lb.
N/acre).
3. 1 per cent trash — 10 tons/acre equivalent.
4. 80 mg. N as (NH4)2S04 4- 1 per cent trash.
* This work will form part of a thesis to be submitted to the University of Natal, Soil Science Department, for the Ph.D. degree.

Proceedings of The South African Sugar Technologists' Association — March 1966 255
Figure I. Immobilization of nitrogen
and rate of carbon mineralisation in
Sugar Belt soils receiving trash and
ammonium nitrogen (100 p.p.m. N).

256 Proceedings of The South African Sugar Technologists Association — March 1966
The trash was carefully mixed with the soils beforehand
after which they were moistened to 30 per cent
of their water holding capacity, either with a solution
of (NH4)2S04 or water, an additional 3 ml. of water
being supplied for each gram of trash, and incubated
in jars at 30° C. After periods of 0, 5, 8, 13, 20, 27,
41, 55 and 69 days, the jars were shaken, and duplicate
samples 25 g. air dry soil were withdrawn from all
treatments and analysed for NH4 - N and NO3 - N.
The soils were aerated daily throughout.
In addition, after the initial moistening, duplicate
samples from each jar 50 g. air dry soil, were taken
and put into macrorespirometers 17 at 30° C, in order
that C mineralisation could be followed in all treatments
throughout the experiment. Daily output of
CO2 was measured and related to that absorbed by
2 N NaOH after 35 and 69 days.
The trash used was a finely ground composite
sample prepared from the dead basal leaves and
sheaths of a large number of cane plants. It contained
42.6 per cent C and 0.28 per cent N, giving a C/N
ratio of 152.
Results and Discussion
Treatment effects on the type and amount of inorganic
N present in the various soils throughout
the incubation period are given in Table II. Figure 1
presents graphically the degree of N immobilization
occurring following the addition of trash plus
(NH4)2S04 to the different soils before the turning
point is reached, and the release of mineral N into the
system that follows. Data for C mineralisation rates
during the 69-day period is also superimposed. The
main effects of the various treatments are now discussed.
Control. All treatments showed a nett mineralisation
of N occurring throughout the incubation period and
varying only with the potential of any particular soil
to mineralise N. In most soils any NH4 - N was
rapidly nitrified shortly after incubation commenced.
Addition of (NH4)2SO4. A variable amount of
ammonium fixation was apparent in all soils. Nitrification
of the added NH4 - N occurred at differing
rates being incomplete in some cases even after 69
days. With an adequate supply of nitrogen, mineralisation
was somewhat reduced in all soils.
Addition of trash. Rapid immobilization of the
existing supplies of inorganic N took place in all soils
from the commencement of incubation, and no nett
mineralisation of N occurred in any soil during the
69 days. This condition still existed after 150 days in
all but two of the soils, namely those of the Clansthal
and Waldene series. In the former at 97 days, 59 ppm.
NO3-N was detected and this had risen to 97 ppm.
at 150 days, while in the latter soil at this time only
29 ppm. NO3-N was present.
Addition of (NH4)2SO4 plus trash. As with the previous
treatment marked N immobilization occurred
in all soils due to the presence of trash, but the
addition of NH4-N provided a supply of N in excess
of that required by the soil micro-organisms, so that
after the initial flush of decomposition, considerable
amounts still remained in the soil. Preferential absorption
of NH4-N rather than NO3-N by the soil
micro-organisms during immobilization, was observed
in all soils. This has been noted by other workers 5 , 9 ,
but it has also been shown that NO2N is also immobilized
to a considerable extent when this is the
only form available to soil micro-organisms 5 . The
data reveals that when immobilization had reached a
maximum in each of the soils the mineral N content
gradually started to increase, thus indicating that the
turning point had been reached and that nett mineralisation
was occurring. The time taken to reach the
turning point in the various soils showed very marked
differences ranging from between 8 to 41 days.
C Mineralisation during Incubation. In the absence
of trash, all soils with or without added N only mineralised
carbon slowly, and the difference between these
treatments was not significant. As anticipated the
addition of trash greatly increased C mineralisation
in all soils, but initially microbial action was far greater
where fertilizer N was also present, so that at the end
of the incubation period, with the exception of the
Clansthal sand, a larger percentage of trash carbon
had been mineralised, than where N was absent. (See
Table III). Analyses for inorganic N after 69 days, of
the small samples continuously aerated in the macrorespirometers,
showed close agreement with those
obtained from the bulk samples incubated over the
same period, thus indicating that the C mineralisation
data had provided an accurate measure of biological
activity within these larger samples.
TABLE HI
Per Cent Trash Carbon Mineralised After
69-Day Incubation Period
The degree of trash decomposition at the turning
point. The information presented in Table IV shows
that the amount of trash carbon mineralised at the
time of maximum N immobilization varies considerably
within the range of Sugar Belt soils examined.
It would appear that generally the more fertile soils
of somewhat heavier texture are able to decompose
larger amounts of trash, and are thus characterised at
the turning point by low C/N ratios in the remaining

Proceedings of The South African Sugar Technologists' Association — March 1966 257
organic material. The sandier soils of lower fertility
decompose far less trash before the turning point is
reached, leaving a residue with a higher C/N ratio. In
this latter group of soils in the presence of N fertilizer,
the cycling of N immobilized during nitrogen
turnover is apparently quite rapid, and under these
conditions cane should be able to utilize the mineral
N released after comparatively short periods.
Soil acidity has been shown to influence the degree
of decomposition 8 , 15 , as it affects the scale of microbiological
activity in a particular soil. This probably
accounts for the low turnover occurring in the mistbelt
soil (Inanda series), which is known to contain
large amounts of undecomposed material inert to
microbial attack which accumulates under moist, acid
field conditions 16 .
Trash decomposition in the Recent sand (Clansthal
series). The comparatively rapid cycling in this soil of
N immobilized when trash was added in the absence
of fertilizer N was noted earlier. On closer examination
it was found that this was associated with the
greatest percentage trash decomposition (46.7), occurring
during the incubation period, in any of the soils,
even where N was added in the presence of trash.
This points to an extremely high degree of microbiological
activity in these coastal sands and helps to
explain the rapid disappearance of trash from them
under field conditions. The additional N released by
this rapid breakdown process may also partially ex-
TABLE IV
Degree of Trash Decomposition at the Turning Point
plain the somewhat indifferent response given by cane
to high applications of fertilizer N on this relatively
infertile soil.
II. THE EFFECT OF TRASH ON NITROGEN UPTAKE
BY SUGARCANE
Introduction
Part I of this paper showed that immobilization of
available N occurs when trash, which has a low N
content, is incorporated with the soil. It also demonstrated
that the addition of fertilizer N to soils containing
trash helps to overcome the depression of
available N and hastens the process of decomposition.
Although immobilization does not result in a direct
loss of N from the soil it is obviously able to compete
with plant uptake, and in the case of young cane
which apparently requires most of its N during the
first few months of growth', competition of this type
could seriously affect N uptake on certain soils, and
if severe ultimately influence yield.
In order to investigate this aspect of cane nutrition
the following greenhouse experiment was undertaken.
Procedure
1,500 g. air dry samples of two soil series (Cartref
and Clansthal sands) and 1,200 g. samples of a third
(Shortlands clay), representing between them a quarter
of the acreage of the Industry, were weighed into
polystyrene pots, and six replicates of each treated as
follows:

258
Proceedings of The South African Sugar Technologists' Association — March 1966
TABLE V
The Effect of Trash on N Uptake by Sugarcane Grown on Three Different Soils†

Proceedings of The South African Sugar Technologists' Association — March 1966 259
Figure 2. The effect of trash on total recovery of nitrogen by sugarcane from three soils after twelve weeks.
1. Control — no fertilizer or trash.
2. 0.5 per cent trash — 5 tons/acre equivalent.
3. 150 mg. N or *120 mg. N as (NH4)2S04 — 100
ppm. (200 lb. N/acre).
4. 150 mg. N or *120 mg. N as (NH4)2S04 + 0.5
per cent trash.
5. 150 mg. N or *120 mg. N as urea— 100 ppm.
(200 lb./acre).
6. 150 mg. N or * 120 mg. N as urea + 0.5 per cent
trash.
* Shortlands clay only.
The trash was uniformly mixed with the soils beforehand,
after which they were moistened to 50 per cent
of their water holding capacity (W.H.C.) with a basic
nutrient solution supplying 100 ppm. K and 80 ppm.
P, plus where appropriate, (NH4)2S04 or urea. Solutions
were applied down a perforated nylon tube,
permanently situated in each pot, in an attempt to
obtain even distribution of water and nutrients
throughout the soil. All pots were sealed at the base
to eliminate leaching losses.
The six replicates of each treatment were then split
and half of them planted with previously germinated
single-eyed cane setts, an inch in length and of approximately
uniform weight. The remaining pots in
each treatment were left unplanted.
For a period of 12 weeks all pots were weighed
daily and maintained at 50 per cent W.H.C. by the
addition of water down the tubes. At the end of this
time all treatments were harvested. Tops were cut
level with the sett, and after rapid air drying the roots
were separated from the soil and washed. Both tops
and roots were oven dried for 24 hours at 90° C. and
then weighed, after which they were finely ground and
stored prior to total N analysis. All soils were also
rapidly dried and ground to pass a 1 mm. sieve before
being analysed for mineral N.
Analytical details of the soils used are given in
Table I.
Results and Discussion
Yield data and N uptake by cane under the different
treatments for the three soils at the end of the 12-week
growth period are summarised in Table V, while
Fieure 2 shows the effect of trash on plant recovery
ofN.
While yields were depressed significantly on all soils
by the addition of trash to the control treatment, no
significant yield differences were found where N was
applied either as (NH4)2S04 or urea in the presence
or absence of trash. Trash however significantly reduced
total plant recovery of N on all treatments,
although there were no significant differences in effects
between control, (NH4)2S04, and urea. The difference
in fertility status of the soils and its effect on crop
available N is clearly seen in those histograms in
Figure 2 showing total plant recovery of N for the
control treatments in the absence or presence of trash.
Though it was not possible to assess accurately the
nett recovery of fertilizer N by the plant as 15 N labelled
fertilizers were not used, an estimate was made by
subtracting the average N uptake of the unfertilized
pots from the pots receiving fertilizer, but otherwise
treated alike, and expressing this figure as a percentage
of the fertilizer added. On this basis a somewhat

260
variable recovery was obtained in the absence of
trash, being least in the Shortlands soil (62-68 per
cent) and greatest in the Cartref sand (up to 86 per
cent). This suggests possible storage of N in the
organic phase of the more fertile soil. These results
showing massive uptake of N by cane within the first
few weeks of growth confirm those of Bishop 3 and
others 18 . The depressing effect of trash on percentage
recovery, which might have been expected was not
generally evident except where urea was added to
trash in the Cartref soil. This implies fairly rapid N
turnover and the utilization of much of the N released
during this process.
Inorganic N status of the soils. At harvest insignificant
quantities of mineral N remained in the soils
from the cropped pots under all treatments, confirming
the rapid uptake of N already shown. In the uncropped
pots, however, much of the added fertilizer
was recovered, and levels of NH4-N and NO3-N
found after the various treatments in the three soils
are given in Table VI.
Nett recoveries of N in the absence of trash ranged
from 75-90 per cent, losses of N occurring due to
volatilization, denitrification or other factors being
kept at a reasonably low level by moistening the soils
only to 50 per cent W.H.C. Recoveries reported else-
TABLE VI
Mineral N Remaining in Uncropped Pots after 12 Weeks
(means of three replicates in ppm.)
Proceedings of The South African Sugar Technologists' Association — March 1966
where in a comparable investigation 5 were as low as
56 per cent on similar sandy soils. The data also shows
the effect of trash in immobilizing part of the fertilizer
N, though without the use of 15 N it was not possible
to establish the size of the fraction held in the soil
organic phase. Turnover would appear to be reduced
in the absence of the growing plant.
Conclusions
From the data reported in this paper it is clear that
the presence of trash may influence to a greater or
lesser extent the availability of N to the cane plant
in all soils of the Sugar Belt. Where large quantities of
undecomposed material are found in the soil at time
of planting or are subsequently incorporated, these
may provide considerable competition between soil
micro-organisms and the developing plant for applied
N, and fertilizer dressings on some of the less fertile
soils might need to be increased to compensate for
this.
When it is recalled that the N applied in the greenhouse
experiment was at a high level (200 lb. N/acre),
while the amount of trash added was comparatively
small, (5 tons acre) the chances of reductions in N
uptake in the field are considerable. Average crop
dressings range from 60-120 lbs. N/acre while up to
20 tons/acre of trash may be present. In these circumstances
depression of N uptake due to trash
could affect yield, especially if it is able to keep plant
N supplies below the threshold value required for
optimum growth. Much of course will depend on the
fertility status of the soil concerned.
The data so far obtained has provided no information
regarding the quantity of fertilizer N immobilized
following an application, the fraction of this
that is rapidly remineralised and absorbed by the
plant, or the amount retained in the organic phase of
the soil. Using l5 N labelled fertilizers it is hoped to
establish more clearly the fate of fertilizer N when
applied to a wide range of soils, and to accurately
assess residual N effects and the importance of these
in the N economy of cane.
Acknowledgment
The author wishes to thank Mr. M. G. Murdoch
for statistical advice during the preparation of this
paper.
Summary
Incorporation of trash (10 tons/acre) into a wide
range of Sugar Belt soils caused rapid immobilization of
existing inorganic N supplies. Addition of (NH4)2S04
(200 lb. N/acre) with the trash, while stimulating
decomposition provided an N surplus in all soils.
As maximum N immobilization was attained a
turning point was reached in each of the soils when
mineral N was once more released into the system.
The time taken to reach the turning point varied from
approximately 8-41 days, broadly depending on the
fertility level, pH and texture of the soil.
Trash added to the Clansthal sand in the absence
of fertilizer N resulted in a high degree of microbiological
activity, which helps to explain its rapid disappearance
from this soil under field conditions.

Proceedings of The South African Sugar Technologists' Association — March 1966 261
In a greenhouse experiment addition of trash at 5
tons/acre significantly reduced N uptake by cane on
three different soils treated with (NH4)2S04 or urea,
though no significant yield differences were found in
the presence or absence of trash. Nett recovery of
fertilizer N was variable ranging from 62-86 per cent
in the cropped, and 75-90 per cent in the uncropped
pots.
References
1. Allison, F. E., and Cover, R. G. (1960). Rates of decomposition
of shortleaf pine sawdust in soil at various levels
of nitrogen and lime. Soil Sci. 89: 194-201.
2. Allison, F. E., and Klein, C. J. (1962). Rates of immobilization
and release of nitrogen following additions of
carbonaceous materials and nitrogen to soils. Soil Sci. 93:
383-386.
3. Bishop, R. T. (1965). Mineral nutrient studies in sugarcane.
Proc. Annual Cong.S.Afr.Sugar Tech.Ass. 39: 128-
133.
4. Broadbent, F. E., and Nakashimu, T. (1965). Plant recovery
of immobilized nitrogen in greenhouse experiments.
Soil Sci.Soc.Amer.Proc. 29: 55-60.
5. Broadbent, F. E., and Tyler, K. B. (1962). Laboratory and
greenhouse investigations of nitrogen immobilization. Soil
Sci.Soc.Amer.Proc. 26: 459-462.
6. Chandra, P. and Bollen, W. B. (1959). Effect of nitrogen
sources, wheat straw and sawdust on nitrogen transformations
in sub-humid soil under greenhouse conditions.
J.Indian Soc.Soil Sci. 7: 115-122.
7. Chandra, P. and Bollen, W. B. (1960). Effect of wheat
straw, nitrogenous fertilizers, and carbon-to-nitrogen ratio
on organic decomposition in a sub-humid soil. J.Agric.
Food Chem. 8: 19-24.
8. Jansson, S. L. (1958). Tracer studies on nitrogen transformations
in soil with special attention to mineralisationimmobilization
relationships. K.LantbrHogsk. Ann. 24:
101-361.
9. Jansson, S. L., Hallam, M. J., and Bartholomew, W. V.
(1955). Preferential utilization of ammonium over nitrate
by micro-organisms in the decomposition of oat straw.
Plant and Soil 6: 382-390.
10. Munson, R. D., and Pesek, J. T. (1958). The effects of
corn residue, nitrogen and incubation on nitrogen release
and subsequent nitrogen uptake by oats: a quantitative
evaluation. Soil Sci.Soc. Amer.Proc. 22: 543-547.
11. Parker, D. T., Larson, W. E., and Bartholomew, W. V.
(1957). Studies on nitrogen tie-up as influenced by location
of plant residues in soils. Soil Sci.Soc.Amer.Proc. 21:
608-612.
12. Stojanovic, B. J., and Broadbent, F. E. (1956). Immobilization
and mineralisation of nitrogen during decomposition
of plant residues in soil. Soil Sci.Soc.Amer.Proc. 20:
213-218.
13. Waksman, S. A. (1924). Influence of micro-organisms upon
the carbon-nitrogen ratio in the soil. J.Agric.Sci. 14: 555-
562.
14. Winsor, G. W. (1958). Mineralisation and immobilization
of nitrogen in soil. J.Sci.Food Agric. 9:792-801.
15. Winsor, G. W., and Pollard, A. G. (1956). Carbon-nitrogen
relationships in soil. 111. Comparison of immobilization of
nitrogen in a range of soils. J.Sci.Food Agric. 7: 613-617.
16. Wood, R. A. (1964). Assessing the potential of Sugar Belt
soils to supply nitrogen for plant cane. Proc.Annual Cong.
S.Afr.Sugar Tech. Ass. 38: 176-179.
17. Wood, R. A. (1965). Mineralisation studies on virgin and
cultivated Sugar Belt soils. Proc.Annual Cong.S.Afr.Sugar
Tech.Ass. 39: 195-202.
18. Yuen, C. H., and Borden, R. J. (1937). Chemical analyses
as an aid in the control of fertilizer application. Hawaiian
Plant.Rec. 41: 353-383.
Dr. Thompson: Mr. Wood and I disagree as to
what is a small amount of trash. Ten tons of trash
per acre would come from forty tons of cane, or if
dry, sixty tons of cane. Nowadays probably only two
or three tons of trash per acre is incorporated in the
soil. During ratoons there is no incorporation, and
action takes place at a mulch: soil interface. I would
suggest that the raising of the C/N ratio and the
immobilisation effect is therefore far less in field
practice than would be found in these greenhouse
trials.
Mr. R. A. Wood: My remarks were directed mainly
towards nitrogen immobilisation occurring under
plant cane. I have seen the topsoil in newly planted
cane fields full of undecomposed plant material
from the previously ploughed out cane crop.
Mr. Wilson: Even if a crop is burned there must be
plenty of organic matter left in the form of stubble,
roots etc.
Dr. Thompson: It would probably amount to about
four tons and would occur everywhere, whether you
practice trash conservation or not.
Mr. K. Alexander: Mr. Wood's paper will indicate
to farmers that where they apply manures such as
filter cake, which are high in carbon and low in
nitrogen, immobilisation does occur and nitrogen
will be needed to supplement the manure.
Dr. Matic: Referring to the previous paper on
nitrification should not other factors such as temperature,
soil moisture, etc. also be considered together
with pH?
Mr. R. A. Wood: Of course they should, but it is
impossible to examine several different parameters
simultaneously in such work, and consequently optimum
conditions must be provided for all the others
to which you referred so that the effects of variable
pH can be measured. I do not think that pH is any more
important than any other factor controlling nitrification
but it had to be examined in detail to provide
an explanation for the set of results obtained.
Mr. Coignet: Is mechanical cultivation of ratoons
justified in order to reduce immobilization of nitrogen?
Mr. R. A. Wood: Cultivation of this type through
aeration and partial drying might stimulate nitrogen
mineralisation and consequently reduce any immobilization
effects.
Mr. Moberly: It has been said that the ammonium
form of nitrogen is immobilised more easily than the
nitrate form. If you are aware that your fields contain

262 Proceedings of The South African Sugar Technologists' Association March 1966
considerable organic material would it not be better
to apply a fertiliser containing more nitrate than
ammonium?
Mr. R. A. Wood: Although the ammonium nitrogen
is preferentially utilised by the micro-organisms
causing decomposition, the nitrate nitrogen will also
be immobilised for this purpose once the ammonium
form has been used up, so that little would be gained
by applying more nitrate than ammonium fertiliser.
Mr. Landsberg: Is it not possible to find out in a
field how much organic material is present and then
calculate the amount of nitrogen required to satisfy
the micro-organisms present?
Mr. R. A. Wood: Yes it would be possible if the
average C/N ratio of the organic material were known
together with the amount of nitrogen supplied by the
soil during mineralisation.
Mr. Pearson: When green manure used to be used
many years ago its rate of disappearance in the soil
was extremely quick, so that it is quite possible that
if there was any delay in planting after it was turned
in there would be no response from the new crop.
Mr. R. A. Wood: This is quite possible as the green
manure would have a high nitrogen content, and
therefore on decomposition nitrogen would be
immediately released into the soil, no immobilisation
occurring.
Mr. du Toit. I agree with Dr. Thompson that there
is very little incorporation of organic matter with
ratoons. Before the days of the trash blanket there
was mechanical cultivation in the line and it was
difficult to get a response to nitrogen in ratoons.
Plant cane crops benefit from nitrogen mineralisation
resulting from ploughing and aeration and consequenty
nitrogen responses are not as good in plant
cane as in ratoons.
Trash does disappear quickly on the coastal sands
but is that the cause of their lack of response to nitrogen,
as there should be quick immobilisation?
Mr. R. A. Wood: Yes, there is rapid immobilisation
but subsequent release of nitrogen is also far more
rapid than in other soils even in the absence of added
nitrogen. I think that nitrogen released in this way
could be utilised fairly soon by plant cane so reducing
the response to fertiliser nitrogen.

Proceedings oj The South African Sugar Technologists' Association — March 1966 263
SOME CHARACTERISTICS OF THE SOILS OF THE SUGAR-
CANE GROWING AREAS AROUND MALELANE-KOMATI-
POORT, EASTERN TRANSVAAL
Introduction
Except for the soils around Nelspruit on which
citrus has been cultivated for a considerable time,
relatively little information is available regarding the
physical and chemical properties of the soils in this
region of the lowveld into which the South African
sugar industry has recently expanded.
The object of the present paper is to remedy this
situation in some degree by providing basic information
about the soils of these new sugarcane growing
areas. It must be emphasised, however, that the results
presented here, for the most part represent
determinations on single soils only. The soils used,
however, were chosen as representative of each type
of soil occurring in the area. Hence these figures should
be used as a general guide only, as it is possible that
other seemingly similar soils may vary quite considerably
as regards their properties and characteristics.
Location
The area under consideration extends at intervals
along the Crocodile River from the vicinity of Schagen
in the west to Komatipoort in the east. In addition
further areas of development are located south and
southwest of Komatipoort on the Komati and Lomati
Rivers.
A very hilly tract of granite terrain east of Nelspruit
conveniently divides the valley of the Crocodile
River into upper and lower portions. The upper valley
around Nelspruit comprises an area of extensive
citrus cultivation, the only area in this portion of the
Crocodile Valley scheduled for sugarcane production
being a relatively small area around Schagen, some
ten miles west of Nelspruit.
The lower Crocodile Valley, however, is a major
area of sugarcane production. This area until recently
has also been one of citrus cultivation and winter
vegetable production. The area of cane production
extends eastward as a long narrow belt from Kaap-
by R. R. MAUD and E.A. von der MEDEN
Table I
Mean Monthly Distribution of Temperature °C
muiden in the west, past Malelane and Hectorspruit
to Komatipoort. It is confined to the south bank of
the Crocodile River as this river here forms the southern
boundary of the Kruger National Park. The main
Johannesburg-Lourenco Marques railway line runs
through the area, with the sugar mill to serve the
region currently being built at Impala siding between
Malelane and Hectorspruit.
In the east, sugarcane growing is also being undertaken
adjacent to the Komati and Lomati Rivers,
above the confluence of the Komati and Crocodile
Rivers. On the Lomati River the area of cultivation
extends into a region of higher rainfall in the hills
south of Malelane, in the Kaalrug area.
A further relatively small area of sugarcane production
is located in the valley of the Kaap River
and some of its tributaries, south of Kaapmuiden.
Elevation in the sugar growing areas varies from
some 500 ft. above sea-level in the east around
Komatipoort, to about 2,500 ft. in the west at Schagen.
Climate
The climate of the region is important in that in
addition to playing a major part in the determination
of the types of soil that are found, it also is a factor
of major agronomic significance. The region is one of
summer rainfall characterised by relatively high
temperatures but relatively low rainfall. Thus nearly
all sugarcane in the region has to be irrigated. Temperatures
generally increase eastwards but rainfall
decreases. To the south, rainfall increases very
markedly with altitude in the Swaziland mountain
lands while temperatures correspondingly decrease.
The mean monthly distribution of temperatures at
four localities in the region is given in Table I. In
Table II is given the mean monthly distribution of
rainfall at three localities, while in Table III is given the
mean annual rainfall at a number of localities in the
region. (Weather Bureau, 1954).

264
Table III
Mean Annual Rainfall (mm.)
Vegetation
The natural vegetation of the region is typically
that of the lowveld and comprises the "Lowveld-
Type, (10)", of Acocks, (1953). It consists typically
of open Acacia nigrescens — Sclerocarya — Themeda
savannah. With increasing elevation and rainfall this
gives way either to open parkland with tall wellspaced
trees in tall grassveld, or else bushveld dotted
with big trees. These constitute the "Lowveld Sour
Bushveld, (11)", of Acocks, (1953).
Geology and Geomorphology
As is the case over much of Southern Africa, the
geology and geomorphology of the region have had
a very marked effect on soil formation. Therefore in
the consideration of the soils of the region it is also
necessary to pay some consideration to the geology
and geomorphology of the area.
Oldest rocks in the region are those of the Swaziland
System which are some 3,000 million years old,
and which are therefore some of the oldest rocks on the
earth. The System is subdivided into the Onverwacht
Series consisting mainly of highly altered lavas, and
the upper Fig-tree Series which consists mainly of
altered shales. The Swaziland System is overlain by
the Moodies System which comprises an alternating
sequence of shales and quartzites. Associated with, but
somewhat younger than these rocks, are the intrusive
rocks of the Jamestown Igneous Complex. These
consist mainly of altered lavas and comprise various
types of basic schists, amphibolites, pyroxenites and
serpentines. As will be discussed later, the rocks of
the Jamestown Igneous Complex yield fairly extensive
areas of characteristic soil in the region.
These ancient rocks are all highly folded and trend
north-eastwards in a broad belt out of Swaziland
towards Malelane, Hectorspruit and Oorsprong.
Many horizons in these rocks, especially the quartzites
of the Moodies System, because of their resistance
to weathering, have given rise to the very hilly country
south of Malelane, Kaapmuiden and Barberton. In
this region the rocks stand almost vertically and
hence there are narrow deep aligned valleys between
Proceedings of The South African Sugar Technologists' Association — March 1966
Table II
Mean Monthly Distribution of Rainfall (mm.)
the high-standing resistant hills. The basic schists
and amphibolites of the Jamestown Igneous Complex,
however, are not resistant to weathering and the
Crocodile River has utilized this fact in determining
its course between Kaapmuiden and Hectorspruit.
The river in this sector flows roughly on the contact
of these rocks with the granite to the north, and hence
there is a relatively broad valley on the south bank of
the river on these rocks. In addition to this main area
of rocks of the Swaziland and Moodies Systems and
the Jamestown Igneous Complex, another smaller
area of these rocks trends eastward from beneath its
covering of younger Transvaal System rocks in the
west of the region, to the vicinity of Schagen. West of
Schagen the Transvaal System rocks give rise to the
very eroded face of the Great Escarpment.
Somewhat younger than the rocks of the Swaziland
and Moodies Systems and the Jamestown Igneous
Complex and intruded into it are a number of granites
of slightly different types. Thus areas of coarse
crystalline granite-gneiss occur very extensively around
Nelspruit, west of Barberton, around Hectorspruit
and south along the Lomati River. There are a number
of types and ages of these granites. These types
of granite usually weather quite readily forming
lowlands. Another type of granite characteristically
more basic and of the migmatite type occurs east of
Nelspruit and extends northwards into the Kruger
National Park. This granite gives rise to very rugged
country because of its resistance to erosion. In places
the granite-gneiss is much intruded by dykes of diabase.
Although in the west of the region rocks of the
Transvaal System make the escarpment, the next
youngest rocks in the area are the Karroo sandstones,
some 150 million years old, which occur as a narrow
north-south trending belt in the east of the region.
The Karroo sandstones, of Ecca and Beaufort age
do not have any marked effect on the topography but
give rise to a number of rocky aligned ridges.
These sandstones are overlain in turn by the Stormberg
lavas, some 100 million years old. These consist
of a lower basalt member, and an upper andesitic
rhyolite member. The basalt overlies the Karroo
sandstones with the intervention of a narrow band of
Cave sandstone, which usually gives rise to a quite
prominent ridge. The basalt however, is readily
weatherable and gives rise to a broad virtually featureless
plain. The rhyolite by contrast is extremely
resistant to weathering and thus gives rise to the
Lebombo range north and south of Komatipoort.

266 Proceedings of The South African Sugar Technologists' Association — March 1966
All these rocks of Karroo age and younger dip eastwards
and the preferential weathering of the sandstones
and basalt has given rise to the broad meridional
lowland in the east of the region. This Lebombo
structure very probably forms part of the Great Rift
Valley system of Africa.
Erosion of the pre-existing rocks has been effected
by the rivers of the region which have selectively
removed and thereby lowered areas of the less resistant
rock, types. Much of this process has been achieved
within the last 2 million years but it has not always
been constant. It has been influenced by such things
as climatic changes and the occurrence of areas of
resistant rocks athwart the courses of the rivers which
slows down erosion upstream temporarily. These
periods of stillstand in the down cutting of the rivers
are reflected in the terraces which flank the major
rivers of the region. They are especially marked along
the course of the Crocodile River in its lower valley.
The alluvial soils deposited on these terraces have
now largely been removed by erosion subsequently,
but the continuing presence of abundant waterworn
cobbles and gravel attest to their former widespread
existence. The youngest terraces, however, survive
as sandy alluvia in the immediate vicinities of the
larger rivers. River terraces are best preserved in the
relatively broad flat valley bottoms of the Crocodile,
Komati and Lomati Rivers. In the narrow valley of
the Kaap River, however, much of the older terraces
has been lost and only the very youngest terraces
survive to any extent.
Nature and Occurrence of Soils
Soils derived from basic schists and amphibolites
In the mountain lands formed by the rocks of the
Swaziland System, soils are skeletal and lithosolic
except in lower slopes and in the valley bottoms where
they tend to be somewhat deeper. These soils tend
to be reddish in colour and of sandy loam to loam
texture for the most part. In a few localities, as in the
tributary valleys of the Kaap River, above Louw's
Creek, these soils are being cultivated for sugarcane.
Fairly extensive areas of reddish-brown clay soils
developed on rocks of the Jamestown Igneous Complex
occur along the south bank of the Crocodile
River from Kaapmuiden past Malelane to about the
vicinity of Impala siding where they give way to
granite-derived soils. The depth of these soils varies
considerably from a few inches to a number of feet.
In general the depth of the soil increases towards
the bottom of the valley. These soils may or may not
have varying amounts of river-terrace gravel incorporated
in them, usually as a stoneline above the
weathering rock. In the immediate vicinity of the
river, the soils become progressively sandier due to
alluvial influence, and lime concretions may appear
in the subsoil on lower slopes. The overall proportion
of soils having these latter characteristics is, however,
small.
These reddish brown clay soils are derived for the
most part from the various basic schists and amphibolites
of the Jamestown Igneous Complex. The
serpentines occur as a narrow band along the foothills
of the highlands on the southern margin of the
valley and have very shallow lithosolic soils developed
on them.
A further area of schist-derived soils occurs near
Schagen but again much of this area is hilly with
shallow soils, but deeper soils occur on the more
moderate slopes.
Figure 1 is a generalised map of the region showing
the distribution of the soils.
Particle size distribution and chemical characteristics
of these schist-derived soils are given in Appendices
1 and 2.
Soil No. 1 is a very shallow soil, Soil No. 2, one of
moderate depth and Soil No. 3 one of considerable
depth, all from Mhlati section, Malelane. Soil No. 4
is a deep soil from near Schagen.
Consideration of the cation exchange data indicates
that these are generally very fertile soils, although
depth, of course, will be a limiting factor in some
instances in their agronomic usage. It is of interest
to note the slightly lower base status and pH of the
soil from Schagen, probably the result of the slightly
increased rainfall there.
This influence of increasing rainfall on the cation
status of soils is also very well marked in the case of
similar soils from the higher rainfall area of Kaalrug.
Soil No. 5 is a very deep red soil derived from schists
at Kaalrug and occurs on a middle-slope, while
Soil No. 6 is a similar soil from a lower-slope at
Lomati. The somewhat higher base status of this
latter soil may be due either to the somewhat lower
rainfall prevailing where it occurs than is the case at
Kaalrug, or it may be that because of its position on a
lower-slope it has not been leached to the same degree.
In its physical characteristics however this soil resembles
the more leached soils.
In spite of its lower base status the Kaalrug soil
is nevertheless a moderately fertile soil.
Soils derived from granites
Extensive areas of soils derived from granite occur
around Nelspruit and west of Barberton. Other areas
of similar soils occur in the vicinity of Hectorspruit,
the Lomati River near the Sterkspruit and in the Kaap
River valley.
The soils derived from these granites are usually
greyish brown to grey loamy sands and sandy loams.
They are of variable depth but shallow soils with a
depth of 18 inches or less predominate. In places
sometimes because of long-continued irrigation, some
of the granitic soils have developed marked hydromorphic
characteristics. This may lead in time to the
development of saline and alkali soils. In the neighbourhood
of soils developed on the basic schists and
amphibolites and the diabase dykes the granitic soils
become more reddish-brown in colour due to migration
of iron.

Proceedings oj The South African Sugar Technologists' Association — March 1966 267
Soil No. 7 is a shallow granitic soil from Hectorspruit
while Soil No. 8 is a similar soil from near the
Sterkspruit. Soil No. 9 is a granite soil with marked
hydromorphic characteristics from the Kaap River
valley, while Soil No. 10 is a similar but deeper soil
from near Schagen. Soil No. 11 is a soil developed on
mixed granite-schist colluvium near Schagen.
The particle size distributions and chemical properties
of a number of these granite-derived soils are
again given in Appendices 1 and 2.
The cation exchange status of these soils reveals that
they are considerably less fertile than the soils derived
from schists and amphibolites. In general these granitic
soils are only of moderate to low fertility. Nevertheless,
with correct agronomic practice, these soils can
be made to yield good crops.
Soils derived from Stormberg Basalts
Although a narrow area of Karroo sandstones
occurs immediately to the west of the basalt in the
east of the region, the soils developed on them are
usually shallow and of little agronomic importance as
it is unlikely that they will ever be cultivated to any
extent. For this reason soils of this type have not been
included in the present study.
The Stormberg basalt underlies a zone some eight
to ten miles wide that trends north-south in the east of
the region around Komatipoort. Soils developed on it
are a characteristic assemblage of red, chocolate and
black coloured clays. In places the soils have been
influenced by former river terraces, and especially
in the case of the red soils, may contain variable
amounts of river terrace gravel.
Depths of these soils vary considerably, but the red
soils tend to be rather shallow. Thus the average
depth to the basalt saprolite at Tenbosch, north of
Komatipoort is some 18 to 22 inches. The chocolate
and black soils are usually somewhat deeper. Some of
the red soils may contain nodular ironstone gravel in
places, and in others may have diffuse calcium
carbonate in the subsoil and weathering rock. The
chocolate clays are intermediate between the red and
black soils. The black clays if they are deeper than
3 ft. usually exhibit the phenomenon of "self-ploughing"
or "self mulching". This is caused by the great
activity of the clay mineral montmorillonite under
conditions of varying moisture. In these soils lime
concretions occur throughout the soil profile, but in
the case of shallower black soils this is not always the
case.
Intrusions of Karroo dolerite of the same age as
the basalt into granitic areas to the west, as around the
Sterkspruit, give rise to similar red, chocolate and
black coloured clay soils.
The andesitic rhyolites which give rise to the
Lebombo range, have for the most part only skeletal
or very shallow soils developed on them because of
the topography to which these rocks give rise. For
this reason they are not cultivated.
In the appendices Soil No. 12 is a shallow red
basaltic clay from Tenbosch, Soil No. 13 is a somewhat
deeper similar soil from south of Komatipoort,
Soil No. 14 is a chocolate coloured clay from the
farm Grimman south of Komatipoort on the Komati
River, whilst Soil No. 15 is a very heavy "selfmulching"
clay from the same area. Soil No. 16 is
a black clay derived from Karroo dolerite in the
Sterkspruit area.
Consideration of the cation exchange status of all
these soils reveals that they are very fertile, the black
soils even more so than the red. This feature is offset,
however, by the more undesirable physical properties
of the black soils such as extreme plasticity when wet.
The good base status of the red soils may also be offset
to some degree by their tendency to shallowness.
Alluvial soils
Areas of alluvial soils are not reflected on the
generalised soil map of the region because of their
impersistence and relatively small areas of occurrence.
It is not possible to delineate these soils at the
scale of the map.
These alluvial soils are always located in the immediate
vicinities of the larger rivers of the region and
are usually a relatively narrow zone associated with the
youngest of the river terraces. These alluvial soils are
characteristicallly brown to reddish in colour and are
deep and sandy. They may or may not exhibit depositional
layering, sometimes with clayey horizons.
In the valley of the Kaap River, however, there occur
in places reddish brown more clayey alluvial soils
associated with a slightly older river terrace, as well
as the youngest sandy alluvia.
In the appendices Soil No. 17 is a sandy alluvial
soil from near Hectorspruit, while Soil No. 18 is the
more clayey alluvial soil from Revolver Creek in the
Kaap valley.
The cation exchange status of the sandy alluvial
soils tends to be low and hence these soils are ordinarily
of low fertility. In the case of the clayey alluvial
soil, however, the moderate to good base exchange
status indicates that these soils are fertile.
Possible development of salinisation and alkalisation
{brak)
As the soils of the region will all be irrigated in
order to produce sugarcane, their tendency to salinisation
with time as a result of irrigation has to be
considered.
It is unlikely that any salinisation will be induced
in the soils as a result of the quality of the applied
irrigation water, as this is for the most part of very
good quality containing only small amounts of dissolved
salts. Of the soils occurring in the region the
most susceptible to salinisation are undoubtedly those
derived from granite. It is unlikely that under normal
agronomic conditions these soils will become saline,
but should in places there be consistent over-irrigation
and lack of provision of drainage in these soils, sali-

268
nisation will very likely occur. This will be more
marked in the hydromorphic bottomland soils than
those on slopes. It is very unlikely that salination
would ever develop to any degree on any of the other
soils of the area because of their cation status in some
cases, and their drainage characteristics in others.
Nevertheless even on these soils irrigation water
should be controlled as far as possible by the lining
of canals and accurate metering of irrigation applications.
In addition adequate provision should be
made for drainage where necessary.
Soil Series
More specific definition of the soils of the region
than is possible on the broad genetic grouping is in.
terms of soil series. Soil series are defined in terms of
the soil profile alone and are named after the locality
where they are first described. All soils with similar
profile characteristics are given the same series name.
Soil series, however, may be modified by "phase"
differences, which indicate that while the essential
characteristics of the series remain, they may be
modified by such features as depth, stoniness, etc.
In order to avoid duplication of series names for
the same soil recognised in different places, in South
Africa there has been adopted a system of registration
of soil series. All soils have to be checked against the
series register to ascertain whether similar soils have
already been described and named. Only if a series'
properties differ from those already registered can a
new series be named and subsequently added to the
register.
The register of South African soil series has been
compiled by Macvicar, Loxton and van der Eyk
(1965), and the characteristics of all series named thus
far are contained therein. In addition in the sugar
belt of Natal, the soil series occurring there have been
described by Beater, (1957, 1959, 1962), many of the
soils occurring there also having been included in the
soil series register.
In the region under consideration, the red clay
soils derived from the basic schists and amphibolites
are referable for the most part to the Glendale series.
The characteristic sand content of these soils serves
to differentiate them from the seemingly similar soils
developed on the Stormberg basalt. There are a number
of phases of the Glendale series present in the
region, including shallow, deep and stony phases.
The shallow stony phase appears to predominate.
The lower cation status of the Kaalrug upper and
midslope soils resulting from the higher rainfall of
this area, is however, more characteristic of the
Bellevue series.
The major soil series of the granite-derived soils is
the Glenrosa. This series is typically shallow. In
places however, as on lower slopes, it may become
deeper and sandier and it then assumes the characteristics
of the Grovedale series. A hydromorphic soil
developed on granite has not as yet been registered,
Proceedings of The South African Sugar Technologists' Association — March 1966
but the Eldorado series of the sugar belt described by
Beater (1959), does have hydromorphic characteristics.
The shallow hydromorphic granite soils then can
provisionally be assigned to this series. The deeper
hydromorphic soil resembles in some respects the
Grovedale series. No name has been established as
yet for the soil developed on mixed granite-schist
colluvium. In the present region the area of such soils
is limited and consequently the introduction of a new
name is not warranted until its occurrence elsewhere
together with its characteristics have been proved.
The red clay developed on Stormberg basalt and
dolerite is referable to the Shortlands series, for the
most part the shallow phase. In the present region
subsoils are non-calcareous as is typical of the Shortlands,
but where calcium carbonate occurs in the lower
profile a new series will have to be named if extensive
areas of such a soil are proved to exist. The intermediate
chocolate coloured clays are referable to the
Kiora series, while the calcareous black clay belongs
to the Arcadia series. Non-calcareous black clays
are referable to the Rydalvale series. Waterlogged
clayey calcareous bottomland and vlei soils belong
to the Rensburg series.
Ordinarily alluvial soils because of their youth and
absence of genuine profile development are not
defined as soil series. In the Natal sugar belt, however,
an alluvial soil exists which is sufficiently constant in
characteristics and of widespread occurrence to warrant
definition as the Shorrocks series. The sandy
river terrace alluvial soils of the present region resemble
the Shorrocks series to some degree but are
not as red in colour and appear to be more sandy.
Again, because of its seemingly limited occurrence
the reddish clayey alluvial soil of the Kaap valley
has not been ascribed a series name as a similar soil
has not as yet been found to occur extensively elsewhere.
Waterlogged acid alluvial and vlei soils are
ascribed to the Katspruit series.
Because of the similarity of some of the soils of the
present region with those occurring in and described
from neighbouring Swaziland, mention is necessary
about the correlation of similar soils from these two
areas. Thus the Glendale series of the present area
closely resembles the Lesibovu series of Swaziland
(Murdoch and Andriesse, 1964), whilst the Bellevue
has its counterpart in the Malkerns series of that
territory. Similarly the Glenrosa series has as its
equivalent the Otandweni series of Swaziland, whilst
the hydromorphic Eldorado series is evidently
paralleled by the Habelo series. Of the soils derived
from basalt, the Shortlands series has as its equivalent
the Rondspring series of Swaziland, the Kiora is
paralleled by the Canterbury series, whilst the Arcadia
has as its equivalent the Kwezi series. The sandy
alluvial soil of the present area resembles the Bushbaby
series of Swaziland, although the redder Shorrocks
soils are referable to the Winn series described
by Murdoch and Andriesse, (1964).

Proceedings of The South African Sugar Technologists' Association — March 1966 269
In summary, the soil series found thus far to occur
in the region under consideration are:
Soils derived from basic schists and amphibolites,
Glendale, (Bellevue).
Soils derived from granites,
Glenrosa, Eldorado, (Grovedale).
Soils derived from Karroo dolerites and Stormberg
basalt,
Shortlands, Kiora, Arcadia, (Rydalvale).
Soils derived from alluvial sediments,
(Shorrocks), (Katspruit), (Rensburg).
Physical Properties of the Soils
(Moisture Characteristics)
The moisture characteristics and bulk densities of
the soils described from the region are given in Appendix
3.
Field capacity was determined on single undisturbed
cores drained at 100 cms. tension. This value has
been found to correspond very closely with field
capacities determined in the field for most soils of
the Natal sugar belt. Triplicate, disturbed samples
were used for wilting point determinations in a pressure
membrane apparatus. In Appendix 3, available
moisture (ins./ft.) is given for each soil horizon, these
values then being used on a pro rata basis to calculate
the total available moisture in the first two feet of the
soil profile. These figures do not constitute an inflexible
guide in the determination of the actual
amount of moisture available to the plant as rooting
depth, for example, varies considerably in different
soils. Thus in soil Nos. 3, 4, 5, 6, 10 and 17, rooting
may take place to a depth in excess of 4 ft., whilst on
shallower soils as Nos. 1, 7, 8 and 12, it may be little
more than 1 ft.
Glendale series:—These soils have in general
relatively high available moisture characteristics,
although this is not the case in soil No. 1. Data from
the first horizon only of this soil is presented as the
stoniness of the underlying horizon precluded the
taking of undisturbed samples. In the estimation of
the total available moisture of this soil, however, data
from disturbed samples were used as a guide.
Bellevue series:—These rather more weathered soils
have a slightly lower available moisture status than
do those of the Glendale series. This is marked in the
case of soil No. 6 with its very high clay content in the
second horizon. This high clay content together with
the relatively high bulk density would probably
restrict rooting depth to some extent in this soil.
Glenrosa series:—Soils of this series have generally
lower available moisture characteristics than those
described above. In the case of soil No. 7, undisturbed
samples could not be taken from the second horizon
of this profile because of its shallowness and hence
the figure for total available moisture has been
estimated from data for the disturbed soil. Because of
considerations of effective rooting depth the available
moisture in these soils of the Glenrosa series may in
practice be less than the amount given in Appendix 3.
Eldorado and Grovedale series:—In the case of soils
of these series the relatively high total available moisture
value of 2.7 ins. per 2 ft. may have to be used
with caution, bearing in mind the tendency of these
soils to waterlogging.
Shortlands series:—These soils have in general
relatively high available moisture characteristics.
Again, shallowness of soil as in the case of soil No. 12
may be a limiting factor.
Kiora. series:—The moisture properties of this soil
are intermediate between those of the Shortlands
and Arcadia series.
Arcadia series:—Soils of this series have very high
available moisture characteristics due to their content
of the clay mineral montmorillonite. As was the case
with their very good cation exchange status, their
high available moisture characteristics are offset by
other undesirable physical properties.
Alluvial soils:—These soils naturally vary very
widely in their characteristics both chemical as well
as physical and the two examples included in this
study provide some measure of the variability in the
properties of these soils that may be expected. Consequently
the application of the data given herein
would have to be made with great caution in its
extension to other alluvial soils.
Although the range in moisture characteristics of
the soils of this region is considerable, provided
cognisance is taken of soil depth and possible waterlogging
in some soils, the data presented here should
provide a reasonable guide for irrigation practice.
Conclusions
It has been shown that the soils of the region vary
quite widely with regard to both their chemical and
physical characteristics. The better soils can be expected
to provide excellent crops of sugarcane in
view of the climatic conditions that prevail. There is
no reason, however, why the poorer soils with correct
agronomic practice should not also yield very good
crops of cane in the light of the results of the present
investigation.
Acknowledgments
The authors would like to acknowledge the assistance
of Mr. E. de Lange in the field and that of Mr.
A. Rowell in the determination of the cation exchange
data.
Summary
The soils of the Malelane-Komatipoort region
reflect very closely the geology of the area and vary
considerably in their chemical and physical properties.
Soil series have been established for the area and data
with regard to the cation exchange status and moisture
characteristics of these soils are presented with a view
to their forming a general guide in agricultural practice.

270 Proceedings of The South African Sugar Technologists' Association — March 1966
References
Acocks, J. P. Veld types of South Africa. Bot. Surv. S. Afr.
Memoir, No. 28, 1953, pp. 47-50.
Beater, B. E. Soils of the Sugar Belt. Part 1. Natal North Coast.
Oxford Univ. Press, Cape Town. 1957.
Beater, B. E. Soils of the Sugar Belt. Part 2. Natal South
Coast. Oxford Univ. Press, Cape Town. 1959.
Beater, B. E. Soils of the Sugar Belt. Part 3. Zululand. Oxford
Univ. Press, Cape Town. 1962.
Loxton, R. F. A modified chart for the determination of basic
soil textural classes in terms of the international classi­
APPENDIX 1
PARTICLE SIZE DISTRIBUTION OF SOILS
fication of soil separates. S. Afr. J. Agri. Sci. 4, 1961.
pp. 507-512.
Macvicar, C. R, Loxton R. F., and van der Eyk, J. J. South
African Soil Series. Parts 1 and 2. Soils Res. Inst. Rep.
107/64. Pretoria. 1965.
Murdoch, G. and Andriesse, J. P. A soil and irrigability survey
of the lower Usutu basin (south) in the Swaziland lowveld.
Dept. Tech. Co-op. Overseas Res. Publ. No. 3. H.M.S.O.
London. 1964.
Weather Bureau. Climate of South Africa. Parts 1 and 2.
Govt. Printer, Pretoria. 1954.

272
Proceedings of The South African Sugar Technologists' Association — March 1966
APPENDIX 3
SOIL MOISTURE CHARACTERISTICS
F.C. = Field Capacity. W.P. = Wilting point. B.D. = Bulk density. A.M. = Available moisture.
Total available moisture in surface 2 ft. of soil. N.B. Allowance must be made in regard to depth of shallow soils in the application
of these figures.
Mr. Halse: I understand that cane is growing very
well on gravelly basalt soils in Swaziland although
you say that this would not be the case on the Transvaal
soils.
Dr. Maud: At Ubombo Ranches in Swaziland the
pebbles and gravel comprise mainly weathered and
porous basalt whereas the Transvaal soils contain very
hard, unweathered and non-porous gravels.
Mr. Turck: At Ubombo Ranches the gravels are
of weathered basalt and the underlying rock is also
porous.
Mr. von der Meden: We did not suggest that cane
would not grow well on these stony Transvaal soils,
but rather that rooting will be restricted to a degree
and total available moisture appreciably reduced.

Proceedings of The South African Sugar Technologists' Association—March 1966
NOTE ON SALINITY LIMITS FOR SUGARCANE IN NATAL
Soil samples from areas in the Natal Sugarbelt are
frequently received by the Experiment Station for
salinity appraisal. The effect of saline soils on plant
growth has been studied extensively 7 , but comparatively
little work with regard to their effect on
sugarcane has been reported, especially in South
Africa.
Maud 4 described two alkali soils in Zululand where
the cane had died and the conductivity* was over
10.0 millimhos/cm. throughout the profile. An
adjacent profile on a field growing good cane had a
conductivity less than 0.5 mmhos/cm. Preliminary
studies in Iran 6 suggested a conductivity of 4 mmhos/
cm. as the threshold value above which cane growth
will be drastically reduced. It was concluded from a
salinity survey in Swaziland 2 that cane will always be
severely restricted or killed above 5 mmhos/cm.,
whereas fair to good growth will normally be recorded
below 2 or 3 mmhos/cm.
From the above it would appear that sugarcane
growth is seriously restricted above a value of 4-5
mmhos/cm., but accurate limits for good cane growth
in Natal were uncertain. An area showing numerous
outbreaks of brackish conditions is the Nkwaleni
Valley in Zululand, where a preliminary examination
of salinity was therefore undertaken.
Soil samples were taken at depths of 0-9 in. and
9-18 in. in transects across areas in which the cane
showed varying degrees of impaired growth due to
saline conditions. Conductivity, pH, and exchangeable
sodium determinations were carried out on all
samples. For comparison with the tentative limits
suggested by Richards 5 , the exchangeable sodium
percentage (ESP) was also determined on the 0-9 in.
samples.
Results
Cane growth at each sampling point was designated
good, fair, very poor, or dead. The means of results
from several quite widely separated sites are given in
Table I. In all cases the soils were fairly heavy and
derived from either Ecca shale or river terrace.
Of the 18 samples in the good category, 15 had
conductivities less than 2.0 mmhos/cm., and all were
* The term "conductivity" refers throughout to the conductivity
of a saturation soil extract at 25° C.
By E. A. von der MEDEN
TABLE I
273
below 3.0 mmhos/cm. All samples in the very poor
and dead groups had conductivities above 5.0 mmhos/
cm. Figure 1 illustrates the relationship between cane
growth and average conductivity within the surface
18 in. of soil.
Discussion and Conclusions
Sugarcane growth is drastically reduced on soils
with conductivities above 4.0 mmhos/cm., this being
in agreement with the findings of Lea, Murdoch and
Dicks 2 and Shoji and Sund 6 . It is also clear that cane
grows well on soils with conductivities below 2.0
mmhos/cm. Between 2.0 and 4.0 mmhos/cm., cane
growth would appear to be adversely affected, but
to a variable degree and this trend is similar to that
found by the above workers. Although data for only
two sets of samples are quoted in the fair group
because the transition from good to very poor
cane was usually rapid, the marked effect on cane
growth of an increase in conductivity much above
2.0 mmhos/cm. is apparent from Figure 1. The
following tentative conductivity limits are therefore
suggested for sugarcane growth in Natal:
On the basis of the U.S.D.A. scale of salinity 5 ,
sugarcane would thus be classified as a sensitive crop.
In the same publication, an ESP of 15 is tentatively
suggested as the boundary between alkali and nonalkali
soils. This compares as follows with the figures
in Table I: For good cane, ESP was never higher than
10 and this could be taken as the limit for normal
growth. Above an ESP of 10, growth seems likely to
be impaired and above 15 seriously reduced.
A highly significant correlation coefficient of
—0.699 between cane tonnage and salinity as measured
by conductivity was recorded in Iran 6 , and a "good
correlation" was also found in Swaziland 2 . From the
Chemical characteristics of soil samples taken in transects across areas of salt damage

274
Proceedings of The South African Sugar Technologists' Association-March 1966
Cane Growth
FIGURE 1. Relationship between cane growth and salinity (Nkwaleni Valley, Zululand)
results of the present survey it is concluded that
conductivity measurements provide a sensitive and
rapid estimate of soil salinity status which correlates
well with the observed sugarcane growth. With reference
to Table I, however, it is clear that such
measurements representing only the top 6 or 9 in.
of soil may be misleading, and for an accurate picture,
samples should be taken to a depth of 18 in. where
possible.
At all sites salinity appears to have developed either
from injudicious irrigation or, perhaps more commonly,
as a result of seepage from unlined canals.
Whatever the cause, the percolating water accumulates
dissolved salts, and where the water table is close
to the surface, these rise and concentrate in the upper
layers of soil, creating a high osmotic pressure against
which the cane must withdraw water. As Marshall 3
points out, the control of salt is mainly a matter of
controlling water movement. The work of Gardner
and Fireman 1 suggests that the movement of salt
near the surface is likely to be serious if the water
table is less than 39 in. and this was clearly demonstrated
in the present survey where the water table
was invariably above this height under poor growth
conditions.
A high water table (which is integrally related to
aeration and the availability of 0., in the soil profile)
and salt accumulation are obviously closely linked.
The removal of the cause of waterlogging therefore,
or the provision of suitable drainage systems, would
solve both problems in many cases.
References
1. Gardner, W. R. and Fireman, M. (1958). Laboratory studies
of evaporation from soil columns in the presence of a water
table. Soil Sci. 85: 244-249.
2. Lea, J. D., Murdoch, G., and Dicks, A. V. R. (1963). Report
on a salinity survey, Swaziland Lowveld, 1962-1963.
3. Marshall, T. J. (1959). Relations between water and soil.
Commonwealth Bureau of Soils, Harpenden, Tech. Comra.
No. 50.
4. Maud, R. R. (1959). The occurrence of two alkali soils in
Zululand. Proc. Annual Cong. S. Afr. Sugar Tech. Ass. 33:
138-144.
5. Richards, L. A. ed. (1954). Diagnosis and improvement of
saline and alkali soils. U.S.D.A. Handbook No. 60.
6. Shoji, K. and Sund, K. A. (1965). Drainage and salinity
investigations at the Halft Tapeh sugar cane project, Iran.
Proc. 12th Int. Sug. Tech. Cong., Puerto Rico. (In press).
7. Wilkins, R. A. and Athesian, K. H. (1965). Relationship
between ground water, salinity, and sugar cane at Rose
Hall Estate. Proc. 12th Int. Sug. Tech. Cong., Puerto Rico.
(In press).

Proceedings of The South African Sugar Technologists' Association — March 1966 275
Dr. Sumner: I wish to sound a word of warning.
Mr. von der Meden proposes these new conductivity
limits for grading soil salinity based on one soil type.
Presumably he is now going to extend these to the
whole of Natal although it has only been worked out
for one soil.
Mr. von der Meden: The area included several soils,
such as Middle Ecca, Lower Ecca, and boulder bed
derived types. Although these covered a range of
textures, they tended to be mostly on the heavier side,
but it is normally only on such soils that salinity
becomes a problem.
Mr. Hill: Good cane growth should also be related to
the amount of sodium present and not just conductivity
as some high conductivities may be due to calcium
and magnesium.
Mr. von der Meden: From our table you can see
that the conductivity reflects the amount of sodium
very closely. A high calcium could be very misleading,
but one would not be looking for salinity on such soils.
Mr. Armstrong: Did you use the soil cup or the
saturation, extract method when measuring conductivity?
Mr. von der Meden: We used the soil cup method.
Then, knowing the saturation percentage, we used
standard graphs to transform the readings to conductivity
of saturation extracts. We subsequently
determined conductivity of saturation extracts directly
on several of these samples covering a range of salinity.
Agreement between these results and those obtained
using the soil cup reading transformed by means of
graphs was very satisfactory.

276
Proceedings of The South African Sugar Technologists' Association - March 1966
AVAILABILITY OF SOIL WATER TO SUGARCANE 1
IN NATAL
Introduction
The theoretical aspects of water availability to
plants have been covered by Philip (1957) and Gardner
(I960), whilst Makkink and van Heemst (1956) and
more recently Denmead (1961) have presented
supporting experimental evidence. In a preliminary
report on water availability in some Natal sugarbelt
soils, Hill and Sumner (1964) briefly discussed the
concept and presented data to indicate that the
quantity of available moisture in certain soils varied
with evaporative demand. Since their results were
affected by the size of the cane plants, the confined
volume of soil, and by advective energy, it remained
for quantitative information to be obtained on this
concept under field conditions.
In irrigation planning two major factors for consideration
are the choice of suitable water duties and
irrigation frequencies or periods. Thus with any
single water duty, say 170 acres per cusec, there is the
choice of applying say applications of 1-inch every
7 days, or one 2-inch application every 14 days. Each
particular irrigation period has its own merits and
disadvantages. In general, longer periods favour
labour utilisation efficiency, but in many cases the
condition of the soil dictates the period. Certain soils
will not accept large applications of water and run-off
becomes a serious problem. On these soils, shorter
irrigation periods become essential, and the introduction
of a remunerative incentive can encourage
good efficiency in operation.
In order to shed further light on this problem, field
trials comprising various irrigation treatments were
laid down. These treatments were designed to yield
quantitative information on water availability to
sugarcane in different soils. It was reasoned that such
information could be used to evaluate the effects of
different irrigation periods on growth and yield of
sugarcane. The experiments and techniques used are
described below.
Methods
During the period 1963 to 1965, there were five
irrigation experiments in operation, four being located
on Windermere clay loam, and the fifth on Clansthal
sand. The experiments were of the randomised block
design with four replications of various irrigation
frequency treatments. The treatments themselves
consisted of a range of water application levels so
designed as to replenish the estimated soil moisture
deficit (from Class A pan evaporation) to field capacity.
Thus when the predicted soil moisture deficit reached
a specified level, then this total amount of water
was applied to bring that treatment back to field
capacity.
1 Part of a Ph.D. thesis submitted to the Department of Soil
Science, University of Natal.
By J. N. S. HILL
Owing to uncertainty in consumptive use data for
cane with incomplete canopy, no irrigation treatments
were started until the measured ground cover was
complete. Until that time, all treatments received
infrequent but heavy irrigations to relieve severe
moisture stress. On two experiments, one on Clansthal
sand, the other on Windermere clay loam, crop
development studies of the variety N :Co. 382 were
carried out. These involved monthly population
density counts, weekly canopy development and daily
stalk height measurements. The effects of increasing
soil moisture deficit could therefore be followed.
Some of these crop development data and the harvest
results of all these experiments are presented below:
Results
(a) Influence of irrigation frequency on stalk elongation
During the growth of the plant cane crops of N:Co.
382 in two experiments, the treatments were imposed
for a period of three months. A comparison between
irrigation frequencies of the effects on stalk elongation
for the Clansthal sand and Windermere clay loam
is shown in figure 1. Growth curves for the highest
FIGURE I: Relation between cane growth, evaporative demand,
and predicted soil moisture deficit in two
Natal soils: solid line represents stalk elongation
in both Clansthal sand and Windermere clay loam
kept at field capacity; broken line is Clansthal dry,
dotted line is Windermere dry.

Proceedings of The South African Sugar Technologists' Association—March 1966 277
frequency (½ inch irrigation when evaporation—E°—
accumulated to ½ inch) and the lowest frequency
(3 inches irrigation when E° accumulated to 3 inches
for Windermere soil; 4 inches irrigation when E°
accumulated to 4 inches for Clansthal sand) are
presented. As can be seen in figure 1, irrigation
frequency has little effect in Clansthal sand but
influences stalk elongation in Windermere clay loam
markedly during the high evaporative demand period.
After the first 18 days when average daily Class A
pan evaporation was 0.18 in., conditions became
much cooler and evaporative demand (average
0.10 in./day) was reduced until irrigation frequencies
lost their effect on stalk elongation.
Stalk elongation measurements were continued on
the first ratoon crops, which experienced one of the
most severe droughts in the history of the sugarbelt.
Thus the total irrigation applications were very
similar in all treatments and on both soil types. In
figures 2 and 3 the growth pattern of sugarcane under
various irrigation IreqiLencies is shown for Clansthal
sand and Windermere clay loam respectively. It can
FIGURE 2: Growth pattern of sugarcane in Clansthal sand as
influenced by irrigation frequency.
FIGURE 3: Growth pattern of sugarcane in Windermere clay
loam as influenced by irrigation frequency.
be seen that prior to harvest, irrigation frequency has
resulted in a wider spread of stalk lengths on Windermere
soil (10 inches between ½ and 3 inch frequencies)
than on Clansthal sand (6 inches between ½ and 4
inch irrigation frequencies). But, perhaps more
important, is the seasonal effect on stalk elongation
indicated in these figures. During the high evaporative
demand period of January, February and up to the
11th March, 1965, the slopes of the growth curves
for various irrigation frequencies are not greatly
different for Clansthal sand whereas far greater
differences are visible on Windermere clay loam. On
both soils, however, growth curves tend to parallel
between irrigation frequencies after about mid-March.
The influence of estimated soil moisture deficiency
on stalk elongation of sugarcane on these two soils is
best illustrated in figures 4 and 5. These figures represent
the high evaporative demand period (18th
January to 15th March, 1965) when mean Class A
pan evaporation was 0.24 in./day. Daily stalk elongation
measurements have been differentiated into
relative growth rate values (ratio of stalk elongation
at any particular soil moisture deficit to that on soil
kept close to field capacity by frequent irrigation).
FIGURE 4: Relative growth rate of sugarcane in Clansthal sand.
FIGURE 5: Relative growth rate of sugarcane in Windermere
clay loam.

278
These figures show that potential cane growth declines
when the estimated moisture deficit in the soil profile
is 1 inch for Windermere clay loam and 3 inches for
Clansthal sand.
(b) Influence of irrigation frequency on cane yield
Some of the earliest work on irrigation frequencies
was conducted at Tongaat by Dr. T. G. Cleasby.
Trials harvested in 1957, 1958 and 1960 (the earliest
years being high rainfall seasons) showed that there
was a tendency for N:Co.310 to respond to high
frequency irrigation on Windermere clay loam. These
harvest results are presented in Tables I, II and III.
TABLE 1
Response by plant cane N:Co.310 to different irrigation frequencies
* Irrigation treatments consisted of replenishing the soil
moisture to field capacity at the intervals noted. Soil moisture
deficit was estimated from progressive Symons tank evaporation
employing a factor of 0.85 to represent evapotranspiration,
whilst irrigation efficiency was taken as 75 %
L.S.D. (1 %) 10.16 T.C.A.; (5%) 7.07 T.C.A.
TABLE II
Response by first ratoon cane N:Co.310 to different irrigation
frequencies
TABLE III
Response by second ratoon cane N:Co.3lO to different irrigation
frequencies
Proceedings of The South African Sugar Technologists' Association—March 1966
Further experiments involved during the period
1963, 1964 and 1965 contained the varieties N.50/211,
N:Co.376, N:Co.382 and N:Co.310 in various stages
from plant cane until 4th ratoon. Harvest results are
presented in Tables IV to VII. Correlation coefficients
between stalk length at harvest and yield in tons
cane are presented in Table VIII. It can be seen that in
general sugarcane responds in tons cane per acre to
higher frequency irrigation on Windermere clay
loam, whereas no such response is noticeable on
Clansthal sand. In terms of response in yield over
dryland production, the results of all trials have been
"averaged" to produce the data in figure 6. In figure
6 the arbitrary linear subdivision of the x-axis is in
terms of average soil suction at the time of irrigation.
These intervals could also represent days of evaporation
(cycle length) or moisture deficit in the soil
profile (inches).
It is quite clear from figure 6 that the potential
response to irrigation is far greater on heavy soils
than on sand. The reasons for this are that sands
accept a higher fraction of natural rainfall (good
infiltration properties), support larger root volumes
and soil moisture is less readily available to sugarcane
in heavy soils during the months of maximum growth.
It is also seen in figure 6 that low frequency irrigation
(with its advantages) is very efficient on sand. Since
these findings project their consequences on irrigation
planning, they are discussed as such in the following
section.
Discussion and Conclusions
One fact which has emerged from these results is
that the potential response to irrigation is greatest on
heavy soils. This of course is due to the fact that it is
on these soils that drought effects are so severe. This
severity of drought is due to several factors. In general,
although there are notable exceptions, heavy clay
soils have poor infiltration properties and thus rainfall
acceptance is reduced, run-off increased. Thus for a
start, the cane crop has to thrive on a lower total
amount of soil moisture during the season. Secondly,
soil moisture in clays is less readily available to sugarcane
during periods of high evaporative demand. Thus
during the months of maximum growth, stalk elongation
is greatly affected by numerous small drought
periods.
The question of real importance is whether the
potential response, or yield of about 60 tons cane per
acre per annum can be achieved under field conditions
with irrigation. The experiments suggest that such a
potential is more easily reached on sands than on
heavy soils. Thus the argument resolves into two
parts:
(a) Factors of academic interest
Soil water is not equally available to sugarcane over
the range field capacity to the wilting points in all
soils and under all climatic conditions. In coastal
Natal, during the summer months December to
March, stalk elongation is reduced on heavy soils
after relatively small deficits of moisture have occurred

Proceedings of The South African Sugar Technologists' Association—March 1966
TABLE IV
Response by plant cane N:Co.382 to various irrigation frequencies on two Natal Soils
TABLE V
Response by N.50/211— and N:Co.376 first ratoons to various irrigation frequencies on Windermere clay loam
TABLE VI
Response by 1st ratoon N :Co.382 to various irrigation frequencies on two Natal soils
279

280
TABLE VII
Response by 4th ratoon N:Co.310 to various irrigation frequencies
on Windermere clay loam
TABLE VIII
Correlation coefficients between stalk length at harvest and yield
in irrigation experiments
in the soil profile. On such soils therefore, potential
yields can only be assured by low water duties and
high irrigation frequencies.
On sands, such a decline in stalk elongation is not
as noticeable and it is believed that this is explicable
in terms of amount and availability of soil moisture
in the root zone. Thus on these soils, close to potential
yields can be achieved with higher water duties and
much lower irrigation frequencies.
(b) Factors of economic importance
The attainment of potential crop yields under
irrigation, and indeed the entire concept of irrigation
planning with water duty considerations, etc., is
deeply involved with economics. Thus the water duty
necessary to produce potential crop yields might well
be economical on one soil but not on another.
It is believed that from these results it is possible
to be categorical only as concerns irrigation frequencies
(or periods). The economics of the water duty
question remain to be solved, although certain
suggestions can be put forward. If the yield results in
Table VI are studied, it will be seen that, although
sugarcane responds to high frequency irrigation on
Windermere clay loam, this response is not great.
When it is remembered that these results were obtained
during one of the driest seasons ever recorded, then
Proceedings of The South African Sugar Technologists' Association—March 1966
FIGURE 6: Response by sugarcane to irrigation in two soils
when the moisture tension at time of water
application varies.
such a difference can be seen in its true perspective.
Therefore it is suggested that on sands, low irrigation
frequencies are suitable. Applications of 2 to 3 inches
of water will be as effective in their own cycles as
would be more frequent irrigations of 1 inch. On
heavy soils, it is believed that any application up to
about 2 inches, within its own cycle, is suitable. On
these soils, however, physical properties often dictate
the application level and irrigations of 1 inch are
probably optimum.
Water duties may be extended on sands to take full
advantage of the large root-zone reservoir and close
to potential yields should be obtainable when other
factors are not limiting. On heavy soils it is believed
that lower water duties are advisable but, in fact,
suitable experimental evidence is not yet available
to substantiate such a remark. In conclusion, an
experimental technique is proposed that should shed
further light on this problem. An irrigation experiment
should have treatments based directly on water
duties. Once it is decided as to the application level
which the soil will safely hold, then different treatments
simply involve the application on different
cycles. Each cycle should furthermore be split for a
climatic effect such that if say, one treatment involves
irrigating 1 inch every 8 days commencing on the 1st
of a month, its conjugate treatment also applies 1
inch every 8 days but commences on the 4th of the
month. In this way, for each treatment, a climatic or
weather effect can be taken out in the same season.
Such an experiment is already under way at Tongaat
and results are awaited with interest.
Summary
Experiments conducted to investigate the availability
of moisture to sugarcane in different soils have
revealed that cane responds to high frequency irri-

Proceedings of The South African Sugar Technologists' Association — March 1966 281
gation on clayey soils but not on sands. This finding
influences the choice of irrigation application levels
and periods. Heavier, less frequent irrigations are
suitable for sands whereas there is a tendency for
lighter, more frequent irrigations to be better on
heavy soils.
Water duties may be extended on sands whilst on
heavy soils information regarding the economics of
different water duties is lacking. An experimental
technique has been proposed for yielding information
on the water duty question in irrigation experiments.
References
Denmead, O. T. (1961). Availability of soil water to plants.
Unpub. Ph.D. thesis. Iowa State Univ.
Gardner, W. R. (1960). Dynamic aspects of water availability
to plants. Soil Sci. 89, 63-73.
Hill, J. N. S. and Sumner, M. E. (1964). A preliminary report
on water availability in some Natal sugarbelt soils. Proc.
S.A. Sugar Tech. Assoc. 38th Congress.
Philip, J. R. (1957). The physical principles of soil water movement
during the irrigation cycle. 3rd Congr. Int. Comm. on
Irrig. and Drain. 125-154.
Mr. van Schalkwyk: Contrary to previously held
opinions, Mr. Hill now tells us that sandy soils require
large amounts of water at long intervals.
Mr. Hill: The main reason for this is that on heavy
soils the available moisture measured by the laboratory
technique can be very much higher than that
measured on sandy soils. If the correct technique is
used, however, that is soil bulk density effects are
taken into consideration, then it will be found that
there are no great differences between total available
moisture values for different soil types. Thus sandy
soils hold much the same amount of moisture for
plants as do clay loams per unit depth. This is where
relative availability of soil moisture comes in. All the
available moisture measured in a sand is easily withdrawn
by plants, whereas only a fraction of the total
available moisture measured in heavy soils is easily
taken up under hot, windy conditions.
Mr. Landsberg: I would like to mention the important
factor of rooting depth in different soils.
Doesn't Mr. Hill believe that his yield and stalk
elongation rate differences measured on different
soils can be explained in terms of different rooting
depths?
Mr. Hill: Rooting depth differences undoubtedly
play an important part in the overall concept of
irrigation. However, the results reported in this paper
seemed to indicate that even that moisture held within
the limits imposed by rooting depth was differently
available, depending on soil type. For example, in the
heavy soil the rooting depth of sugarcane was seen
to be approximately 2 feet. Yet this soil which held
1½ inches of available moisture per foot of soil, that
is about 3 inches of total available moisture, could
not support the potential growth rate when only as
little as 1 inch of water had been removed.
Dr. Thompson: I would like to mention the ratio
of potential evapotranspiration to Class A pan
evaporation for fully canopied sugarcane for a 24-day
period last October. These quantities were measured
daily and the mean ratio was found to be 0.98, and
this includes the effects of very stong advection on
several days.
In the succeeding first ratoon crop the ratio had
increased to approximately 1.0 by the time the vertical
ground cover was fifty per cent. It is therefore apparent
that, because of sun angle, effective canopy must
always be in excess of vertical ground cover.
Secondly, I think you will agree that the decision
to go for a low frequency of irrigation on a sandy soil
will depend on the slope of the response curve of the
type shown in Figure 6. Actual increased yield must
always be measured against the saving in labour
owing to less frequent irrigation.
Mr. Halse: By increasing the frequency more capital
costs are also incurred.
Mr. Hill: Not necessarily. If one is dealing with a
dynamic scheme—that is one already designed and
operating—no extra capital costs are involved. If,
however, one were to design a scheme with this
concept in mind, water duties might be altered and
this could incur additional capital costs.
Mr. du Toit: From Figure 5 it would appear that
growth of cane on Windermere soil could actually
increase between field capacity and a deficit of one
inch. Were any such measurements recorded?
Mr. Hill: On certain days we found that cane in
Windermere and Clansthal soils at relatively small
moisture deficits did exceed in stalk elongation, cane
growing in plots kept as close as possible to field
capacity. This is not easy to explain. I can only
suggest that photosynthesis and the production of
growth substances in the plant is less susceptible
to water stress than is actual production of dry matter
as reflected by stalk elongation. Thus on days of high
evaporative demand there is a net build up of photosynthetic
products which are not used up in actual
growth due to loss of turgidity in the plant cells. If the
following day is of a slightly lower evaporative demand,
then on this day those plants containing a carryover
of growth substances actually grow more than do
plants which had used up all their growth substances
the day of production.
Mr. Browne: In places like the Eastern Transvaal
where daily evaporation sometimes exceeds 0.40
inches, can we apply some factor to reduce the 1:1
ratio as an estimate of consumptive use?
Mr. Hill: I know of no evidence which puts an
upper limit on the transpiration rate of sugarcane.
However, I have seen data on maize grown in the
United States which puts the maximum transpiration
rate for this crop at 26 parts per day. This of course
implies that there is a maximum diffusion pressure
deficit which can be developed in the leaves. There
are several reports in the literature concerning this

282 Proceedings of The South African Sugar Technologists' Association—March 1966
maximum D.P.D. and, again for corn, I think 1 have
noticed a value of about 100 bars. Whether or not
we can use this value for sugar cane I am not sure.
Mr. Browne: Are investigations being pursued
anywhere to determine this value ?
Mr. Hill: I personally do not have the equipment
to do so, but perhaps Dr. Thompson can help us.
Dr. Thompson: We are following it up by looking
at the two things which can effect the ratio, namely
the relative advective effects in different areas and
the relative degrees of physiological control of transpiration.
Mr. Collings: Mr. Hill mentioned the possible
increase and decrease in capital investment and it was
said we could put 2 inches on a sandy soil instead of
1 inch. If you apply 1 inch on a seven day cycle or
2 inches on a fourteen day cycle you do not change
your capital investment or equipment. However,
1 inch applied on a 14 day cycle means a different
water duty and a reduction in field equipment per
unit area. Your mains and pumping unit, however, do
not necessarily change in proportion.
Mr. van Schalkwyk: It is the excess wear on equipment
through constant movement that is costly, not
the capital expenditure.

Proceedings oj The South African Sugar Technologists' Association—March 1966 283
THE USE OF PERFORATED PIPES FOR IRRIGATION
EXPERIMENTS
Early in 1963 a number of irrigation trials were
planned at Hulett's, Mt. Edgecombe to study and
improve the efficiency of water utilisation in sugar
cane on our irrigated Estates.
The practical difficulty of applying water to trials,
by overhead irrigation, with quarter rainers soon
became evident, due to the windy conditions that prevail
during the summer months. Even the slightest
breeze distorts spray pattern within the plot and creates
drift into adjacent treatments.
Simple trials with few treatments allow some latitude
in time to permit acceptable delays until calm
conditions return. More complicated and precise
trials leave no latitude for irrigation delays, and
consequently a more efficient and accurate method of
applying water was sought.
Thoughts were turned to underhead irrigation, and
a system utilising light portable 2 in. perforated piping,
commercially known as Perf-O-Rain, proved an
immediate success. The pipes are drilled to the
Perf-O-Rain 2 in. type F pattern and adapted to
operate at pressures indicated in Table 1.
Table 1
Flow rates of 2 in. perforated piping adapted from Perf-O-Rain
Type F Pattern
Irrigating a ten line plot
with eleven lengths of
piping.
Note Manifold in the
foreground.
by P. J. M. de ROBILLARD ond M. J, STEWART
The inability of this system to simulate overhead
irrigation by applying water to the foliage after canopy
was appreciated, but an even distribution of water
without fear of drift more than compensated for this
disadvantage.
The perforated pipes are laid along the length of
each interline in the plot, with additional lengths to
cover the two outside guard rows. A conventional size
plot of 1/40 acre, having six lines 40 ft. long requires
seven lengths of piping to ensure an even distribution
of water. Each length of pipe is designed to extend into
the guard row on either side at least 18 in. in order
to reduce end effect. For ease of handling, two pipes
21.5 ft. long are linked to provide a total length of
43 ft. Furthermore, the low operating pressures favour
the use of Bauer type couplings to avoid water leakage,
which is likely to be experienced with the Ames
system.
Water which enters the system under pressure is
metered to a 2 in. manifold with a 2 in. Saunders
valve at each take-off point for controlling pressure
and ensuring an equal flow of water to each interrow.
Adjustments are effected by observing the height of the
spray pattern, and testing pressures from a Schroeder
valve on each perforated pipe coupling. The Saunders
valve is particularly suitable for fine adjustments
which are to be made at each take-off point.
When a scheme is designed it is important to ensure
adequate line pressure in order to cover friction
losses and provide sufficient working head to each
perforated pipe at any point within the experiment.
Flow rates through a 43ft. length of perforated piping

284
Perforated pipes are laid
4 ft. 6 in. apart along the
centre of each interrow.
adapted from the 2 in. Perf-O-Rain type F pattern
are provided in Table 1.
The optimum operating pressure is one where the
spray pattern covers the interrow at a steady pressure,
which, in the case of 4 ft. 6 in. planting, falls within
the range of 2-3 lbs. per square inch.
The quantity of water to be applied is calculated in
gallons and simply metered to each plot. A plot of
6 lines 40 ft. long with a 4 ft. 6 in. spacing has a wetted
area of 1/32 acre covered by seven pipe lines 43 ft.
long. The quantity of water required to apply 1 in.
per acre is therefore 708 imperial gallons which is
delivered in 21 minutes at 3 lbs. per sq. in. working
pressure.
In view of the relatively high flow rates obtained
from this method of irrigating experimental plots,
water application techniques should be based on time
lapse applications for soils with low infiltration
capacities.
Water entering the system is piped from the meter
box by portable piping, or flexible plastic piping, to
the manifold. The distance of the plots from the meter
box is immaterial because measurements only commence
when water starts to flow from the perforated
pipes in the plot.
Proceedings of The South African Sugar Technologists' Association — March 1966
Various methods of assembling and moving pipes
from one plot to another can be developed to save
effort and time. After irrigating a plot, the water
contained in the perforated pipes should be exhausted
through a by-pass valve on the main supply system,
or a drainage pipe leading outside the experimental
block to avoid interfering with the water regime of
other plots.
Consideration was given to planting cane lines on
a gradient so that water would automatically drain
back into the manifold after irrigation, but at low
operating pressures of 2-3 lbs/sq. in. this would lead
to an uneven distribution of water within the plot.
Consequently the current method of draining involves
manual raising of the perforated pipe at the opposite
end to the manifold.
When pumping dam water the presence of debris
or sand is unavoidable, even if a fine mesh screen is
fitted to the foot valve. No trouble has so far been
experienced, because the sand remains at the bottom
of the pipes and any floating debris is washed to the
end of the pipe and exhausted when emptying the
lines.
Having drained the perforated pipes, it would be
unwise to uncouple the two pipes within the plot
and so induce soil compaction. Therefore, pipes are

Proceedings of The South African Sugar Technologists' Association—March 1966 285
uncoupled at the manifold and drawn through the
lines into the adjacent plot until the centre coupling
is reached. The two pipes can then be disconnected
and re-assembled in the next plot for irrigation.
Normally two people are required to operate the
irrigation system which entails the moving and assembling
of pipes, and water measurement. In an emergency,
one person can be left to perform these operations,
provided the heavy manifold assembly can be
disconnected into two or more units. However, the
loss of irrigating time is a hindrance to schedules,
and such situations should be avoided whenever
possible.
A take-off point on the
manifold showing a
Saunders valve, a Schroeder
valve for testing pressures
and a Bauer coupling to
perforated pipe
Dr. Thompson: We have used the author's data and
copied their equipment in laying down our first irrigation
experiment at the Experiment Station for six
years.
Perforated pipes will also be used for irrigating
seedlings and for the root laboratory.
We wish to conclude by adding that two experiments
at Hulett's Mt. Edgecombe, have been conducted
with complete success using this method of
irrigation. One experiment on Clansthal sands showed
no run off in plant or ratoon cane with an application
rate of 3 in/hour. The second experiment on a
Windermere series also showed no run off under a
trash blanket in the 1st ratoon. Control of soil moisture
levels and irrigation timing by means of the Neutron
Probe, conducted by the Research Agronomists
of the S.A.S.A Experiment Station, showed complete
coverage and even distribution throughout the cycle
of the two experiments.
Mr. Hill: Tongaat has also adopted this method
for irrigation experiments and we find it labour saving.
We have modified their method to apply a type of
perforated pipe down alternate rows and thus have
almost halved the cost, with all the rows still getting
water.

286 Proceedings of The South African Sugar Technologists' Association—March 1966
SOME FACTORS AFFECTING THE TRANSLOCATION OF
RADIOACTIVE PARAQUAT IN CYPERUS SPECIES
Introduction
Paraquat (1, 1'-dimethyl-4,4'- bipyridilium dichloride
or dimethyl sulphate) has been used successfully
on an extensive scale for the post-emergent control of
Cyperus rotundus L. and Cyperus esculentus L. in
sugarcane. It is primarily a contact herbicide and has
no residual effect in the soil. Regrowth of Cyperus spp.
(watergrass) after scorching of the foliage always
occurs, the rate and extent depending on the stage of
growth and conditions under which paraquat is
applied. Experiments in trays have indicated that best
control of watergrass, together with maximum reduction
of tubers and roots can be achieved by spraying
with paraquat about 3 weeks after emergence
(S.A.S.A. 1964, 1965). This coincides with the beginning
of the flowering stage.
Since regrowth originates from plant parts which
escape contact with the paraquat spray, namely the
tubers, rhizomes, basal bulbs and portions of the
shoots still below ground level, spraying under conditions
favouring maximum translocation of paraquat
to these parts would be of advantage. Research with
this objective was undertaken using radioactive
paraquat.
An investigation into the starch reserves of tubers
at different stages of growth was also undertaken,
as it is was thought that this might have a bearing
on recovering from the effects of paraquat. (Thakur
and Negi 1954).
Experimental details
General
Watergrass tubers were planted singly in 5 x 5
inch earthenware or plastic pots filled with a Clansthai
sand. The pots were kept in the open, watered
daily, and the plants were subsequently treated with
radioactive paraquat.
Radioactive paraquat labelled with Carbon-14 in
the methyl radical was received in two consignments
with specific activities 2.02 and 10.1 millicuries per
millimole. Preliminary work had shown that translocation
of paraquat could be traced satisfactorily when
a watergrass plant was treated with a 0.01 ml. droplet
of aqueous paraquat solution containing 0.5 microcuries
of carbon-14. Hence, stock solutions were
prepared from the low and high specific activity consignments
containing 6330 and 1272 p.p.m. respectively.
The former was used for all except experiment
D.
The standard method of treatment was to apply the
0.01 ml. droplet to the primary tiller in each pot, one
third of this being spread over the upper surface of
by G. H. WOOD and J. M. GOSNELL
the middle inch of each of three adjacent mature
leaves. To confine the solution to the treatment site,
a band of lanolin was placed across the leaf above and
below it. No surfactant was required, as Cyperus
leaves are able to retain and spread aqueous solutions
efficiently (Ennis et al. 1952).
After treatment, the plants were either removed
directly into the open or greenhouse, or kept in the
dark for a 24-hour period prior to this. They were
harvested intact about a week after treatment by
carefully removing the soil from the roots. The treated
tiller and a few adjacent tillers and any attached
tubers were selected and kept intact for autoradiography.
These were carefully inserted between sheets
of blotting paper and dried in a plywood-hardware
cloth press at 80° C for approximately 24 hours.
After mounting the dried plants on 10 x 12 in. sheets
of white paper and covering them with Melinex film
of 6 microns thickness, they were autoradiographed
following the method developed by Yamaguchi and
Crafts (1958). Ilford Industrial G film was used, as
previous work had shown this to give entirely satisfactory
and "unambiguous autoradiographs. The
exposure time varied from 3 to 6 weeks.
Three replicates were used for each treatment together
with a control to check for pseudoautoradiography.
The treated plants were selected, for
uniformity from a large number of pots.
Treatments
Experiment A—Application of paraquat to C. rotundus
at different stages of growth
Uniform emergence had occurred by 25 days after
planting and this was taken to be the reference date.
Radioactive paraquat solution was applied at 1,2, 3,
4, 6 and 8 weeks after the reference date of emergence.
The pots were left in the dark for 24 hours after treatment
and then placed in the open for 5 days before
harvesting.
Experiment B—Application of paraquat to C. rotundus
at various times of the day
Three weeks after emergence, paraquat was applied
at the following times of the day: 6.00 a.m., 10.0 a.m..
2.00 p.m. and 6.00 p.m. Following treatment, the
plants were placed in the open. In addition, a "standard"
treatment (as used in most of the other experiments)
was included; paraquat was applied at 2.00
p.m. and the plants left in the dark for 24 hours before
being placed in the open. Harvesting took place a
week after treatment.
Experiment C—Combined application to C. esculentus
of paraquat and bromacil (5~bromo-6-methyl-3-(1methylpropyl)
uracil)

Proceedings of The South African Sugar Technologists' Association-—March 1966 287
Comparison of translocation of paraquat in C. rotundas at A: 1 week, B:4 weeks and C:6 weeks after emergence. (Autoradiographs
above, plants below).
The following treatments were applied 3 weeks
after emergence:
(i) Paraquat only
(ii) Paraquat + bromacil in the proportion 1:2
(iii) Paraquat + bromacil in the proportion 1:6
Bromacil in suspension was added to the paraquat
solution as required. The plants were placed in the
open immediately after treatment and harvested after
6 days.
Experiment D—Application of paraquat to C. rotundus
growing under various soil moisture regimes
Six weeks after emergence, the plants were removed
to a greenhouse, where the following water treatments
were applied:
(i) High — sufficient water for optimum growth.
(ii) Medium - sufficient water to maintain a
reduced rate of growth, with relatively slight
wilting.
(iii) Low—just sufficient water to keep the plants
alive.
Treatments were controlled by daily weighing and
addition of water to constant weight. Plastic pots were
used to avoid errors inherent with the use of earthenware
pots. After ten days the application of radioactive
paraquat was made. The plants were left in the
dark for 24 hours and then returned to the greenhouse,
where the same watering treatments were
carried out as before, until they were harvested a week
later.
Experiment E — Starch in C. rotundus tubers at
different stages of growth
A bulk sample of tubers was divided into 7 subsamples
each comprising 24 tubers. One subsample
was prepared for subsequent starch analysis by slicing
the tubers, drying them at 80°C for 24 hours and
grinding, while the remaining six were planted out into
trays. Harvesting was carried out at weekly intervals
from 1 to 6 weeks after planting, and the tubers were
prepared as before.
To extract the starch, I gram of the ground sample
was boiled with 40 ml. calcium chloride solution
(S.G. = 1.3). The subsequent procedure was identical
to that used for cane juice. (Wood 1962).

288
Proceedings of The South African Sugar Technologists' Association — March 1966
Comparison of translocation of paraquat in C. rotundus when applied A:at 10.00a.m., Brat 6.00p.m. and C:at 2.00 p.m. (standard
treatment). (Autoradiographs above, plants below).
Results and Discussion
The autoradiographs of both species of Cyperus
showed massive translocation of paraquat to the tip
of the treated leaf with a smaller amount moving
downwards. With the exception of the treated tiller
the difference in growth between the control and treated
plants was not large. This was due to the short
treatment time (1 week) and to the fact that only a
relatively small amount of paraquat moved out of
the treated tiller to other tillers and underground
parts, even when optimum translocation occurred.
This small quantity was unlikely to retard new growth
seriously. It must be borne in mind, however, that in
these experiments only a small proportion of the
total leaf area of a single tiller was treated. It is
reasonable to assume that the amount absorbed and
translocated by the plant as a whole would increase
with the area of leaf covered, and, hence, in the
field, where all the foliage is sprayed, the quantity of
paraquat going into certain underground parts might
well be lethal.
Accumulation of paraquat was very seldom ob­
served in the tubers of either species. This confirms
field observations that tubers from treated and untreated
plants are equally viable.
The effect of increasing age on paraquat movement in
C. rotundus
Little or no difference was found in the extent of
translocation occurring from treatment 1 to 4 weeks
after the date of emergence.
During this period, the greatest amount of translocation,
apart from that to the tips and base of the
treated leaves, was in general either to the younger
leaves or to the inflorescence of the treated tiller,
depending on the stage of growth. The older leaves
of the treated tiller, the foliage of connected tillers
and underground parts excluding the tuber accumulated
paraquat to a lesser extent.
At 6 to 8 weeks, the extent of translocation was
much reduced with little or no movement into attached
tillers and roots. However, some accumulation
was observed in the few new shoots that were present.
Maximum translocation into the inflorescence occurred
from 4 to 6 weeks. This corresponds to the

Proceedings oj The South African Sugar Technologists' Association — March 1966 289
The effect of bromacil on paraquat translocation in C.esculentus A:paraquat only, B: paraquat + bromacil (1:6). (Autoradiographs
above, plants below).
period immediately preceding maturity of the seeds.
Representative plants and corresponding autoradiographs
are shown in Plate 1.
Although the quantity of paraquat translocated per
unit area of leaf remained approximately the same
from 1 to 4 weeks after emergence, the number and
size of tillers rapidly increased and total ground
cover was achieved at about 3-4 weeks. Thus the total
amount translocated must increase rapidly during
this period. This, together with the fail-off in translocation,
after the fourth, week, is probably the main
reason for optimum control at this stage.
Increased translocation to the inflorescence from
4 to 6 weeks could have practical significance in
reducing seed production, as it is possible that paraquat
spray falling directly onto the seed may simply
scorch the seed coat without affecting its viability.
The effect of application of paraquat to C. rotundus
at different times of the clay
The autoradiographs showed that most extensive
translocation and uniform distribution in all plant
parts including the inflorescence, occurred in the
"standard" treatment where application of paraquat
was immediately followed by a period of darkness.
The extent of translocation was very similar when
paraquat was applied at 6.00 p.m. but was considerably
smaller in the case of the earlier applications
where no paraquat accumulated in the roots and new
shoots. Paraquat was deposited preferentially in the
older leaves when the plants were treated in the morning
or early afternoon, while uniform distribution
in the foliage together with good accumulation in the
inflorescence took place when paraquat was applied
shortly before dark.
These observations agree with the results of field
experiments which showed that the extent of scorching
of watergrass increased the closer to evening the
plants were sprayed (Gosnell 1965). Baldwin (1963)
also found that more efficient translocation of dipyridilium
compounds was obtained when a period of
darkness followed spraying. The reduction of paraquat
to an active radical only occurs when photosynthesis
is actively proceeding (Boon 1964). Either
absence of light or the presence of a photosynthetic
inhibitor therefore delays the scorching and allows
greater translocation to occur.
Representative plants and corresponding autoradiographs
are shown in Plate 2.

290
Proceedings of The South African Sugar Technologists' Association — March 1966
The effect of moisture stress on paraquat translocation in C. rotundas. A: low watering treatment, B: high watering treatment. (Autoradiographs
above, plants below).
A further interesting observation from the autoradiographs
is the patchy distribution of translocated
paraquat that occurred in all cases, where the plants
were exposed to light immediately after treatment.
The paraquat appears to have a tendency under these
circumstances to accumulate in isolated cells or groups
of cells. The same phenomenon was frequently
observed when paraquat was translocated to the older
less turgid leaves, irrespective of the light intensity
following treatment. Absence of light after treatment
usually resulted in more even distribution of paraquat
in the turgid or actively growing tissue into which it
moved.
The results of this experiment indicate that optimum
control of watergrass should be obtained with
evening spraying. However, no significant differences
were obtained in a field experiment comparing these
times of spraying (Gosnell 1965), and it is probable
that the increased control of watergrass was countered
by increased damage to the cane. In addition,
practical considerations, especially the effect of wind,
favour morning spraying.
The effect of bromacil on the translocation of paraquat
in C. esculentus
Bromacil, being a photosynthetic inhibitor, might be
expected to improve the translocation of paraquat
by delaying the scorch. This has in fact been observed
in field experiments, but did not occur in the work with
radioactive paraquat. The lower level of bromacil
had no noticeable effect on the translocation of the
paraquat applied with it. Slightly improved translocation
did, however, occur with the higher level.
The following plant parts are listed in the order of
decreasing paraquat accumulation in this experiment.
Tip of applied leaf >> root tips > older leaves of
treated tiller > younger leaves of treated tiller =
roots of treated tiller > roots and older leaves of
connected tillers > rhizomes and new shoots >
younger leaves of connected tillers; this was the overall
picture for all treatments.
Representative plants and corresponding autoradiographs
are shown in Plate 3.

Proceedings of The South African Sugar Technologists' Association — March 1966 291
From, the autoradiographic results it can be concluded
that in this experiment bromacil had only a
slight effect on the translocation of paraquat, even
at the highest rate of application.
The effect of water stress on paraquat translocation
in C. rotundus
The plants which were subjected to severe water
stress showed patchy accumulation of translocated
paraquat, mainly in the older leaves, with some
movement into the basal bulb. In addition, small
isolated concentrations of paraquat occurred throughout
the above ground and underground portions of
the plant system.
With adequate watering, the paraquat tended to be
distributed evenly in the leaves and other plant parts
into which it was transported. Accumulation took
place mainly in the younger leaves of the treated
tiller, with a somewhat lower amount going into the
older leaves and basal bulb of the treated tiller and
the younger leaves of attached tillers. A small amount
was deposited in the roots and rhizomes and in one
replication appreciable accumulation occurred in the
tuber. This was the only instance noted of any deposition
of paraquat in. tubers.
The translocation pattern in plants which received
the medium watering treatment fell somewhere
between the two extremes described above.
Representative plants and corresponding autoradiographs
are shown in Plate 4.
The starch content of tubers at different times after
planting
The starch contents of the samples of tubers taken
at different times after planting are listed in Table I.
The results show that initiation of growth appreciably
depleted the starch reserves in the tubers, as
demonstrated by Thakur and Negi (1954). The
lowest content was found at 4 to 5 weeks after planting,
which coincides with the period of maximum
control of watergrass obtained at 3 weeks after
emergence (S.A.S.A. 1964, 1965). It is therefore
possible that recovery from paraquat application is
related to the starch reserves in the tuber.
Conclusions
The foregoing results and discussion indicate that
translocation of paraquat is most extensive and
effective when watergrass is sprayed approximately
3 weeks after emergence and in the evening. These
findings agree well with the results of field and greenhouse
trials.
By far the greatest movement of paraquat took place
in the xylem towards the tip of the treated leaf. However,
there was generally appreciable translocation
downwards to the basal bulb, and movement also
occurred to a smaller extent to the young shoots,
rhizomes and root tips. This type of movement is
not associated with xylem transport, but rather with
phloem movement, and the evidence from these
experiments thus tends to confirm that of Thrower
et al. (1965) who found that diffusive movement
occurred out of the xylem. van Oorschot (1964) has
also found accumulation of paraquat in root tips,
believed to be due to phloem movement. This symplastic
movement probably accounts for the small
residual effect which is obtained with paraquat on
Cyperus spp.
Although bromacil in combination with paraquat
is undoubtedly useful in extending control of watergrass
for considerably longer periods than is possible
with paraquat alone, it only increased the translocation
of paraquat slightly in these experiments.
Adequate soil moisture appears to favour even
distribution of translocated paraquat and deposition
in younger rather than older leaves. The phenomenon
of patchy deposition associated with plants growing
under severe moisture stress and those which were
exposed to light after treatment, has also been frequently
observed when paraquat was deposited in
older, less turgid leaves. A possible explanation may
be the collapse of large numbers of cells when the
leaves become old or start to wilt as a result of moisture
stress, while paraquat continues to be transported
through the conducting tissues to groups of
functional cells where accumulation can still take
place.
From these studies no firm conclusion can be drawn
regarding the possibility that low starch reserves
may influence paraquat translocation, although this
appears possible.
Summary
In a series of pot experiments to study the translocation
of paraquat in Cyperus spp., this herbicide
labelled with radioactive carbon was applied at
different stages of growth, various times of the day,
alone or mixed with the photosynthetic inhibitor,
bromacil, and under various soil moisture regimes.
Autoradiographic methods were used to determine
the translocation pattern.
The autoradiographs showed that optimum paraquat
translocation occurred when the plants were
sprayed 1 to 4 weeks after emergence. When the

292
development of ground cover was taken into consideration,
the quantity of paraquat transported to underground
parts was estimated to be greatest at approximately
3-4 weeks, agreeing with the results of field
and greenhouse trials.
The extent of translocation following evening
application was similar to that of the "standard"
treatment, where 24 hours of darkness followed
application, and was much greater than when treatments
were applied at 6.00 a.m., 10.00 a.m., or 2.00
p.m.
Bromacil increased the translocation of paraquat
only slightly in these experiments.
Patchy deposition of paraquat was associated with
plants growing under moisture stress, whereas
adequate moisture favoured even distribution of the
translocated herbicide. In general, plants growing
vigorously tended to accumulated paraquat preferentially
in new growth. In only one case was accumulation
in the tubers noted.
Investigation of the starch reserves of the tubers
at the different stages of growth indicated the possibility
that recovery from paraquat application could
be related to the starch reserves of the tuber.
References
Baldwin, B. C, (1963). Translocation of diquat in plants.
Nature, 198, 872-3.
Boon, W. R., (1964). The Chemistry and mode of action of the
bipyridylium herbicides Diquat and Paraquat. Outlook
on Agriculture, IV, 163-170.
Ennis, W. B., Williamson, R. E. and Dorschner, K. P., (1952).
Studies on spray retention by leaves of different plants.
Weeds, 1, 274-286.
Gosnell; J. M., (1965). Herbicide trials in Natal Sugar Cane,
1964-65. Proc. Annual Cong. S. Afr. Sugar Tech. Ass.,
39, 171-181.
South African Sugar Association Experiment Station, Annual
Report, 1963-64, 34.
South African Sugar Association Experiment Station, Annual
Report, 1964-65, 57.
Thakur, C. and Negi, N. S., (1954). Organic reserves in relation
to eradication of nutgrass. Ind. Sci. Cong. Ass., 41, 239.
Thrower, S. L., Hallam, N. D. and Thrower, L. B., (1965).
Movement of diquat in leguminous plants. Ann.App. Biol.
55, 253.
van Oorschot, J. L. P. (1964). Personal communication.
Proceedings of The South African Sugar Technologists' Association — March 1966
Wood, G. H., (1962). Some factors influencing starch in sugarcane.
Proc. Annual Cong. S. Afr. Sugar Tech. Ass., 36,
123-135.
Yamaguchi, S. and Crafts, A. S., (1958). Autoradiographic
method of studying absorption and translocation of herbicides
using C-14-labelled compounds. Hilgardia, 28.
161-191.
Mr. Gilfillan: Has Mr. Wood mixed bromacil with
paraquat? In my experience paraquat will dominate
any mixture of herbicides.
Mr. Wood: The bromacil was applied as a suspension
in the paraquat solution. Field tests to determine
the effect of bromacil on paraquat carried out by the
Agronomy Section indicated a delay in scorch and
a greater residual effect. I am not sure whether they
were applied mixed together or separately.
Dr. Thompson: In one instance they were applied
together in a ratoon crop and we did get a delayed
reaction which might have contributed to the translocation
of paraquat but in subsequent experiments
we have gained no advantage by mixing these two
herbicides.
Mr. Glover: Your summary would make it reasonable
to associate movement of paraquat with the
metabolised products in the phloem.
Cyperus seems to accummulate starch in leaves
during the day and may translocate it as sugar at
night so that the paraquat may be flowing in linkage
with the sugar down the transportation stream towards
the sink. The meristematic tissues will have a
high demand and therefore you will get a very fast
shift particularly to active cells which you have
clearly described. I would like to see if you are collecting
more starch during the day than you are discharging
sugar to find out if the light inhibiting and
dark promoting reactions are associated with the
carbohydrates.
Mr. Alexander: Did Mr. Wood experience any difficulty
owing to radiation due to the presence of potassium
in the plants in establishing where translocation
of the paraquat had occurred?
Mr. Wood: Preliminary work showed that the
amount of potassium in the leaves is not sufficient to
produce any image on the X-ray film.

Proceedings of The South. African Sugar Technologists' Association — March 1966 293
SUMMARY OF AGRICULTURAL DATA: SUGARCANE
CROP 1965
Introduction
The last Annual Summary of Agricultural Data for
the Sugarcane Crop 1958-59 appeared in the Proceedings
of the Thirty-fifth Annual Congress of the South
African Sugar Technologists' Association of 1961.
The data discussed were very largely obtained from
the Special Census of Sugarcane Plantations 1958-59,
supplemented by information obtained from Survey
of Cane Production by the Sugar Industry Central
Board. Indeed the Government Census of Sugarcane
Plantations was for very many years the only source
of information available for this summary. In recent
years, surveys by the Central Board have gradually
assumed greater importance as a source of information.
They have enabled us to calculate the yield for
the various groups which make up the industry including
European growers, European Miller-cum-Planters,
Indian growers and Bantu growers. They also
provided data on acreage under irrigation but unfortunately
no information has been obtained on the use
of varieties, the proportion of plant and ratoon crops,
or the age of the crop at harvest. These items had to
be extracted from returns from the government census
and as a result the information was somewhat dated
by the time it became available to the industry. The
two sources of information were supplementary but
also a little confusing where they dealt with the same
data, e.g. total yield per acre and where the results did
not always agree.
During 1965 it was decided to extend the scope of
the Central Board surveys so that they would provide
not only all the data needed for this report but also
pertinent information for use by the Mechanisation
Committee. In order to expedite returns and at the
same time ensure greater accuracy enumerators were
appointed to collect data and to assist where necessary.
As a result of this action we now have for use
relevant agricultural data obtained by the industry
itself for its own use. This means that the data can be
processed quickly and while still of interest to the
industry.
Total Areas and Yields
The Sugar Industry Central Board Survey of Cane
Production 1964 to 1967 CB 46/19 shows on the 1st
May, 1965, an estimated 833,328 acres were under cane,
376,075 acres of which were scheduled to be cut during
the 1965/66 season. The exceptionally low percentage
area to be cut reflects the expansion that is now in
progress, but it must also be borne in mind that area
under cane includes fallow land which thus inflates
the area under cane. The following figures reflect the
actual situation during the two years before 1965 and
provide estimates for two years hence.
By J. L. DU TOIT and M. G. MURDOCH
These figures reveal clearly the tremendous expansion
within the industry during the years 1964 and 1965
and the marked manner in which the corresponding
percentage area harvested was depressed. It is interesting
to note further that during period 1966 to 1967,
and here only estimates are now available, the expansion
is likely to level off while the percentage area
harvested is expected to reach unprecedented levels
which indicate, apart from the necessary readjustment
after a period of expansion, an apparent expected fall
in the age of the cane at harvest.
At the time when this survey was conducted, the
duration and the effect of the drought which was then
only beginning to be felt, could not be foreseen and
consequently the estimated yield for the 1965/66
season was unrealistic and will not be considered here.
Neither would it serve a useful purpose to enumerate
estimated yields for future years. However, when it
comes to yields obtained in the past there is no reason
to doubt the accuracy of the average figure as revealed
in this and earlier surveys. The average yield for the
industry was 35.8 tons cane per acre for the 1964/65
season and 36.6 tons cane per acre for the 1963/64
season.
Rainfall and Yield
In South Africa the yield of cane is of course
extremely sensitive to rainfall and rainfall distribution.
The following table provides a comparison of yields
of cane as recorded in Central Board surveys and
average rainfall data as compiled by the Experiment
Station from 54 centres scattered throughout the sugar
belt.

294
The rainfall for the year ending the 31st May, 1965,
was only 29.02 inches compared with a 41 years'
computed mean of 38.20 inches. The drought was in
fact very much worse than is suggested by these annual
figures. Thus it is known that some 75 per cent of our
annual cane growth takes place during the six months
from November to April and during this period only
14.68 inches of rain fell compared with a mean figure
of 26.05. Evaporation at this time was very high being
34.56 inches compared with a mean of 28.92 inches.
This disastrous drought must depress the yield for the
1965/66 season very greatly.
Irrigation
Of the 833,328 acres under cane on the 1 st May,
1965,98,754 acres or 11.9 per cent are under irrigation.
Although the area under irrigation has expanded
considerably during recent years, the percentage area
irrigated has not altered a great deal.
The estimates here given do not include the sugar
areas of the Eastern Transvaal which are now being
developed and which will be entirely under irrigation.
They do, however, cover areas such as Pongola where
cane cannot be grown without irrigation and also
areas where supplementary irrigation is practiced. The
following table indicates the main areas being irrigated
during 1965.
Burning and Trashing
Although the areas of cane burned or trashed are
not readily available from this survey, growers did
indicate to what extent trashing was practised on
individual farms and for the purpose of this report a
grower who practises trashing to the extent of more
than 50 per cent will be classified as one who trashes.
Thus it was found that for the industry as a whole,
Proceedings of The South African Sugar Technologists' Association — March 1966
858 growers out of 1,535 stated that they burned or
intended to burn crops which were to be ratooned.
The remaining 678 trashed these crops. We can therefore
conclude that 56 per cent of the individuals burned
crops which were to be further ratooned while 44 per
cent trashed these crops. These figures include between
250 and 300 returns from Jaagbaan and Union Cooperative
growers whose cane is mostly in the plant
crop stage. If we subtract these from the returns, the
percentage of growers practising burning falls to 47
per cent. It is perhaps reasonable to conclude that
about half the growers burn and half trash crops
which are to be ratooned.
When it comes to crops to be ploughed out after
cutting, i.e. not intended for further ratooning, burning
is far more general. Here some 85 per cent of the
returns showed that burning is practised. In fact
burning of crops to be ploughed out predominates in
all major subdivisions of the industry.
There is, however, no such general practice for crops
intended for further ratooning. In the coastal areas of
the North Coast, only 14 per cent of growers indicated
that they burned, whereas in the Jaagbaan and Union
Co-op. areas over 96 per cent of the growers intend
burning crops for further ratooning. The following
table gives some pertinent data for these crops.
Percentage burned
before ratooning
Whole Industry 47-56
Jaagbaan-Union Co-op. . . 96
Pongola 90
Umfolozi 82
Nkwaleni 85
North Coast (Coastal below
1,000 ft.) 14
North Coast (Inland above
1,000 ft. but excluding Jaagbaan-Union
Co-op. . . 21
Zululand (Excluding Umfolozi
and Nkwaleni) . . . . 29
South Coast (Coastal below
1,000 ft. excluding Umzimkulu)
28
South Coast (Umzimkulu) . . 49
It is clear therefore that there are very large differences
in the treatment of crops which are to be
ratooned and as a rule there are sound reasons for
trashing or burning. Burning is obviously preferred in
areas of low temperatures and where there may be a
frost hazard: Jaagbaan-Union Co-op. and Pongola.
It is also commonly practised in areas regularly
irrigated: Pongola and Nkwaleni. Similarly burning
is preferred on the Umfolozi flats because of the
existence of a high water table and trash can also be
troublesome in case of floods. On the other hand,
where temperatures are not so critical but moisture
conservation is, a trash blanket can be most valuable
and consequently trashing predominates on the coastal
non-irrigated areas of Zululand, the North Coast and

Proceedings of The South African Sugar Technologists' Association — March 1966 295
on the South Coast. Even on inland areas of the
North Coast (excluding Jaagbaan-Union Co-op.)
and in Ntumeni in Zululand, trashing predominates.
The fact that Umzimkulu practises more burning than
is normal at that altitude, is probably due to low
temperatures prevailing at this, the southern extremity
of the sugar belt.
Fertiliser Usage
The survey reveals that the sugar industry used
approximately a quarter of a million tons of fertiliser
for the year ending 30th April, 1965, the average
amount of fertiliser applied per acre under cane being
700 lbs. Where the approximate age of the crop at
harvest is 18 months it means that an individual crop
receives on average rather more than a 1,000 lbs. of
fertiliser per acre.
The following table reflects some details of fertiliser
usage in the industry and the yield of ratoon cane per
acre per month is also given.
The term "Midlands" refers here to the Jaagbaan
and Union Co-operative growers, as well as growers
delivering cane to Illovo and Mount Edgecombe mills
but where the cane is grown at an altitude higher than
1,000 feet. This area therefore covers some inland
areas of both the South and the North Coast and has
a lot in common in terms of temperatures, rainfall, etc.
In fact this area appears in many respects quite distinct
(e.g. varieties, burning and trashing) from other high
altitude areas such as Upper Tongaat or Ntumeni.
The table reveals that the most productive area
Pongola also uses the largest quantity of fertiliser per
area under cane although not per unit of cane production.
The relatively high yield is of course mainly
the result of irrigation and favourable temperatures.
The relatively low amount of fertiliser used in the
Zululand area is partly the result of understandably
low applications on the fertile alluvial flats of Umfolozi,
and the somewhat more surprisingly, low applications
at Nkwaleni.
In the following table fertiliser usage and yield for
these two districts are compared with those of Ntumeni,
as well as with the production from Umzimkulu
which is the southern limit of cane production. Ntumeni
is of course a high altitude area and in the case of
Umzimkulu the coastal (below 1,000 ft.) and high
altitude (above 1,000 ft.) areas have been separated.
Data are also provided for all the areas in this high
altitude group but excluding the Midlands. It is clear
that high altitude areas are using fertiliser liberally and
getting good yields.
The survey reveals vast differences in fertiliser
practice between individuals in the same area. There
are growers who use adequate or perhaps even excessive
quantities of fertiliser but there are those also who
still use far too little fertiliser.
Although only the total amount of fertiliser used
was asked for in this survey, data obtained from the
Fertiliser Society of South Africa show that nearly 1
lb. of P was applied for each ton of cane cut during
the 1964 season and that the ratio of N:P:K was about
3.5:1:3.5.
Varieties
Trends in the changing use of varieties are probably
best studied by comparing the varieties to be ploughed
out with those being planted or in the absence of
these data comparing the varietal position of areas
under plant cane and ratoons.
In the following table the percentage area under
different varieties expressed in terms of total cane,
plant cane and ratoon cane is given for the main
divisions of the industry.
It is obvious from this table that the newer varieties
N :Co.376, N:Co.382 and N.50/211 are gaining ground
rapidly while the old varieties Co.331 and N:Co.310
have lost much of their former popularity. This,
however, is not the case with N:Co.310 at Pongola
where this variety forms 93 per cent of all plant cane
nor is it to a lesser extent in the Midlands where
Co.331 forms 13.4 per cent of all plant cane. In fact
over 80 per cent of all Co.331 as plant cane is now
found in the Midlands and this variety is now seldom
planted elsewhere. N:Co.293 continues to find favour
in the high altitude areas and particularly in the
Midlands, where it forms about a third of all the plant
cane. N:Co.339 is now disappearing fast and the
varieties N:Co.292 and N:Co.334 never really took on
in the industry. N:Co.376 is now the most widely
planted variety and seems to do well under the most
varied conditions. It seems, however, to be favoured
mainly on the South Coast where it constitutes 78 per
cent of the area under plant cane.

296
* Above 1,000 ft. above sea level but excluding Midlands.
Proceedings of The South African Sugar Technologists' Association -•- March 1966
PERCENTAGE AREA UNDER
The following table based on percentage area of the main varieties established as plant cane shows how
far the Midlands differ from the High Altitude area and Pongola from the Low Ahitude area.
Although it is tempting to analyse the yields obtained
from different varieties in these regions, conclusions
so drawn are likely to be of little value since the main
varieties grown are bound to be rather similar in yield
and exceptional yields obtained on small areas must
necessarily be discarded as unrepresentative.
Plant Cane, Ratoons and Age of Cane
The area under plant cane on the 1st May, 1965,
forms an exceptionally high proportion of the total
area under cane. This is the result of expansion and
bears little or no relation to the number of ratoon
crops grown. The expansion was of course greatest in
the Midlands area where on the 1st May, 1965, no
less than 72.4 per cent of all land under cane was under
plant cane. During the year ending 30th April, 1965, the
same area however harvested 19.8 percent of their cane
areas as plant cane and 80.2 per cent as ratoons. The
comparable results for the whole industry are as
follows: 38.7 per cent of the area was under plant
cane on the 1st May, 1965, but during the year ending
30th April, 1965, 17.1 per cent of the cane harvested
was plant cane and 82.9 per cent ratoon. Not only did
rapid expansion take place during the year under
review, but there was also a larger than usual
replanting programme.
Plant cane expressed as a percentage of the total
area under cane on 1st May, 1965.
South Coast .
Midlands .
North Coast
Zululand
Pongola .
High Altitude
Low Altitude
Industry
48.8
72.4
28.9
31.6
24.9
40.0
31.8
38.7
This survey enables us for the first lime to estimate
with a fair degree of accuracy the age of cane at
harvest. The average for the whole industry for both
plant cane and ratoons for the year ending 30th April,
1965, was 18 months. The data reveal a rather surprisingly
small age difference between plant cane and
ratoons. On the average plant cane outyields ratoons
by some 5 to 9 tons cane per acre per harvest or 0.3
to 0.5 tons cane per acre per month.
The following table shows the yields obtained for
plant cane and ratoons for our main areas.

.
Proceedings of The South African Sugar Technologists' Association—March 1966 297
DIFFERENT VARIETIES OF CANE
Total
P + R
0.4
0.0
0.7
0.2
0.0
0.6
0.3
0.3
N:Co.334
Plant
0.3
0.0
0.4
0.2
0.0
0.8
0.2
0.2
Rtn.
0.4
0.0
1.0
0.1
0.0
0.6
0.4
0.4
South Coast
Midlands
North Coast.
Zululand
Pongola
High Altitude
Low Altitude
Industry .
Total
P + R
62.9
25.1
46.0
27.7
6.0
52.1
37.4
36.6
N:Co.376
Plant
78.0
29.1
63.4
34.8
5.1
69.8
49.7
45.9
Rtn.
49.0
14.3
39.1
24.3
6.3
40.4
32.0
30.9
Total
P + R
2.6
16.3
6.7
8.3
0.2
3.0
7.7
8.0
N:Co.382
Plant
3.4
20.7
10.8
15.6
0.6
3.7
13.3
13.7
Rtn.
1.9
4.4
5.1
5.0
0.1
2.5
5.1
4.5
t Below 1,000 ft. above sea level but excluding Pongola.
T.C.A.
41.1
37.7
39.6
42.3
61.5
43.2
40.7
41.6
J I.ANT CAN1
Age
21.0
22.8
19.0
17.2
14.0
18.1
18.3
18.3
T.C.A.
per month
1.96
1.65
2.08
2.46
4.39
2.39
2.22
2.27
A summary of the data submitted shows that the
highest yields are obtained in the extreme North at
Pongola where the average age of the crop is only 14
months and 2.97 tons cane per acre per month were
produced during the year 1964/65. The Midlands gave
the lowest yields at 33.3 tons of cane per acre at 22.2
months or 1.50 tons cane per acre per month. With
the exception of the Midlands, an inland high altitude
area covering parts of the South Coast and the North
Coast, the yield of cane increased progressively from
South to North and the average age at which cane is
cut decreased from South to North. Although cane
grown in the high altitude area (excluding the Midlands)
is on the average older at cutting than cane
grown on the coast, there is indeed but little difference
in the average yield of cane per acre per month between
these two areas.
General Information
According to the Sugar Industry Central Board
Survey of Cane Production CB 46/19 European growers
occupied 581,977 acres out of a total of 833,328
acres under cane on the 1st May, 1965, i.e. European
growers were responsible for 69.8 per cent of the area
under cane. Their yield 37.8 tons cane per acre was
somewhat higher than the average for the industry,
35.5 tons per acre, and European growers produced
Total
P+R
1.8
1.1
8.0
5.9
0.6
6.6
6.0
5.2
T.C.A.
32.6
32.3
33.3
34.3
39.0
36-4
33.3
34.1
N.50/211
Plant
2.4
1.3
13.8
10.5
1.0
8.5
10.1
7.2
RATOONS
Age
20.1
22.0
18.2
17.3
14.0
19.8
17.7
17.9
Rtn.
1.2
0.5
5.8
3.9
0.4
5.1
4.2
3.9
Total
P+R
0.4
0.2
1.1
0.8
0.3
1.9
0.6
0.7
T.C.A.
per month
1.62
1.47
1.83
1.98
2.79
1.84
1.88
1.91
Others
Plant
0.1
0.0
0.9
0.6
0.3
0.4
0.6
0.4
Rtn.
0.1
0.4
0.4
0.1
0.3
0.7
0.2
0.2
ALL CANE
(PLANT AND RATOON)
T.C.A. Age per month
34.0
33.3
34.4
35.7
41.6
37.7
34.6
35.4
20.2
22.2
18.3
17.3
14.0
19.5
17.8
18.0
1.68
1.50
1.88
2.06
2.97
1.93
1.94
1.97
69.0 per cent of the 1964/65 crop, indicating that
expansion was largest in this sector of the industry.
The following table gives the percentage areas
under cane on 1st May, 1965, for the main producing
groups and the mean yields obtained by each group
for the year 1964/1965.
European growers
Miller-cum-Planter
Indian growers . .
Bantu growers . .
Industry . . . .
Per cent of area
under cane on
1st May, 1965
69.8
17.9
8.8
3.5
100.0
Summary
Yield T.C.A.
1964/65 Season
37.8
36.3
25.6
19.0
35.8
This report deals with data from the Sugar Industry
Central Board Survey of Cane Production CB 46/19
and shows that on the 1st May, 1965, there were
833,328 acres under cane of which 376,075 acres were
expected to be cut during the 1965/66 season. This
exceptionally low percentage area to be harvested

298 Proceedings of The South African Sugar Technologists' Association — March 1966
reflects the rapid expansion within the industry during
the preceding year.
Rainfall has a pronounced effect on yield and the
disastrous drought during the summer of 1964/65 will
severely reduce the 1965/66 crop but this survey was
made too early to venture an estimate of the crop.
On the 1st May, 1965, there were 98,754 acres
under irrigation in the industry, i.e. about 12 per cent
of the total area under cane.
Burning is generally practised where crops are to
be ploughed out but when crops are to be ratooned
trashing predominates particularly along the main
coastal belt. Even for the latter crops, however,
burning is preferred where it is very cold, where
irrigation is practised, or where there is a high water
table.
Fertiliser usage within the industry was heavy during
the year under -review and amounted to approximately
1,000 lbs. per acre per crop, or an estimated
total of about a quarter of a million tons for the
whole industry.
N:Co.376 is by far the most popular variety being
planted at present and on the South Coast 78 per cent
of all plant cane is N:Co.376. At Pongola, however,
N:Co.310 still constitutes 93 per cent of plant cane
and in the Midlands N:Co.293 remains the most
popular variety with 32.7 per cent of all plant cane.
Both N:Co.382 and N.50 211 are increasing rapidly
in popularity.
The average age at cutting for the season 1964/65
was 18 months with but a small difference in age
between plant and ratoon cane. The average yield of
cane per acre per month was 2 tons, being highest at
Pongola at 2.97 and lowest in the Midlands at 1.50
tons cane per acre per month.
Less than 18 per cent of the area under cane belongs
to miller-cum-planter companies while private European
growers account for more than 69 per cent.
Where the average industrial yield for the season was
35.8 tons cane per acre, that of Bantu growers averaged
only 19 and Indian growers 25.6 tons cane per
acre.
Dr. Cleasby (in the chair): It surprised me that the
average age of cutting in the 1965 season was eighteen
months.
Mr. Wyatt: Can Mr. du Toit explain the very large
drop in yield from plant cane to ratoon cane for
Pongola.
Mr. du Toit: We were also very surprised and
checked our figures thoroughly so unless the returns
from the farmers in that somewhat small area were
incorrect, which is possible, attention should be paid
to decreasing the number of ratoons to avoid this
drop in yield.
Mr. King: The table for fertiliser usage on page
3 shows the South Coast and Midlands using more
than the North Coast and Zululand but this is almost
certainly due to the present high proportion of plant
cane in those first two areas. When conditions return
to normal I think all these areas will be using between
600 and 700 lbs. per acre of fertiliser.
Mr. Perk: It is noticeable that the sucrose content
of the cane started higher than usual due to premature
ripening but then failed to increase. Were maturity
tests carried out to determine if the cane was ripe in
June and July?
Mr. du Toit: Tests were not carried out.

Proceedings of The South African Sugar Technologists' Association—March 1966
SOME FACTORS INFLUENCING SUCROSE % CANE
AT HIPPO VALLEY ESTATES
Introduction
Sugar cane was first planted at Hippo Valley in
1959, and the first crushing season started in 1962.
The period 1962/64 may be considered as a rapid
development stage, the main effort being concentrated
on the expansion of cane sections to cope with the
anticipated crushing capacity of 330 tons cane per
hour programmed for the 1966 season. In consequence,
little attention could be paid to the finer points of
control, and cutting orders were issued on the basis
of cane age and variety.
During the 1964 season difficulty was experienced
with fluctuations in sucrose per cent cane. Hence for
the 1965 season it was decided to introduce a system
of maturity testing over the whole Estate, combined
with moisture determinations and drying-off routines,
in an attempt to obtain a more uniform quality of
cane at the gantry, and thereby higher yields of sucrose
per acre.
by C. A. JOHNSON
299
Figure I shows the sucrose per cent cane for the
two seasons, 1964 and 1965. Climatically the seasons
were comparable except for a wet February in 1964,
which probably had an adverse effect on cane quality
duiing the early stages of that season. It will be seen
that a much better overall sucrose content of cane was
obtained in 1965, but there still existed periodic
fluctuations in sucrose per cent cane which will be
dealt with in a section of this paper.
Method
To determine the degree of ripeness of the cane,
representative samples were taken section by section
and field by field, at regular intervals, for analysis.
The preliminary results permitted the fields to be
placed in categories, irrespective of age or variety.
Promising fields were then sampled at more frequent
intervals, and before the crushing season started a
cutting programme was prepared based on sucrose

300 Proceedings of The South African Sugar Technologists' Association—March 1966
per cent cane and juice purity. This maturity testing
continued throughout the season, a total of some
3,000 laboratory determinations being made at the
rate of 12 to 15 samples per day.
Date of
Sample
2/3
29/3
5/4
12/4
16/3
31/3
7/4
15/4
Table I
Pre-season Maturity Tests
Field
No.
7B
3C
Variety
310
310
Age
Months
11.00
11.75
12.00
12.25
10.00
10.50
10.75
11.00
Purity
/o
75.0
77.1
83.8
85.4
81.2
84.2
83.8
84.8
Sucrose
% Cane
9.7
11.9
12.7
14.7
11.2
12.0
12.6
13.8
Maturity Testing
A random sample of 20 stalks per field was cut early
in the morning. If the field contained arrowed or
lodged cane, a representative proportion of these
canes was included in the sample. After dividing the
canes into top, middle and bottom thirds, a composite
third was processed in a "Jeffco" shredder. Taking
a pre-calculated weight of 327 grams of shredded
cane, blending for 10 minutes at 8,000 r.p.m., with
1,000 mis. water and 10 mis. of a 5 per cent solution
of Sodium carbonate, after normal clarification, the
direct polariscope reading gave the sucrose per cent
cane. The remaining two-thirds of the sample was
crushed in a small mill, and brix, pol and purity were
determined in the normal manner.
Tables I and II show typical progressive readings
for individual fields before and after the start of the
crushing season.
Date
Sample
20/5
27/5
1/7
17/7
25/6
16/7
22/7
28/5
16/7
13/8
23/7
6/8
13/8
1/7]
26/8
2/9
10'9
18/9
1/10
Field
No.
5C(R)
IrrigationIrrigationHarvested
6D(P)
7E(R)
2C(R)
1E(R)
8A(R)
Table II
Maturity Tests
Variety
376
ReducedStopped
376
376
310
310
310
Age
Months
12.50
12.70
14.00
14.50
16.00
16.75
17.00
10.00
11.75
12.50
12.00
12.50
12.75
12.00
13.75
14.00
12.50
12.75
13.25
Purity
/o
82.4
85.5
86.5
94.1
89.0
89.9
90.1
85.6
89.6
90.1
90.0
90.8
91.0
88.7
90.4
90.0
89.3
89.5
92.5
Sucrose
% Cane
12.1
12.3
13.6
16.1
14.6
15.2
15.9
12.9
14.2
15.0
15.3
16.3
16.5
14.9
16.3
16.5
15.4
16.0
17.0
Drying-off
From the preliminary maturity tests, selected promising
fields were programmed for harvesting.
Depending upon the time of year, the system of irrigation
(overhead or furrow), the soil type, and the
climate, a tentative reduction in frequency of irrigation
was made. This was followed by further maturity
tests, and as the ripening proceeded a decision was

Proceedings of The South African Sugar Technologists' Association—March 1966 301
taken on when the last irrigation should be applied.
Water was withheld entirely for periods of two to
six weeks prior to harvesting. Table III shows some
typical drying-off routines.
To assist in this timing of drying-off, moisture
determinations were made using the 4th/5th intemode.
Representative samples of six tops were cut, as far as
possible from stalks of identical stages of top growth.
The 4th/5th intemode tissue consists of the section of
the stalk between the intercalary meristem immediately
above the 4th and 6th leaf attachments. The
samples were quartered, weighed, oven-dryed at
85° C and a direct moisture determination made.
With ample water at the time of maximum growth,
the moisture content is approximately 92 per cent.
With drying-off the moisture content falls slowly to
90 per cent, after which dehydration becomes rapid
as irrigation is withheld. Table IV shows some typical
moisture determinations, together with the sucrose
per cent cane of the samples.
Soil
Type
Sandy Clay
Loam.
Derived from
Paragneiss
Heavy Active
Clay.
Derived from
Basaltic Larvas
Average
depth of
soil in
feet
2.0
1.5
Available
moisture
per foot
of soil in
inches
1.7
3.0
May
.092
.092
Table HI
Drying-off routines: Furrow Irrigation
Pan
Factor
Oct.
.218
.218
Temperature Effects
There is no doubt that cutting on a maturity basis
has helped to eliminate some of the big fluctuations
in sucrose per cent cane, but a number of smaller
variations are still to be seen in Figure 1. Analysis
of temperature data has shown, for the past two seasons,
a strong correlation between sucrose per cent
cane and diurnal temperature range one calendar
month before harvesting.
Figures 2 and 3 show the positions for the 1964
and 1965 seasons respectively.
If the temperature graph is plotted one month in
advance, it is possible to predict a drop in sucrose
per cent cane, and this was done on five occasions
this last season. With the temperature range recorded
in advance and the field maturity figures,
amendments will be made to the cutting orders next
season as follows:
Normal irrigation cycle
at 1.8 in./application
May
14 days
19 days
Oct.
6 days
8 days
May
1.8 in. — 14 days
1.8 in. — 21 days
Nil — 28 days
1.8 in. — 19 days
1.8 in.— 28 days
Nil — 28 days
Drying-off cycle
Oct.
1.8 in. — 6 days
1.8 in. — 12 days
Nil — 14 days
1.8 in. — 8 days
1.8 in. — 14 days
Nil —21 days

302
For periods where the diurnal range is less than
25° F, cutting will be switched to suitable mature
fields of N:Co.310 of known high sucrose.
For periods where the diurnal range in greater than
30° F, cutting will be switched to mature fields of
N:Co.376.
At temperature ranges between 25° F and 30° F
cutting of both varieties will continue on a normal
maturity basis.
In this way it is hoped to be able to maintain cane
at the gantry with a more constant sucrose per cent
cane, which in turn will be reflected in smoother
running of the factory, and a higher overall recovery.
6C
7A
4A
Sample
2A
Table IV
4th/5th Internode Moisture Determinations
a
b
c
d
e
f
a
b
c
d
e
f
a
b
c
d
e
f
a
b
c
d
Moisture
0/
/o
88.8
89.2
88.3
89.3
88.8
90.7
90.0
88.6
88.9
90.0
87.5
87.5
88.0
88.0
84.6
90.0
91.7
86.7
83,3
82.3
81.0
80.9
Results and Discussion
Average Moisture percentage
and sucrose %
cane
89.0% 13.66%
88.2% 13.39%
87.8% 15.06%
81.4% 17.06%
Due to the many variable factors involved, it is
always difficult to assess the benefits of one individual
practice in agriculture, but there is no doubt that by
careful routine analysis and correct drying-off, the
yield of sugar per acre can be raised. Correcting for
mill efficiency and overall recovery, there has been an
8 per cent increase in sugar production from the same
acreage of cane, with slightly lower yields of cane per
acre in the last season compared with the 1964
results. How much of this can be attributed to the
above exercise is difficult to say.
Conclusions
To obtain maximum yields of sugar per acre per
month, correct drying-off is essential. Maturity testing
and plant moisture determinations play an essential
part in the drying-off routine.
Proceedings of The South African Sugar Technologists' Association—March 1966
Summary
Routine maturity testing for sucrose per cent cane
continued throughout the season. Stalk moisture
determinations were used in conjunction with the
maturity figures in planning drying-off cycles. A
correlation was found between diurnal temperature
range and sucrose per cent cane.
Acknowledgments
The author wishes to thank Mr. I. Bales-Smith,
Field Manager, Hippo Valley Estates, for close cooperation
and patience in dealing with crop sampling
and many alterations in the cutting programme; also
my colleague Mr. P. Koenig, Agronomist, Hippo
Valley Estates, for handling the day-to-day analyses
and crop logging.
References
Tanimoto, J. (1961). 4-5 Joints as indicators of moisture
tension of the sugarcane plant. Proc. 20th Ann. Meeting
Haw. Sugar Tech. pp. 265-274.
Johnson, C.A. (1966). Sucrose % cane and diurnal temperature
range, Hippo Valley Estates. Int. Sugar J. (in press).
Mr. du Toit: The sucrose percent cane is low at the
start of the season, rises in the middle of the season
and then falls again. Mr. Johnson says they start the
season by cutting N:Co 310, which has a consistently
high sucrose content.
But it is difficult to decide when to cut a variety
which starts with a fairly poor sucrose content and
never increases appreciably.
It is very important to establish correct drying-off
times and the pattern of ripening of the different
varieties.
Mr. Johnson: Figure I indicates exactly what happened
in the 1964 and 1965 seasons.
There were carry-overs of cane in both seasons but
in 1964 it was all milled in the first month and caused
the big drop in sucrose. In 1965 the carry-over was
spread more evenly throughout the season, which
was started with N:Co 310, then N:Co 376 and back
toN:Co310.
Mr. Andries: I do not see how you can phase
varieties unless you cut at 12 or 24 months, If you cut
at 14 months you will be out of phase with that
variety the following season.
I agree with Mr. du Toit that you will get your best
average sucrose by cutting the high sucrose varieties
at their optimum periods.
Mr. Johnson: Last year we cut 8,000 acres of plant
cane at 14 months but this year we will cut 17,000
acres mostly at 12 months so phasing will not be
difficult.
I would like to mention that by maturity testing
and drying off it is quite possible to induce both higher
sucrose and higher purity in 10 months old cane.

Proceedings of The South African Sugar Technologists' Association—March 1966 303
Mr. Andries: In the Big Bend area of Swaziland it
would be quite impossible to cut cane at 9 months,
particularly if it was well topped.
I think it would only be feasible in our case to ripen
the cane between 9 and 12 months if it was all harvested
between August and October.
Mr. Johnson: It is in September that we have the
widest diurnal temperature range which should have
an effect on sucrose in October and November but
cloud effect must also be considered.
Mr. Glover: I think that temperature is also important
because when the temperature drops at night
growth slows down a lot and sugar demand for growth
may not be as high as normal. If a bright day follows,
considerable sucrose is made which may not again be
distributed for growth on a succeeding cold night
and so the sucrose will build up in store.

304 Proceedings of The South African Sugar Technologists'' Association—March 1966
THE RESULTS OF HERBICIDE SCREENING TRIALS
IN SUGARCANE DURING 1965
Introduction
Herbicide screening trials for sugarcane lands have
been conducted in Natal each season since 1961/62,
and these have led to the gradual development of an
almost complete system of chemical weed control for
commercial plant cane. The efficacy of 2, 4-D formulations
in controlling a wide spectrum of weeds when
used as a pre-emergent spray had long been recognised,
but the persistence of watergrass (Cyperus esculentus
and Cyperus rotundus) from tubers severely limited
the reliance which farmers could place on herbicides.
Paraquat, used as a post-emergent contact spray,
gave excellent temporary control of watergrass
(Thompson and Gosnell, 1963) when applied three or
four weeks after emergence. If the herbicide was used
at a sufficiently early stage of crop development,
damage to the sugarcane could practically be avoided
(Gosnell and Thompson, 1965).
When soil moisture conditions were favourable,
and particularly where the cane stools were well
developed with leaves openly exposed over the weed
growth, the use of diuron with surfactant gave excellent
control of a number of grasses as well as Cyperus
esculentus, and could be recommended as a substitute
for paraquat (Gosnell and Thompson, 1964). The
extremely variable rainfall in Natal, however, limited
the general applicability of substituted urea com­
Treatment
by J. M. GOSNELL and G. D. THOMPSON
Table I
Pre-emergent treatments compared in Experiments I and II
Trifluralin
Trifluralin
Trifluralin
Norea
Norea
7175
7175
Norea + 7175 . . . .
DCPA
DCPA
DCPA + 2, 4-D amine . .
DCPA -I- 2, 4-D amine . .
2, 3, 6-TBA + MCPA . .
Fenac
Fenac
Atrazine 80
2, 4-D glycol ester.
2, 4-D glycol ester + silvex .
No Weeding
Hand Weeding . . . .
Lb. a.i. per
full acre
1.0
3.0
6.0
3.2
4.8
2.0
3.5
3.2 + 2.0
4.5
9.0
4.5 + 3.75
9.0 + 3.75
1.44 + 4.5
2.0
4.0
3.2
3.0
2.0 + 1.0
—
pounds. The uracils, and particularly bromacil,
were found to be more effective than the substituted
ureas over a much wider range of soil moisture conditions
(Gosnell, 1965). The phytotoxicity of these
compounds to sugarcane was acknowledged to be
an important limitation.
The results of three post-emergence experiments
harvested during 1965 have served to elucidate the
effects of bromacil on sugarcane yields and also to
indicate the considerable value of uracil-substituted
urea combinations. In two pre-emergence experiments
a wide range of herbicides were compared when
applied shortly after planting. In all of these experiments
the herbicides were applied on an 18-in.c.h swath
over the cane row only, interrow weed control being
effected either by tractor-mounted or mule-drawn
cultivators. Cane and sucrose yields are expressed
throughout in short tons per acre.
Pre-emergence Experiments, I and II
Description
These two trials were identical in design. Each
comprised three replications of a 4 X 5 rectangular
lattice. The treatments are shown in Table I and
details of the herbicide formulations are given in
Appendix I.
Quantity of product
per acre,
row only
i Pt-
2 Pt.
4 Pt.
I] Lb.
2 Lb.
1} Lb.
2j Lb.
1> Lb. + liLb.
2 Lb.
4 Lb.
2 Lb. + 2 Pt.
4 Lb. + 2 Pt.
8 Pt.
3 Pt.
6 Pt.
liLb.
1+ Pt.
1 Pt. + If Pt.
—
Cost of
herbicide'ac.
R
—
—
—
—
—
—
3.00
3.39
6.78
2.93
0.78
2.39
—

Proceedings of The South African Sugar Technologists" Association — March 1966
Experiment I was located on a Williamson sandy
loam at the Chaka's Kraal Experimental Farm. It
was planted in October, 1964 and harvested at thirteen
months of age in November, 1965. Supplementary
overhead irrigation was applied until February, 1965,
when the local river ran dry, and was resumed in
June, 1965. The major weed species were Cvperus
esculentus, Portulaca oleracea, Digitaria achcendens
and Eleusine inclica.
Experiment II was planted on the Mtunzini Propagation
Farm on a Rosehill sandy loam in September,
1964, and harvested in October, 1965 when the cane
was 13 months old. The experiment was not irrigated,
but rainfall records and gravimetric soil moisture
Soil series
Depth
Per cent Clay
Per cent Fine sand
Per cent Coarse sand
Per cent Organic matter
Moisture content at 100 nib
Moisture content at 15 bars
Table II
305
determinations indicated that conditions were ideal
for soil-applied herbicides during the first three months
after' planting. The heavy rains which fell during this
period caused water-logging to occur in a portion of
the experiment, and the cane growth in Replication II
was so poor that the results from these plots were
omitted from the experimental data. The main weed
species were Cyperus esculentus, Paspalum distichum
(dominant in the poorly drained areas), and Digitaria
zeyheri (see Plate I).
The mechanical analyses and moisture retention
characteristics of the soils in Experiments I and II
are shown in Table II.
Mechanical analyses and moisture retention characteristics of soils on the experiment sites
0-3 in.
19
14
46
22
2.44
26.4
12.6
5.9
Williamson
Results
Mean yield data from Experiments I and II are
3-6in.
18
12
50
19
2.72
27.6
13.6
5.4
Table III
(Win.
15
12
50
23
2.02
22.3
T5.5
5.1
Rosehill
3-6in.
16
12
46
26
2.09
23.4
12.1
5.8
0-3 in.
8
i
26
62
1.12
8.2
3.1
2.8
Fernwood
3-6in.
8
3
33
56
1.12
7.6
4.9
2.8
presented in Table III, together with the cane yields
from the individual experiments.
Mean results of Experiments I and II and cane yield data from the individual experiments

306
Discussion
Visual scorings of weeds were carried out in these
experiments on the basis of 0 (= no weed control)
to 9 (= complete weed control). A scoring of 7 using
this system reflects adequate weed control, such that
further control is not immediately necessary.
In Experiment I, at one month after planting only
three treatments were scored at an average greater than
7. These were fenac at 3 and 6 pints per acre row only
and trifluralin at 4 pints per acre row only. However,
cane vigour scorings carried out at the same time
showed that both fenac treatments had resulted in a
slight suppression of cane growth, and the trifluralin
treatment had almost eliminated cane growth. The
plots treated with 7175, which gave cane yields com-
The spectrum of weed control was very similar
for all of the effective herbicides. There was generally
good control of most broadleafed weeds and annual
grasses, with a moderate to slight effect on Cyperus
esculentus. DCPA (Dacthal) had previously given
good control of grasses but not of broad-leafed
weeds, and in these experiments it also gave good
results when used with 2, 4-D.
Some interesting effects were observed in the plots
treated with trifluralin. At Chaka's Kraal the chemical
was sprayed into the furrow before planting and
immediately covered with soil. The control of Cyperus
spp. and grasses was good at the highest rate of application,
but damage to the cane was so severe at all
rates that no millable stalks were obtained from any
of the trifluralin-treated plots. At Mtunzini trifluralin
was sprayed as a normal pre-emergent herbicide
on to the soil surface after planting and little
effect was apparent for a period of two months. During
the third month, however, a substantial improve-
Proceedings of The South African Sugar Technologists' Association — March 1966
Table IV
parable with those from hand-weeded plots, were
scored at 6.3 and 5.3 for the higher and lower
treatment levels respectively.
In Experiment II a number of treatments gave
adequate weed control after 5 weeks, but the effect
was lost after 8 weeks and weed infestation was
generally severe after 12 weeks. The progressive loss
of weed control in these treatments is shown in Table IV,
where the mean visual scorings are given. The low
cane yields obtained from the plots treated with the
high rate of fenac were the result of obvious cane
damage. Once again relatively high yields were obtained
from plots treated with 7175, without weed
control ever having been adequate on the basis of
visual scorings.
Mean visual weed control scorings and yields for effective treatments in Experiment II
Weeks after planting
Rate of product
Treatment row only
Hand Weeding —
2,3,6-TBA + MCPA 8 pt.
Norea + 7175 1± + 1+Jb.
DCPA + 2, 4-D 4 + 2 lb.
DCPA + 2, 4-D 2 + 2 lb.
Fenac 6 pt.
STorea 2 lb.
Atrazine 1+ lb.
5
8.3
8.0
7.3
7.3
7.0
7.0
7.0
Mean Scoring
8
6.7
6.3
5.7
6.3
5.3
4.0
3.7
12
4.2
3.8
3.3
1.8
4.0
1.7
1.3
Mean
cane
yield
30.9
31.1
26.7
30.8
25.0
19.1
20.0
23.3
ment in the control of Cyperus spp. and grasses was
noted, and the highest and lowest rates of application
resulted in fairly good cane yields.
The large yield reduction from 33 tons to 11 tons of
cane per acre, caused by complete lack of weeding,
could be ascribed to the reduction in stalk populations
from 62,000 per acre in hand weeded plots to
29,700 per acre in unweeded plots. Herbicide treatments
giving similar yields to hand weeding generally
had much lower stalk populations, however.
There were no consistent effects of treatment on
sucrose content.
Post Emergence Experiments, III and IV
Description
Experiments III and IV were also identical in
design, each comprising three replications of a 4 x 5
rectangular lattice with the treatments shown in
Table V. Details of the herbicide formulations are
given in Appendix I.

Proceedings of The South African Sugar Technologists' Association — March 1966 307
Table V
Post-emergent treatments compared in Experiments III and IV
Experiment III was located close to Experiment I Results
at Chaka's Kraal and Experiment IV was adjacent Mean yield data from Experiments III and IV are
to Experiment II at Mtunzini. The conditions in these presented in Table VI, together with the cane yields
experiments were therefore virtually identical to those from the individual experiments,
already described for Experiments I and II.
Table VI
Mean results of Experiments III and IV, and can yield data from the individual experiments

308
Discussion
The outstanding post-emergent herbicide in these
trials was bromacil (see Plates I and II), used either
alone or in combination with other chemicals. The
prolonged period of good weed control obtained from
the various treatments is illustrated by the weed
control scorings shown in Table VII. Cane vigour ratings
were also carried out on the basis of 0 (= complete
suppression of cane) to 9 (--- no effect of herbi­
Proceedings of The South African Sugar Technologists Association — March 1966
Table VII
cide on cane). The results, also presented in Table VIII
show that a combination of diuron and bromacit
can be used to avoid most of the phytotoxicity whilst
still effecting adequate weed control. It is apparenl
that potential yields, following adequate weed contro,
four months after planting, were obtained by using
1 lb. diuron and 1/2 lb. bromacil per acre on the row
only at a cost of R4.77 per acre. A most remarkable
result, however, was the recovery of the cane sprayed
Mean visual weed contro! and cane vigour scorings in Experiments III and IV
with high rates of bromacil alone (14 and 2 lb. per
acre on the row only). Despite the low cane vigour
scorings shown in Table VII for these treatments, the
final yields were almost equal to those in hand-weeded
plots. Weed control was almost perfect with these
high rates of bromacil, and was always very good
with the mixtures of diuron and bromacil. Lower
rates of bromacil gave fairly good weed control which
was usually improved by the addition of paraquat.
Yields of cane were not appreciably affected, the
improved weed control probably being offset by
increased damage to the cane. The addition of paraquat
to the higher rates of bromacil served only to
reduce yields.
None of the remaining herbicides gave commercially
acceptable degrees of weed control during the
first three months of the experiments. It became noticeable
later, however, that cane growth in the 7175,
ametryne and prometryne treatments was surprisingly
good in spite of the abundance of weeds present.
The effects were confirmed at harvest when the plots
treated with these chemicals gave yields only slightly
below those of the hand weeded plots. Atrazine failed
as a post-emergent herbicide while dalapon applications
resulted in appreciable cane damage without
much weed control being effected.
The general relationship between yield and stalk
counts was quite marked, and it was observed that
applications of bromacil, especially at high rates,
resulted in the development of populations almost as
high as those in hand weeded plots. This was probably
due to the high degree of weed control obtained at
an early stage of crop growth, which encouraged
tillering. In contrast, treatment with 7175, ametryne
and prometryne resulted in relatively low populations
due to poor weed control in the early stages.
Treatments generally had relatively little effect on
sucrose content, but increasing levels of bromacil
tended to cause a decrease in sucrose content in Experiment
IV. However, no such trends were apparent
in Experiment III, and the mean results shown in
Table VI are not conclusive.
Post Emergence Experiment V — Herbicides on
Panicum Maximum
Description and Results
An area of plant cane on a Fernwood sand, where
Panicum maximum and Cyperus esculentus were
dominant, was selected for this experiment. Panicum
maximum, or Ubabe, is a serious weed problem in
many areas, particularly on sandy soils. It frequently
occurs in competition with Cyperus esculentus, and
experience has shown that where paraquat is used to
control Cyperus, the Panicum problem increases.
The analytical results shown in Table II indicate
that this soil was extremely poor. Cane growth was
very slow following planting in March, 1964. The
herbicides were applied in November, when there was
an average of 11 leaves per stalk, and the crop was
harvested in November, 1965. The design was a
4 x 4 partially balanced lattice, but the growth of
the cane in Replication IV was so poor and uneven
that the yield data from this replication were omitted
from the results which are presented in Table VIII.

Proceedings of The South African Sugar Technologists' Association—March 1966 309
0.1 per cent Agral surfactant used with all treatments.
Discussion
Dalapon applied at a rate of 3 lb per acre on the row
only eliminated Panicum but caused very severe cane
damage, and TCA at 9 lb also gave good control of
Panicum at the expense of the cane. The damage
caused by dalapon was visually much more severe
than that caused by TCA. The lower rates of dalapon
(2 lb) and TCA (6 lb) caused far less cane damage
but did not control Panicum adequately. The combination
of 2 lb of dalapon and 6 lb of TCA, at a cost
of R3.37 per acre, was effective and could be recommended
providing that it was applied accurately to
avoid local excesses of dalapon. Bromacil applied at a
rate of 1-} lb per acre on the row only gave good
control of Panicum without causing severe cane
damage. Although the lower rates of this chemical
did not control the grass adequately, yields were
nevertheless high.
Of the remaining herbicides, only paraquat was of
interest since it caused a reduction in cane yield due
to the profuse tillering of Panicum which followed the
scorching effect of the chemical.
Conclusions
A number of herbicides such as 2,3,6-TBA -f
MCPA, fenac, 7175, atrazine and diuron have given
better results than 2, 4-D when applied pre-emergent,
both in the current and in previously reported trials.
However the costs of these herbicides range from five
times the cost of 2, 4-D upwards. Because of soil
moisture variability under dryland conditions, present
recommendations for pre-emergent weed control
include only the various 2, 4-D and MCPA formulations.
However, under irrigated conditions, atrazine
and 2, 3, 6-TBA |- MCPA are worthy of field trial on
a limited scale, when the economics of their use can
be more accurately assessed.
The visually severe effects of bromacil on sugarcane
need not apparently result in significant suppressions
of yield. In view of the excellent weed control obtained
consistently with this compound, it is important
to know that such tolerance exists since accurate
application rates are not always achieved in the field.
However, the best results were obtained by using mixtures
of diuron and bromacil. The use of these mixtures
or of their analogues linuron and "Sinbar", will
probably be the most important development in
chemical weed control in the immediate future.
No entirely satisfactory chemical control of Panicum
maximum in cane fields has yet been found.
Where labour is available, hand-weeding is still to be
preferred; where labour is not available the following
treatments can be used, but some cane damage can
be expected:
1. A mixture of 4-5 lb dalapon and 15 lb of TCA
per acre full cover, or 2 lb of dalapon and 6 lb
of TCA on the row only.
2. Bromacil at a rate of 3-4 lb per acre full cover
or 1^ lb row only on sands, and higher rates on
heavier soils.
Summary
In two experiments where pre-emergent herbicides
were compared, the best results were obtained with
7175, 2,3,6-TBA + MCPA, DCPA + 2,4-D,
atrazine and norea. In two trials to compare postemergent
herbicides, bromacil was outstanding.
Mixtures of diuron and bromacil gave the best results
but bromacil alone at relatively high rates (l-J-2 lb
per acre on the row only) gave excellent weed control
and the cane recovered well from initial damage.
In general, the addition of paraquat to bromacil
served little purpose.
The best control of Panicum maximum was obtained
by using either a mixture of dalapon and TCA
or high rates of bromacil. Cane damage was evident,
however, in all treatments which controlled Panicum.

310 Proceedings of The South African Sugar Technologists' Association—March 1966
Effect of bromacil, at 1/2 lb./acre row only,
in Experiment IV 3 months after planting.
The plot was adjacent to the No-Weeding plot
which can be seen in the background.
Uncontrolled weeds in Experiment IV 3
months after planting.

Proceedings of The South African Sugar Technologists' Association—March 1966 311
Short Chemi cal Name
2, 3,6-TBA 4 MCPA
2, 4-D amine
2, 4-D glycol ester
'7175'
A me try ne
Atrazine
Bromacil
DCPA
Dalapon
Din ron
Fenac
Linuron
Norca
Paraquat
Prometryne
Silvex
Trifluralin
Uracil 629
Uracil 732
Uracil 733
Uracil 766
Uracil 767
Urea 1318
Appendix I
Herbicides used
Commercial product
Fisons 1815
Fernimine 5
Esteron 10-10
Hercules 7175
Ametryne 50W
Atrazine SOW
Hyvar X
Dacthal
Dowpon S
Karmex
Weedac
Afalon
Hercules 7531
Gramoxone
Prometryne 50W
Kuron
Treflan
Uracil 629
Sinbar
Uracil 733
Uracil 766
Uracil 767
Urea 1318
a.e. -- acid equivalent; a.i. ~ active ingredient
Composition
0.48 lb. a.e. + 1.5 lb.
5 lb. a.e. per gall.
4.8 lb. a.e. per gall.
50 per cent a.i.
50 per cent a.i.
80 per cent a.i.
80 per cent a.i.
75 per cent a.i.
85 per cent a.i.
80 per cent a.i.
1.85 lb. per gall.
50 per cent a.i.
80 per cent a.i.
2 lb. a.i. per gall.
50 per cent a.i.
3.33 lb. per gall.
3.33 lb. per gall.
80 per cent a.i.
80 per cent a.i.
80 per cent a.i.
80 per cent a.i.
80 per cent a.i.
80 per cent a.i.
Surfactants used
Agral 90 (non-ionic) contains an alkylated phenol-ethylene oxide condensate.
WK (non-ionic) contains the dodecylether of polyethylene glycol.
References
Gosnell, J. M. (1965). Herbicide trials in Natal Sugarcane,
1964-65. Proc. 39th Cong. S.A. Sugar Tech. Assn. pp.
171-181.
Gosnell, J. M. and Thompson, G. D. (1964). The results of
herbicide trials, 1963-64. Proc. 38th Cong. S.A. Sugar
Tech. Assn. pp. 166-174.
Gosnell, J. M. and Thompson, G. D. (1965). The effects of paraquat
on the growth and yield of sugarcane. Proc. 12th
Int. Cong. Sug. Cane Techn. Puerto Rico.
Thompson, G. D. and Gosnell, J. M. (1963). The results of herbicide
trials conducted in the cane belt of Natal, 1962-63.
Proc. 37th Cong. S.A. Sugar Tech. Assn. pp. 143-152.
Mr. Wyatt: In the last slide it was mentioned that
3 pints of paraquat per acre were applied, in line
only, and this seems rather a lot.
Dr. Thompson: The figure should have been one pint
on the row only.
Mr. Gilfillan: Why were 3 pints pre-selected instead
of 2, which is the usual application?
Dr. Thompson: Mr. Gosnell chose 3 pints full cover,
but 1 would have preferred one pint, row only.
Mr. Coignet: Is not a lot of damage done to the cane
by weeds by the time a post emergent treatment is
applied? Why not use chain harrowing?
Under drought conditions weeds of course will not
grow, but they are a big problem under irrigated
conditions.
a.e. per gall.
Dr. Thompson: We would always recommend
pre-emergent spraying after planting or burning and
harvesting a ratoon, and it is agreed that weeds which
survive this treatment should be eliminated as soon
as possible. Suitable herbicides can be used but if
the soil conditions favour the use of a chain harrow,
use it by all means.
Mr. Wardle: Is bromacil a systemic? If so, could it
not be combined in very small quantities with 2,
4-D in order to produce a more effective control of
water grass?
Dr. Thompson: Bromacil is systemic in action and
we have not used it with 2, 4-D.
Mr. Browne: We had residual damage when bromacil
was used at Muden.
Dr. Thompson: We are not sure how long bromacil
effects may persist in the soil, but Mr. Anderson tells
me that the final applications on citrus at Muden
were made only four months before the trees were
taken out and sugarcane planted. From the analyses of
the Muden soils where recent damage to cane was
done, it appears that the recommended levels of
bromacil were too high for the sandy nature of these
soils and their low organic matter contents.
Mr. Gilfillan: We have found bromacil and diuron
very effective against weeds but I must issue a warning
that they can cause stunting in old ratoons, when
used together under certain conditions.
Dr. Thompson: Damage to cane, I believe, is always
related to soil texture and organic matter content,
and these should be carefully evaluated.

312 Proceedings of The South African Sugar Technologists' Association—March 1966
WEED CONTROL ON A NEWLY DEVELOPING ESTATE
AT MAZABUKA, ZAMBIA
Introduction
In view of the high capital cost of a raw sugar factory,
it is considered essential to produce the highest
possible output right from the first season. The Tate
& Lyle Groups' new Nakambala Estate will be commencing
its first crop at full production level in 1968,
which means that 90 per cent of the acreage at that
date will be in plant cane. Labour requirements on
the plant crop are always considerably higher than on
subsequent ratbons and in order to reduce the peak
demand on labour at planting, provision is being
made for the use of herbicides for weed control on a
large scale. Herbicide trials were carried out on the
existing 300 acre trial block at Nakambala and at the
nearby Kafue Pilot Polder during the summer seasons
of maximum weed germination in 1964 and 1965.
Herbicide Trials
1. Soil Types and Weed Flora
(a) Previously cultivated Sandy Clay Loams
The most serious weed problem occurs on areas of
old arable land on Sandy Clay Loams at Nakambala
which are liable to heavy infestations of grasses, the
dominant species being Eleusine indica and Rotboellia
exaltata, together with occasional broad-leaved "dicot"
weeds.
(b) MontmoriUonitic Clays (Polder Soils)
Heavy, wet MontmoriUonitic clays at the Kafue
Pilot Polder are also liable to massive infestations of
grass weeds. Eleusine indica is present in abundance,
together with sedges of Cyperus spp.
(c) Virgin Sandy Clay Loams
The potential weed growth on virgin lands at
Nakambala is less severe than on the old arable areas,
and the weed flora consists largely of broad-leaved
"dicots".
2. Herbicides Tested
The following herbicides have been evaluated on
the two major soil types and at different timings with
respect to planting and irrigation:
(a) For pre-emergent control
(i) Prometryne at 4 and 6 lb/acre
(ii) Prometone at 4 and 6 lb/acre
(iii) 2, 4-D at 1 lb a.e./acre every 3 weeks
(iv) Diuron at 4 and 6 lb/acre
(v) Atrazine at 4 and 6 lb/acre
(vi) Ametryne at 4 and 6 lb/acre
(vii) Atratone at 4 and 6 lb/acre.
(b) For post-emergent control
(i) Paraquat at 2 pints/acre
(ii) Dalapon/2, 4-D at 3 lb Dalapon + U lb acid
equivalent 2, 4-D per acre.
By D. S. HUGHAN AND D. R. C. BOOTH
The chemicals were applied by hand operated knapsack
sprayers at a volume of 30 gallons per acre.
3. Trial Results
The most satisfactory results were obtained from
the following treatments:
(a) Pre-emergent treatments
On Sandy Clay Loams
(i) Prometryne at 4 lb/acre
(ii) Prometone at 4 lb/acre
(iii) 2,4-D at 1 lb acid equivalent per acre repeated
at 3 weekly intervals.
On heavy MontmoriUonitic Clays
(i) Atrazine at 6 lb/acre
(ii) Prometryne at 6 lb/acre
(iii) Ametryne at 6 lb/acre.
(b) Post-emergent treatments
On both of the above soil types — Paraquat at 2
pints/acre.
4. Detailed Assessment of Pre-emergent Herbicides
Tested
The preliminary assessment of each chemical tested
is as follows:
(i) and (ii) Prometryne and Prometone. These chemicals
both gave very good results, lasting up to 8 weeks
after spraying, against a wide range of weeds on the
Sandy Clay Loams. Somewhat better control was
obtained from applications of 6 lb/acre, but at this
rate neither material would be an economic proposition.
On the heavy clays of the Kafue Pilot Polder Prometryne
gave good results at 6 lb/acre, though not
comparable to those obtained from Atrazine. The
effect of Prometone was disappointing on the Polder
soils.
(Iii) 2,4-D (Amine). Repeated applications of 2,4-D
(amine salt) at 1 lb acid equivalent per acre, at 3
weekly intervals, gave very good results indeed on the
sandy clay loams. All broad-leaved weeds and a fairly
wide range of grasses were kept under good control
by this treatment, the only species showing any resistance
being the grass Eleusine indica. This 2,4-D
treatment was not included in the Polder trials and
was only evaluated on the sandy clay loams.
(iv) Diuron. This chemical gave excellent results
both at Nakambala and on the Polder when applied
at 6 lb/acre but was very disappointing at the 4
lb./acre rate.
At 6 lb/acre this costly material is not an ceonomic
proposition.

Proceedings of The South African Sugar Technologists' Association—March 1966 313
(v) Atrazine. This chemical, applied at 6 lb/acre,
proved to be the best pre-emergence herbicide of all
on the heavy Montmorillonitic clays. At 4 lb/acre
results were variable, and on the lighter soils Atrazine
generally failed to give any control at either 6 or 4
lb/acre.
(vi) and (vii) Ametryne and Atratone. These chemicals
gave fair degrees of control on the Heavy Clays
when applied at 6 lb/acre, but were still inferior to
Prometryne and Atrazine. They were not successful
on the Sandy Clay loams.
5. Timing of Application of Pre-emergent Herbicides
The timing of pre-emergent herbicide spraying in
relation to irrigation water application was found to
be important. All of the pre-emergence herbicides
tested have been most effective when applied between
the first and second irrigations of newly planted lands.
Both spraying and the second irrigation need to be
carried out within a week of planting and before any
emergence of weed seedlings occurs. Spraying on to
dry land immediately following planting and before
any irrigation, has generally resulted in very poor
weed control.
6. Detailed Assessment of Post-emergence Herbicides
(i) Paraquat. This chemical gave excellent results
when used as a post-emergence (contact) herbicide on
the inter-rows. One of the drawbacks of this material,
in common with most post-emergence herbicides, is
that owing to high phytotoxicity to cane it cannot be
applied directly to the rows. Even when applied between
the rows, there is danger of damage to cane by
Area
(a) Virgin lands .
(b) Old arable lands
Rows
2, 4-D (overall spray by
aircraft)
Prometryne or Prometone
(band spiay)
Initial Treatment
Table I
Inter-rows
2, 4-D (overall spray by
aircraft)
Nil
(a) Virgin Land
The use of 2,4-D is advocated for the main block
of virgin land, where experience to date has shown
that the potential weed growth is not only less severe
than on the old arable lands of the property, but is
also comprised largely of broadleaved "dicots". An
overall pre-emergence spray could be applied by aircraft
at 3 weekly intervals. Any weeds getting away
in the rows would be dealt with by Ihe available
manual labour, and in the inter-rows by tractor cultivation.
A more expensive alternative in case of continuous
rain will be by limited "spot spraying" with
Paraquat.
spray drift. Such damage occurred, in varying degrees,
on most of our trial plots, however this was confined
to localised scorching of the outer leaves and recovery
was fairly rapid.
(ii) Dalapon/2,4-D Mixtures. The application of a
mixture containing 3 lb Dalapon plus U lb acid
equivalent 2,4-D (amine) per acre gave good postemergence
weed control on the inter-rows at approximately
half the cost of using Paraquat. However,
damage occurred, and it is considered that phytotoxicity
hazards associated with the use of the cheaper
Dalapon are liable to be serious. This chemical is
translocated within the affected plant, and can have
long lasting deleterious effects which may not become
apparent for some considerable time after spraying.
7. Timing of Application: Post-emergence Herbicides
Both Paraquat and the Dalapon/2,4-D mixture were
found to be most effective when applied to weeds at
the early post-emergence stage when weeds are no
taller than three inches, and before they are allowed
to become well established. One well timed application
was found to be capable of giving good control for
6 to 8 weeks. Where applications were delayed and a
heavy cover of weeds had become established, then
a second application was necessary at 2 to 3 weeks
after the initial one. This applied to both Paraquat
and Dalapon.
Future Planning
A. proposed scheme of weed control, combining the
three basic methods of chemical, mechanical and hand
weeding, each to its best advantage, is shown in
Table 1:
Rows
Hand Weeding
Hand Weeding
Follow-up Treatment
Inter-rows
(i) Tractor cultivation,
(ii) Paraquat (spot spray)
(i) Tractor cultivation
(ii) Paraquat (spot spray)
(b) Old Arable Land
Areas of old arable land, which are liable to heavy
infestation of the grasses Rottboellia exaltata and
Eleusine indica, will receive a pre-emergence "band"
spray over the cane rows using either Prometryne or
Prometone at 4 lb per acre sprayed. The choice of
materials here will depend largely on availability and
costs at the time required. At present, although Prometryne
has given slightly better results than Prometone
at the 4 lb rate, the cost is 18 per cent higher.
Follow-up treatment of both rows and inter-rows will
be as for the 2,4-D treated areas. Some tractor cultivation
of the untreated inter-row spaces will be essential.

314 Proceedings of The South African Sugar Technologists' 1 Association—March 1966
Treatment
Area
Herbicide used
Type of Application .
Spraying Rate
No. of Applications .
Cost of Chemicals per
acre treated
2, 4-D
Virgin Lands
Overall spray by aircraft
1 lb a.e. per acre
3 at 5s
15s. = R1.50 . . .
Table II
Pre-emergent
Prometryne . . . .
Band spray by knapsack
or tractor
4 lb. per acre
1
38s. =- R3.80 . . .
Cost of Spray Materials
Approximate costs of herbicides used per acre
treated under the foregoing recommendation, and at
current prices, are summarised in Table II above:
Summary
Results of trials carried out over the past two years
have shown that the most promising herbicides are:
(a) For pre-emergence control on the lighter red
soils — Prometryne at 4 lb per acre and Prometone
at 4 lb per acre.
Old Arable Lands
For discussion see page 316
Prometone . . . .
Band spray by knapsack
or tractor
4 lb per acre
1
32s. = R3.20 . . .
Paraquat
Post-emergent
Virgin and Old
Arable Lands
Spot spray by knapsack
2 pints per acre
1
50s. - R5.00
(b) For pre-emergence control on heavy Montmorillonitic
Clays — Atrazine at 6 lb per acre.
(c) For post-emergence control on all soil types —
Paraquat at 2 pints per acre (taking into account
phytotoxicity hazards associated with the use of
the cheaper Dalapon).
A proposed scheme of field scale weed control combining
the three basic methods of chemical, mechanical
and hand weeding is given.

Proceedings of The South African Sugar Technologists' Association — March 1966 315
PROBLEMS IN WEED CONTROL ON AN ESTATE
Introduction
With the increasing use of herbicides in the fields
of The Tongaat Sugar Company Limited it has
become possible to make general observations on the
influence of chemical control on the character of the
weed population and the resultant change in the
nature of the weed problem.
Broad-leafed Weeds
(a) 2, 4-D amine. Broad-leafed weeds are controlled
largely by the use of 2, 4-D amine (7.2 lb a.e.*)
applied at 2.7 lb a.e. per acre, for both pre and post
emergent spraying. Pre-emergent spraying has not
proved successful in the cooler months and as a result
it is not practiced from May until the first spring
rains. Susceptible weeds such as Blackjack (Bidens
pilosa), St. Pauls weed (Siegesbeckia orientalis) and
Pig weed (Amaranthus spinosus) have, under chemical
control, almost been eradicated from many fields.
With the lack of competition so engendered, certain
weeds are appearing which are less susceptible to
2, 4-D. Foremost amongst these problem weeds is
Mint weed (Australina acuminata).
(b) 2, 4, 5-T. An observation trial was conducted
in conjunction with Kynoch-Capex in which several
levels of diuron, linuron, 2, 4-D ester, paraquat,
2, 4, 5-T and some mixtures of these were tried on a
stand of mint weed. Of the eighteen treatments, the
four most effective proved to be 2, 4, 5-T at 1.15 lb
a.e. per acre (cost R2.00), diuron at 4 lbs p.f per
acre (cost R9.88), linuron at 3 lbs p. per acre (cost
R8.10) and a mixture of 2 lb. diuron p., 1 pint
paraquat p. and 1.15 lbs. 2, 4, 5-T a.e. (cost R9.22).
2, 4, 5-T was then tried on a field scale at 1.15 lb.
a.e. per acre full cover, where it has proved to be
completely effective in controlling mint weed. Increasing
use is being made of 2, 4, 5-T in the control of
other broad-leafed weeds which have grown too old
to be controlled by normal field applications of 2,4-D
amine. Field results indicate that some control is also
achieved on watergrass (Cyperus sp.). It has up to now
only been regarded as a brush killer. It would appear
from these limited observations that more intensive
investigations are warranted on this herbicide with
regard to its possible uses, both pre and post emergent
on broad-leafed weeds and possibly with certain
admixtures as a herbicide for the control of watergrass.
* a.e. = acid equivalent
t p. ~ product
By E. C. GILFILLAN
Cyperaceae
Watergrass (Cyperus esculentus) in plant cane is
easily controlled with paraquat at 2 pints p. per acre.
The watergrass comes away ahead of the plant cane.
Because of this the paraquat spray achieves maximum
burn on the watergrass with minimum burn on the
cane. In ratoon cane however the position is reversed
so that, on spraying severe scorching of the cane
results, with only an ineffective burn of the watergrass.
Linuron at 4 lbs. per acre was tried as an alternative
measure for watergrass control in ratoons. Initial
results were most encouraging and excellent control
was achieved for three months, with no visible damage
to the cane. Subsequently however very disappointing
results were obtained under very nearly identical
conditions, and the herbicide was temporarily dropped
as a recommendation. Moisture is believed to be the
limiting factor but it is felt that critical investigations
should be conducted on this herbicide, so as to
establish the causes for its erratic performance.
Gramineae
As better control is gained over broad-leafed weeds,
it is becoming increasingly apparent that effective
control must be gained over grasses as well. Of the
perennial grasses, Ubabe (Panicum maximum) and
Mqangabodwe (Sorghum verticillifiorum) are the
greatest problem. Ubabe is the worst invader of cane
fields and competes fiercely with cane. Heavy treatments
of Dalapon (up to 20 lbs. per acre) have been
found to effect good control; however such severe
treatments cannot be used in fields because of damage
to the cane. A successful treatment for all grasses
has been developed and is being used on a limited
scale. The treatment consists of a light application of
Dalapon (up to 3 lbs. per acre) followed by paraquat
at 2 pints per acre from seven to ten days later. This
treatment effects a total kill of Ubabe but also burns
the cane quite severely. The cane does recover however.
The treatment is meant specifically for use in
young cane where it is possible to spot spray the
Ubabe with very little damage to the cane. It has also
been used with success in cane up to four feet tall but
it was not possible under these conditions to spot
spray and the cane was severely burnt and took about
three weeks to recover. Complete kill of the Ubabe
was achieved. This can only be achieved by hand
weeding when the Ubabe stools are actually carried
out of the field to prevent regrowth.
Discussion
It would appear from the above observations that
problems in weed control are getting worse. As
herbicides and techniques are developed which kill

316
susceptible weeds, so they are replaced by less susceptible
ones and a vicious circle is developed. This leads
to the comment from some that it would have been
better never to have started in the first instance. This
is not the case at all. The introduction of herbicides
into regular cane farming practice has assisted
enormously in reducing the weed population. Herbicides
are only used when their use is economically
justifiable as compared to mechanical control or hand
weeding, or when shortage of labour or machinery
or adverse conditions warrants their use. The attempts
made to control other weeds apart from broad-leafed
species, merely serves to illustrate that these weeds
are now "seen" and that weed control is advancing to
its ultimate goal of complete control.
Summary
Weed control problems are discussed from a
practical estate view point, with reference to three
main groups of weeds: Broad-leafed, Cyperaceae
and Gramineae. The uses and limitations of 2, 4-D,
2, 4, 5-T and paraquat are described and a new
technique for grass control is advanced which involves
a split treatment of dalapon and paraquat.
Mr. Pearson: From experience in the Natal Midlands
it is apparent that 2, 4-D is very effective if it
is applied at the right time i.e. under correct weather
conditions. On one occasion on a misty day 2, 4-D
was applied in the furrow after planting and complete
control of watergrass and all other weeds was obtained.
The next day was bright and sunny and 2, 4-D was
again applied but did not effect control of water grass.
Mr. Wyatt: I feel that the authors of the papers
could have paid more attention to the methods of
application of herbicides.
Mr. Gilfillan: Work is being done along these lines
and the following figures have been obtained so far:
1. One labourer using a knapsack spray can
cover an area of two acres a day, at a cost of
60 cents/acre.
Proceedings of The South African Sugar Technologists' Association — March 1966
3. A tractor mounted sprayer under good conditions
can spray 30 acres per day at a cost of
only 30 cents per acre.
3. Aircraft spraying has proved difficult owing
to the necessity to fly only six feet above the
crop but new emulsions being used will enable
a twenty feet clearance to be used.
Mr. Tucker: Has Mr. Booth any information about
helicopter spraying in Zambia?
Mr. Booth: It has been carried out but it is very
expensive and not entirely satisfactory.
Mr. Aucock: Has anyone tried herbicide in irrigation
water.
Dr. Thompson: The application of overhead spray
is not uniform enough for herbicide distribution.
Mr. Andries: We did one season of aerial application
and the cost of application alone was 85 cents
per acre, compared with 55 cents per acre for tractor
sprayings, both on an irrigated lay-out. There are
advantages in aerial spraying in being able to choose
the optimum period for application and because of
the enormous areas which can be covered quickly.
One day, after 1/2 inch of rain we did a blanket spray
of 385 acres in one day—6.00 a.m. to 4.00 p.m.
We have also used for spraying a method used by
Crookes Bros, on the South Coast, namely a portable
boom with a small motor drive and two three hundred
feet hoses. It is an eighteen feet boom carried by two
men and is highly effective for hand application.
Mr. Johnson: The cost of helicopter spraying is
two and half times that of a conventional aircraft
and the down-draught of air from the rotors bounces
back from the land and prevents proper penetration
of the herbicide.
Mr. Wardle: If nozzles are correctly positioned in
helicopters the disadvantage of the down draught can
be overcome.
On tractor sprays the mist blower is used at present
in the Free State on wheat crops. It has the advantage,
for instance, of getting the same results from 2, 4-D
with approximately half the recommended rate of
application and doing 100 acres per day. The rig
would have to be modified for our terrain.
\

Proceedings of The South African Sugar Technologists'' Association—March 1966 317
PEST CONTROL PROBLEMS AT MAZABUKA, ZAMBIA
Introduction
The recording of four insect pests of sugar cane on
the newly developing Nakambala Estate at Mazabuka
during the current summer season (1965/66) illustrates
the importance of conducting cane trials over as long
a period as possible before embarking on full scale
development, and of being prepared for such outbreaks
when working in a new and unfamiliar
environment.
Since November 1965 the following four pests
causing damage to cane have occurred:
(1) Trash Caterpillar (Cirphis sp.)
(2) Top Borer (Lepidoptera)
(3) Leaf Roll Caterpillar (Marasmia trapezalis)
(4) Black Beetle (Heteronychus sp.)
The only pests of cane previously recorded in the
district were:
(a) Mole Cricket (Gryllotalpa sp.)—A localised
attack on germinating cane plants at the Kafue
Pilot Polder in September, 1963, was successfully
controlled by the application of 2% Dieldrin
dust at 50 lbs/acre.
(b) Top Borer (Lepidoptera)—Occasional very light
infestation at the Kafue Pilot Polder and
Nakambala Estate.
Of the recently occurring pests, Trash Caterpillar
and Heteronychus Beetle have caused intensive damage
and could possibly develop into major pests. Top
Borer, though rather more widespread in 1965 than
over the previous two years, has not assumed serious
proportions and can still be rated as a minor pest.
The same applies to Marasmia trapezalis the damage
from which, though unsightly in appearance, is
largely superficial and of short duration.
Brief descriptions of these pests, with notes on
methods of control, follow below.
1. Trash Caterpillar (Cirphis sp.)
This caterpillar, a type of "army-worm", causes
serious damage in young ratoon cane under a trash
blanket—completely stripping the leaf blades and
cutting off emerging shoots. At Chirundu Estate, in
the Zambezi Valley, attacks occur annually in July
and August and are most effectively controlled by
dusting with 2 1/2 % D.D.T. at 50 lbs /acre. The dust is
usually applied during early mornings and at dusk
to take advantage of calm wind conditions, and is
applied directly on to the young ratoon cane rows
using small hand dusters, one labourer covering 5
acres per day.
By D. S. HUGHAN and D. R. C. BOOTH
A very serious outbreak of this pest occurred at
Nakambala towards the end of November, 1965,
much later in the season than anticipated. 200 acres of
ratoon cane were affected and over most of this area
a single application of 1\% D.D.T. dust brought
about complete control; however, in the areas where
the infestation commenced cane growth was set back
by two months.
2. Top Borer (Lepidoptera)
Incidence of a Lepidopterous larva causing "dead
heart" in both plant and ratoon canes has increased
somewhat, but has not yet assumed serious proportions.
No control measures have been taken so far
but careful observation is being maintained and the
possibility of introducing Tachinid parasites has been
considered.
3. Leaf Roll Caterpillar (Marasmia trapezalis)
Almost the entire plant cane acreage of the estate
became infected by caterpillars of the moth Marasmia
trapezalis early in December, 1965. This caterpillar,
which is an occasional pest of Graminaceous crops,
feeds on the upper surface of the leaf tips. The edges
of the leaf are drawn together and stitched in place
by silk threads to form a protective tunnel. The lower
epidermis is left intact. The larval phase is over in
about ten days and leaves the affected area looking
unsightly because of the brown colour of the damaged
leaf tips. Fortunately the crop does not appear to
suffer any serious setback, since the damage is restricted
to a few inches of the leaf tip.
A portion of the infested plant cane area was
dusted with 2{-% D.D.T., but this was discontinued
owing to the natural termination of the attack. A
slight recurrence of this pest took place towards the
end of January, 1966. It has been observed that
Marasmia is very common on areas of grass
surrounding the estate.
4. Black Beetle (Heteronychus sp.)
The first noticeable attack on cane by Heteronychus
Beetle occurred on heavy, wet Montmorillonitic soils
at the Kafue Pilot Polder at the beginning of January,
1966. Towards the end of the same month it was
found in. a rather wet clay area on Nakambala Estate.
The original outbreak at the Polder commenced in isolated
pockets wherever cane was at all stunted through
poor drainage, but soon afterwards it spread throughout
the entire field. The adult beetles form a network
of tunnels just below ground level from which they
move up into the cane shoots and stems. The cane
shoots are completely severed from the stool through
the eating away of the centres of the stems. Stands of
cane at all stages of growth are attacked and some

318 Proceedings of The South African Sugar Technologists' Association—March 1966
Summary
Trash Caterpillar has been controlled cheaply and
effectively by single applications of 2 1/2 % D.D.T. dust.
Top Borer is rated as a minor pest and no control
measures have been taken so far, though the possibility
of introducing Tachinid parasites has been considered.
Leaf Roll Caterpillars of the moth Marasmia
trapezalis have caused some damage to cane, but are
likewise at present rated as a minor pest.
Heteronychus Beetles have caused serious damage to
cane at all stages of growth on heavy, wet soils, but
have been brought under control by spraying with
Dieldrin at the rate of 2 lbs active material per acre.
Further insecticide trials are being carried out against
this pest.
Dr. Dick: Experiments in Natal have shown that a
severe outbreak of trash caterpillar can cause a loss
of 7 tons cane per acre. The caterpillars are well
controlled by parasites, especially Tachinid flies and,
if insecticides are to be used at all, they should be
applied as early as possible. Once the cane has been
damaged, treatment may do more harm than good
since the parasites as well as the caterpillars will be
killed.
The top-borer found in Natal is Sesamia calamistis
and it is very likely that the one recorded from Mazabuka
is the same. This insect is well controlled by
Braconid wasp parasites and does not usually cause
economic damage in South Africa.
The leaf-roller caterpillar is also well controlled
by Braconid parasites and is of no real importance.
Beetles of the genus Heteronychus, of which H.lieas
is the most important in South Africa, are the most
dangerous of the insects mentioned. The larvae, which
feed on the roots of sugarcane, are probably even
more harmful than the adults. Biological control
appears to be ineffective and, in areas such as Swaziland
where outbreaks have taken place, insecticide
applications have been undertaken. Dieldrin is the
most suitable insecticide available and emulsions
are preferred to wettable powders since they penetrate
the soil more effectively. Control is aimed at the
adults which feed near the surface of the soil and are
most numerous from October to January. Larvae,
which occur much deeper in the soil, generally escape
contact with the insecticide.
Mr. Carnegie: Have you tried leaving infestations
of borer and trashworm untreated to find out whether
biological control will take over?
Mr. Booth: This has not been tried.

Proceedings of The South African Sugar Technologists' Association — March 1966
THE PROGRESS OF AN UNTREATED OUTBREAK
OF NUMICIA VIRIDIS, MUIR
by A. J. M. CARNEGIE
The green leaf-sucker, Numicia viridis, Muir (Homoptera:
Tropiduchidae) was first recognised as a pest of
sugarcane in 1962, when its activities caused conspicuous
damage on an estate in Swaziland, (Dick,
1963). It is an indigenous insect which has been recorded
now from 32 grass species, and it can be found
on cane, if only in small numbers, in most cane growing
areas of South Africa and Swaziland.
In April 1965 an opportunity arose of following the
course of an infestation of N. viridis in a field of
irrigated cane (N:Co.310) at a large sugar estate in
Swaziland. At the time of writing the infestation has
been visited ten times, at approximately monthly
intervals, and notes have been made and material
collected for subsequent laboratory examination.
At the time these observations started numbers of
N. viridis adults and nymphs suddenly became very
numerous in four fields on the estate. Three of these
were treated by aerial application of malathion
5 per cent dust, but the management of the estate
kindly agreed to leave one field of five month old
cane of variety N:Co.310 untreated, and placed it at
our disposal. Each time the untreated field was visited
the general state of the cane was noted, samples
taken and counts made. Regular samples included
the following:
1. Adults and nymphs
In six randomly selected places cane was shaken over
a one yard square piece of black plastic sheeting, and
all. N. viridis adults and nymphs which dropped were
collected for subsequent counting and examination.
2- Eggs
Egg samples were taken both in cane and in grasses
growing in the vicinity of the field. Grasses included
the following species: Panicum maximum, Pennisetum
thunbergii, Pennisetum clandestinum, Rhynchelytrum
repens, and the sedge Cyperus sexangularis. Whenever
available, egg samples from grasses included some
from leaves and some from inflorescence stems.
Subsequent treatment of samples
Adults and nymphs collected by shaking were
counted, examined for parasites, and the adults sexed.
Table I
Numbers of N. viridis collected by shaking cane. Total shown is
for six randomly chosen sites on each occasion
Date
8.6.65
29.6.65
20.7.65
18.8.65
15.9.65
11.10.65
14.11.65
14.12.65
17.1.66
Total
N. viridis
201
194
350
395
445
114
82
13
17
%
Nymphs
6.9
39.1
94.3
99.7
98.2
88.6
34.1
38.4
29.4
Adults
% Female
52.9
53.4
35.0
0.0
37.5
38.4
42.6
12.5
75.0
% Male
47.1
46.6
65.0
100.0
62.5
61.6
57.4
87.5
25.0
»/
Parasitised
by Dryinids
3.5
2.6
2.3
0.0
1.5
0.9
3.6
30.8
11.6
Egg batches from all media were dissected and
divided into the following categories: hatched,
unhatched, parasitised by Oligosita sp. (Trichogrammatidae),
parasitised by Ootetrastichus beatus
Perkins (Eulophidae), degenerated through causes
other than parasites, still likely to hatch.
Before these observations started and until the time
of writing, similar weekly or fortnightly sampling
for eggs, nymphs and adults in cane has been done
also by the agronomy section of the estate. All leaf
material containing eggs has been forwarded to the
Experiment Station and has been treated as mentioned
above; their figures for nymph and adult counts have
also been made available.
From data accumulated it has been possible to
obtain figures for N. viridis populations, as eggs,
nymphs and adults from May 1965 until the present
time.
Eggs
Numbers of unhatched eggs per leaf are shown in
Fig. la. There was an increase in numbers while
adults were plentiful, and a decrease as numbers of
adults fell and nymphs became plentiful (Fig. lb).
The slight increase in egg numbers in September is
unexpected since numbers of adults were low then,
but the increase in numbers in October and November
followed the appearance of a second generation of
adults. Numbers of eggs then decreased.
Nymphs and adults
Numbers of both nymphs and adults were high
when sampling began in April, with numbers of
nymphs falling rapidly as the adult stage was reached
(Fig. lb). This was followed by a fall in numbers of
adults, which had died after copulation and oviposition.
There was then a further increase in nymph
numbers as a second generation arose, and this was
followed by a small increase in adult numbers between
October and mid-November, with a subsequent drop
in numbers of both nymphs and adults.
Factors causing reduction in numbers
A reduction in numbers with metamorphosis and
from adults dying after copulation and oviposition
is to be expected, but the extent to which numbers of
all stages were reduced, and the failure in production
of even larger following generations cannot be explained
only in these terms, and attention was paid
to a number of other factors.
1. Egg parasitism
Two egg parasites of N. viridis have been noted in
most areas where the host has been collected. These
two minute wasps, Ootetrastichus beatus and Oligosita
sp. were present in the field under discussion
throughout the period, and notes were made of their
activities. Both parasites exercise a controlling effect
on their host's numbers by laying their own eggs
inside those of the host.

320 Proceedings of The South African Sugar Technologists' Association — March 1966

Proceedings of The South African Sugar Technologists' Association — March 1966 321
FIGURE 2 : Relative importance in biological control of two egg parasites, Ootetrastichus beatus and Oligosita sp.
(a) respective numbers of parasites (relative to total eggs).
(b) eggs parasitised by each parasite (as percentage of total eggs).

322
The young grub hatching from the egg of Oligosita
sp. feeds on the contents of the host's egg in which it
completes its larval development and in which it
pupates. The adult wasp emerging from the pupa
bites its way to the exterior through the egg chorion
and the leaf midrib. Thus one wasp larva accounts
for one N. viridis egg.
The life cycle of O. beatus is slightly different, and
has not yet been worked out in detail; but that of a
similar species, O. pallidipes, has been described
(Williams, 1957) and the pattern is similar. The young
larva feeds on the contents of the host's egg until
the last instar is reached, by which time it fills the egg.
It then leaves the host's egg and may pupate immediately
in the surrounding plant tissue, or it may
feed on adjacent eggs, consuming as many as six,
before pupating. The resulting adult wasp bites its
way through the midrib to the exterior. It is of
interest that during its final instar the larva of this
parasite may eat eggs already parasitised by its own
species or by Oligosita sp.
From regular leaf samples sent to us by the estate
it was found that as a biological control factor
O. beatus was the more important of the two egg
parasites (Figs. 2a and b). Even where numbers of
the two parasite species were similar, more host eggs
were accounted for by O. beatus than by Oligosita sp.
It is of interest to compare the state of egg parasitism
in cane leaves with that in adjacent grass species
(Fig. 3a and b). The figure shows percentages of
parasitised eggs in cane leaves, leaves of the grasses
Panicum maximum, Pennisetum clandestinum and
Pennisetum thunbergii, in inflorescence stems of
Panicum maximum, Pennisetum thunbergii and Rhynchelytrum
repens, and in stems of the sedge Cyperus
sexangularis. All these were growing in the immediate
vicinity of the cane field, and the same patches of grass
were sampled each visit. In the figure readings represent
all available grass stems and all available grass
leaves respectively. Figures for the sedge are plotted
separately.
Egg parasitism followed the same trends in grass
stems and leaves and in the sedge, being generally
higher in the sedge than in any of the grasses, and
always higher in grass leaves than in stems. The
explanation for this could lie in the texture of the
plant tissue concerned. In the course of field observation
it has been noted that the wasps insert their eggs
through that side of the midrib or inflorescence stem
which is opposite the egg operculum. This necessitates
penetration of the plant tissue, and softer plant
tissue may result in more successful parasitism.
Inflorescence stems of grasses are tougher than leaf
midribs, and in the case of the sedge, the hosts' eggs
were always laid along one of the ridges of the sixangled
spongy stem, through which the wasp parasite
should easily insert her egg.
In cane the same general trend was not followed
(Fig. 3 b) and parasitism was generally lower than in
grass leaves and in the sedge. In Figure 3b the unbroken
b'ne represents egg parasitism in green cane
leaves selected at random throughout the infested
area of the field, and the broken line represents egg
Proceedings of The South African Sugar Technologists' Association — March 1966
parasitism in green cane immediately adjacent to the
sedge C. sexangularis, in which a similar egg parasitism
pattern is not reflected.
Wherever egg parasitism in grasses has been compared
with that in cane (including other localities
and other grass species) figures have shown it to be
less successful in cane. There are probably several
reasons for this including: the relatively tough texture
of cane leaf midribs, periodic harvesting (often accompanied
by burning) destroying parasite populations,
the greater height of cane may make host's
eggs more difficult to find for parasites which are
adapted to a more horizontal and closely-packed
grass community.
In the course of egg examinations records were kept
of all parasites which had died in the plant tissues
without reaching the exterior, and it was found that
more parasites died in cane than in grasses (Table II).
In cane Oligosita sp. in particular experienced difficulty
in biting its way out of the leaf midrib. A high
proportion died also in grass stems, but in grass leaves
most survived. The rather high percentage of dead
parasites in the sedge was unexpected; parasitism,
there was also high (Fig. 4a). Figure 3 suggests that
there was a "rhythm" of parasitism in indigenous
host plants which was not found in cane.
Table II
Parasites which died without leaving plants; expressed as percentage
of total number of each species.
Medium
Cane leaves .
Grass leaves .
Grass stems .
Cyperus stems
Ootetrastichus
beatus
%Dead
12.6
13.7
7.9
21.2
Oligosita sp.
% Dead
54.8
5.3
30.3
25.0
2. Further causes of egg mortality
It was noted also that a number of eggs had
degenerated through causes other than parasitism.
Some had become translucent but hard, some had
collapsed and appeared to consist of a chorion only,
and others appeared to have been destroyed by a
fungus. Numbers of such eggs from various plant
hosts are shown in Table III. On nearly all occasions
figures from grasses were higher than from cane.
Fewer still were destroyed in the sedge, but parasitism
there was so high (compare Figure 3a) that the sample
of eggs degenerated through causes other than by
parasitism was too small to give significant results.
In cane the rigid structure of the leaf midrib protects
eggs of N. viridis from mechanical damage.
It has been noted that eggs will hatch even from
completely dry cane leaves which have been lying on
the soil in a crumpled condition for more than a week.
3. Parasitism of nymphs and adults
A wasp parasite, Lestodryinus sp. (family Dryinidae.
has been reared from both nymph and adult N. viridis)
The female wasp inserts her egg into the host's body,

Proceedings of The South African Sugar Technologists' Association — March 1966 323

324
and the hatching parasite larva feeds at first within
the host's body and soon appears as a protrusion of
the host's integument. The protrusion grows, the
parasite feeding on the body contents of the host
without immediately killing it. When mature, the
caterpillar-like parasite larva leaves the host and
pupates nearby, usually on a cane leaf. The host dies.
From the pupa the adult wasp emerges.
The impression is that when outbreaks of N. viridis
first occurred these parasites were more common than
they are now. Certainly they can have had little effect
on the infestation under consideration. In the course
of collecting nymph and adult N. viridis, records were
kept of dryinid parasites, and there was no appreci­
Medium
Cane leaves . . . .
Grass leaves . . . .
Grass stems . . . .
Cypenis stems
! No figure available.
Proceedings of The South African Sugar Technologists' Association — March 1966
Table III
able build up as the infestation grew, (Tables I and IVa).
The parasite's activity is checked to some extent by a
secondary parasite, Chiloneurus sp. (family Encyrtidae).
In Table I latest figures for percentage parasitism
by Dryinids are relatively high but they are
calculated from such small total numbers as to be of
little significance.
4. Other causes of reduction in numbers
As this infestation proceeded it became apparent
that the general arthropod fauna of the field was
becoming richer. While sampling for N. viridis by
shaking cane the numbers of general predators and
omnivorous insects which fell on to the sample sheet,
including numerous spiders, some ladybird beetles
Eggs destroyed through causes other than parasitism; expressed as percentage of total eggs
8-6-65
0.4
10.3
13.9
*
30-6
0.1
13.8
20.8
0.9
22-7
1.2
13.2
13.5
6.2
19-8
9.7
18.2
22.4
2.3
Table IV
Date
15-9
2.2
14.6
13.9
2.6
12-10
7.6
17.2
9.2
2.7
15-11
12.4
12.4
11.6
4.9
14-12
4.3
10.6
13.1
3.8
18-1-66
Average number of Arthropods collected by fumigating cane (3 to 4 stools at a time) in (a) an untreated field, and (b) a field dusted with
Date
15- 9-65
12.10.65
15.11.65
15.12.65
18. 1.66
12-10-65
15-11-65
18- 1-66
N. viridis
101.0
116.0
31.4
4.0
4.0
2440.5
92.2
16.0
Spiders
& predacious
mites
46.0
38.0
25.7
13.0
2.0
53.5
21.6
17.0
malathion 5 per cent dust
A
Ladybird
beetles
0.0
2.0
0.0
0.5
0.0
B
10.0
7.0
0.0
and the small predacious mite Anystis baccarum
(Linn.), appeared to increase. To investigate the
fauna more thoroughly a method was devised of
sampling by fumigation.
A plastic canopy was placed over three or four stools
of cane at a time in the form of a tent. Ground sheets
were placed under this, and insecticide fumigant
tablets ignited within it. After a suitable period
everything which fell was collected and preserved for
later examination. Results of preliminary sortings are
listed in Table IVa. From the figures available (September
1965 onwards) it can be seen that while numbers of
Dryinid
parasites
0.5
0.0
0.3
1.0
0.0
0.5
0.03
0.0
Other
insects
135.0
103.0
78.0
36.0
15.5
100.0
64.3
10.0
/o
Other
insects
47.8
40.6
57.6
92.6
72.1
3.8
34.6
23.3
0.2
19.2
14.6
6.7
/a
Predators
16.3
15.8
19.0
22.2
9.3
2.4
15.4
39.6
N. viridis were high so were numbers of other arthropods,
and it might be assumed that the action of
predators accounted for the gradual fall in numbers
of N. viridis. But in Table IVb fumigation figures are
shown for another field, which was dusted with insecticide
in May 1965. In this field (discussed in the
next section) numbers of arthropods other than
N. viridis were proportionally lower and could not
account for the drop in numbers which occurred
there between October and November (Fig. 4b).
This drop in numbers may have been due to natural
mortality following copulation and oviposition

Proceedings of The South African Sugar Technologists' Association — March 1966 325

326
(only 11.1 per cent of N. viridis in October's sample
were nymphs), but it seems too great and sudden a
reduction to be so explained. If mortality was natural,
i.e. followed normal oviposition, a gigantic following
generation might be expected.
Other natural causes of mortality which deserve
mention are the actions of insectivorous birds and
climatic factors. In the field which was left untreated,
during the period of greatest infestation several species
of small birds were active in the cane, and it is reasonable
to suppose that they took their toll. Owing to an
outbreak of foot-and-mouth disease it was not
possible to shoot any birds for identification and
examination of stomach contents. Heavy rain,
especially when accompanied by high winds must have
an adverse effect on nymphs and adults, and may have
contributed towards the drop in numbers in the treated
field.
A note on numbers of N. viridis following insecticide
treatment
Figures for N. viridis populations were available
for one of the fields which were treated on the 28th
May, 1965 (at about the time these investigations
started). The field was dusted from the air using malathion
5 per cent dust at the rate of 40 lbs. per acre.
Population trends for this field are illustrated in
Figure 4.
Following dusting there was an immediate drop in
numbers of nymphs and adults, as was to be expected.
But by then many eggs, which are not affected by
malathion dust, must have been laid, for by the end of
August numbers of nymphs had reached the 4,000
mark (per 10 square yard sample), and six weeks
later over 2,000 adults were recorded.
Insecticide dusting, in the presence of unhatched
eggs could account for the sudden rise in numbers of
nymphs and adults. Parasites within the host's eggs
or within the leaf tissue would not have been affected,
but any adult parasites on the cane at the time of
dusting would have been very vulnerable, and many
adults emerging shortly after dusting would have been
killed by contact. In addition, natural enemies which
do not spend a long egg or pupal period in a protected
state, such as spiders and ladybird beetles, would
also have been affected. There would therefore have
been very little to prevent numbers of nymphs and
adults from becoming very great in the following
generation, some time being necessary for a buildup
of natural enemies.
Should insecticide be used?
Whether or not it is better to leave an outbreak of
Numicia viridis completely untreated is still uncertain.
It certainly depends to some extent on local circumstances
such as, among others, the size of the affected
area, age of the cane, and presence of natural enemies.
Where insecticide is applied it must be done thoroughly
and be well timed and should usually be done
more than once in the space of a few weeks. In most
outbreaks investigated generations of N. viridis
have not been staggered, so that good timing of
application should be possible. There is evidence that
insecticide dusting is frequently followed by marked
increases in numbers of unwanted insects, including
Proceedings of The South African Sugar Technologists'' Association — March 1966
N. viridis. Certainly in the case of the fields discussed
above the infestation in the untreated field dwindled
while that in the treated field went from strength to
strength; but this is an isolated case and should at
this stage be regarded as an indication rather than as
conclusive evidence. The fields discussed were of
different cane varieties (N:Co.310 and N:Co.376),
and the comparison therefore is not entirely fair. It
must be remembered also that the first recognised
outbreaks of N. viridis which occurred in 1962,
(Dick 1963), did not follow insecticidal treatment.
When contemplating insecticidal action the variety
of cane affected should also be considered. The matter
of varietal susceptibility is being investigated, and it
seems that the insect affects some cane varieties more
adversely than others. Regarding the fields discussed
in this paper, the untreated field (N :Co.310) looked in
extremely poor condition and almost beyond recovery
in late July and August, although with the
subsequent drop in numbers of TV. viridis it recovered
quite rapidly. In the treated field (N:Co.376) it required
the presence of very large numbers of the insect
over a long period before the cane appeared to be
really adversely affected (in November 1965).
The fields concerned have not yet been harvested,
and it is impossible therefore at this stage to know the
loss in yield resulting from the untreated infestation.
Yield is expected to be about 5 tons per acre less than it
would otherwise have been, but the loss may be less
severe.
Summary
The progress of an untreated infestation of the leaf
sucker Numicia viridis Muir (Homoptera : Tropiduchidae)
in a field of sugar cane is discussed. It
was found that after fluctuations in numbers of all
stages of the insect, the infestation gradually dwindled
and the cane, which had suffered badly, recovered.
Causes of the drop in numbers are discussed, mention
being made of two egg parasites, Ootetrastichus
beams Perkins (Eulophidae) and Oligosita sp. (Trichogrammatidae);
a parasite of nymphs and adults,
Lestodryinus sp. (Dryinidae), and various arthropod
predators. A comparison is made of natural control
factors in grass and in cane and possible reasons given
for control being generally better in grasses. A brief
comparison is made between arthropod populations
in an untreated cane field and in one dusted with 5 per
cent malathion, and this is followed by a discussion of
the merits of insecticide use.
Acknowledgments
The generous help and co-operation of the management
and staff of Ubombo Ranches, Swaziland, is
greatly appreciated. Messrs. Kirby and Roodt of the
Experiment Station helped accumulate much of the
data. The paper includes material being used for postgraduate
work in the Entomology Department of the
University of Natal.
References
Dick, J. 1963. The green leaf-sucker of sugarcane Numicia
viridis, Muir. Proc. S. Af. Sug. Technol. Ass. 153-157.
Williams, J. R. 1957. The sugarcane Delphacidae and their
natural enemies in Mauritius. Trans. R. ent. Soc. Lond. 109(2):
65-110.

Proceedings of The South African Sugar Technologists' Association—March 1966 327
Mr. du Toit (in the chair): Referring to figure 4b,
why are you reluctant to accept that parasites could
have caused the fall in numbers of N. viridis between
October and November?
Mr. Carnegie: Because I consider that, judging by
the tremendous increase in Numicia numbers following
insecticide application, most parasites, excepting
those actually in the hosts' egg, must have been killed
by the insecticide, and I do not think that by October
their numbers could have increased sufficiently to
have brought about this control of the host. There
may be some other control factor which we do not
yet appreciate which may be density dependent, such
as the production of infertile eggs, for instance.
Dr. Dick: Ubombo Ranches are to be congratulated
for having the courage to leave one field untreated.

328
Proceedings of The South African Sugar Technologists' Association—March 1966
THE SUGARCANE NEMATODE PROBLEM
A paper presented to this Association in 1961
summarised the results of investigations, carried out
during the previous four years, on the role of nematodes
in South African sugarcane fields. 7 This project
has continued to receive attention, and the present
paper is an attempt at summarising and generalising
upon the results of more recent investigations. Since
much of the experimental work involved has been
described in Annual Reports of the Experiment
Station, it is unnecessary to repeat details such as
counts of nematodes and yields of cane obtained from
particular nematocide trials, although a few are
quoted as examples. If a single generalisation could
be made at this stage, it would probably be that,
despite investigations here and in other countries, the
role of nematodes in sugarcane fields remains in many
aspects a problem.
Types and status of Nematodes Concerned
Examination of soil samples from South African
sugarcane fields reveals the presence of many different
species of nematodes, and the task of determining
which, if any, actually injure the cane plant is by no
means easy. As a rule, the majority of individuals
belong to the Order Rhabditida and other groups of
which no member is known to be parasitic on plants.
Most of these feed on decomposing organic matter or
associated micro-organisms. Nematodes which attack
the roots of plants may be internal parasites, such as
the various species of Meloidogyne and Pratylenchus,
external feeders, such as Criconemoides, Xiphinema
and Trichodorus, or intermediate in being able to feed
either externally or within the roots as do some species
of Hoplolaimus and related genera. The plant parasites
found in our cane fields, most of which have as yet
been identified only to the genus, are listed in Table I.
Table I
Plant parasitic nematodes recorded from South African
sugarcane fields
by J. DICK
Their presence in this environment does not, of course,
prove that they attack sugarcane since weeds are
often present in sufficient numbers to provide them
with nutriment. For species which occur in the roots
of sugarcane it might appear safe to assume that they
are parasitic on this plant, although the extent of their
effect on it might be difficult to determine. Even in
this environment, however, some saprophagous worms
are usually present since the root system, especially of
older cane, includes a considerable amount of dead
and decomposing tissue. The improved yields which
generally follow applications of nematocides to the
soil of growth-failure areas suggest that nematodes
have been responsible for poor growth prior to
treatment. In interpreting these results, it must be
remembered that most nematocides exercise some
control over other organisms such as soil arthropods,
molluscs, or even weeds, and that they may change
the nature and amount of available plant nutrients
through their action on soil micro-organisms.
Direct evidence has been obtained in some other
countries by planting sugarcane in sterilised soil
inoculated with more or less pure cultures of particular
species. This can, unfortunately, only be done in
pots and the results are qualitative rather than quantitative.
Such experiments have produced evidence that
the following species do affect the growth of sugarcane:
Meloidogyne incognita acrita, Helicotylenchus
dihystera, Tylenchorhynchus martini, Pratylenchus zeae
and Trichodorus christei. 1 - 8 Our own experiments have
given similar results for Meloidogyne javanica while
ciicumstantial evidence casts suspicion on Pratylenchus,
various Hoplolaimids, Trichodorus and, in a
few localities, Xiphinema.
Root-knot nematodes (Meloidogyne spp.) are most
numerous, and presumably most injurious, in sugarcane
roots during early growth. Examination of
samples collected at monthly intervals showed that
populations in the roots attained their maximum
from four to six months after planting, after which
they decreased fairly rapidly (Fig. 1). Although numbers
increased again in young ratoons, they did not
reach the level attained in the plant cane. In at least
one field, it was found that Meloidogyne was progressively
replaced by a species of Pratylenchus (lesion
nematode) as the cane roots became older.
In addition to saprophagous and plant parasitic
nematodes, soil samples may yield specimens of
predacious types which feed on small soil animals
including other species of nematodes. The most
noticeable of these belong to the family Mononchidae;
they seldom occur in our cane field soils in sufficient
numbers to be of importance in biological control of
the plant parasites.
Association with organisms causing disease
Certain virus diseases of plants have been shown
to be transmitted by nematodes, all known vectors

Proceedings of The South African Sugar Technologists' Association — March 1966 329
being members of the Order Dorylaimida. At present
no instance of this type of transmission is known for
sugarcane although Martin et al. (1960) attempted to
transmit ratoon stunting and chlorotic streak by this
means. 8
Synergistic association between nematodes and
parasitic micro-organisms, including fungi and bacteria,
has been demonstrated in a number of host
plants. Up to the present, no such association has been
proved to occur in sugarcane, although a suspected
interaction between nematodes and root-iot fungi
has been investigated in Hawaii. In Natal, Roth found
a fungus, Mortierella sp., in sugarcane roots in which
infestation by Meloidogyne javanica had produced
typical nodules. Possible association between these
two organisms was investigated by planting sugarcane
cuttings in replicated pots of sterilised soil, untreated
or after the addition of pure cultures of Meloidogyne,
of the fungus, or of both. Aftei live months, when cane
in some of the pots had begun to be root-bound, the
shoots which had been produced were cut and weighed
Statistical analysis of the results (Table II.) indicated
a highly significant reduction in growth due to the
nematodes. The fungus had produced no observable
effect and there was no significant intereaction.
Examination of the roots showed that nodules occurred
only in pots inoculated with nematodes; they
were not produced in the presence of the fungus
alone. The co-existence of these two organisms in the
same roots appears to be fortuitous.
Table II
Nematode-Fungus Trial
Total weight in grams of six replicates
No Fungus
Fungus (waxed)*
Fungus
Total
Control
2509
2620
2612
7741
Nematode
1744
1653
1495
4892
Total
4253
4273
4107
12633
*The ends of cuttings were coated with wax to
prevent direct infection by the fungus.
Tests with Nematocides
In Natal, twelve field trials involving soil fumigation
have now been completed. In most cases the fumigant
used was EDB, since this was the material most
readily available at the start of investigations, but the
few trials using DBCP and D-D gave no evidence
that results with these were significantly different.
Early tests indicated that 40 gallons per acre of a
solution containing 2.25 pounds per gallon, or 20
gallons per acre of one containing 4.5 pounds per
gallon of EDB were approximately the correct amounts
to apply, and most subsequent trials were carried out
at these rates, which correspond to 90 pounds of
active ingredient per acre.
Even at rates of application high enough to be
completely uneconomic, nematocides cannot entirely
eliminate plant parasitic nematodes from the soil
although treatment generally reduces their numbers
very significantly. In our trials, mortalities have
varied from less than 60 per cent to about 90 per cent.
Reproduction by survivors leads to recrudescence\
so that populations eventually equal or exceed those
in untreated plots. The time required for this to happen
varies considerably, ranging from under a year
to over two years in our experiments. In Australia,
intervals as short as six months have been recorded. 1 "
The tendency of populations in fumigated plots
eventually to exceed those in untreated plots is thought
to be due to the vigorous root system which develops
in treated soil. This may provide conditions which
encourage rapid increase of nematodes. It may also
account for the phenomenon of residual yield responses
in ratoons which often occur even when nematode
populations have completely recovered during the
life of the plant cane crop.
In the twelve field trials which we have cairied out,
mean yield differences between untreated and fumigated
plots have varied from nil to 21 tons cane per
acre, with a general average of 9 tons. In four trials
the yield increases were statistically significant at the
one per cent level and in another four at the five per
cent level. In seven trials which wei e harvested as first
ratoons mean differences between treated and untreated
plots varied from nil to 13 tons cane per acre,
with a general average of about 6 tons. Three of these
trials showed statistically significant residual effects
resulting fiom fumigation which was applied befoie
planting.
On account of the high cost of the chemicals used,
the difficulty of applying them effectively and the somewhat
variable results obtained in experiments, it has
not been possible to recommend soil fumigation of
sugarcane fields with confidence that the results will
be economically profitable. Consequently, here as in
other countries, commercial application of soil fumigants
has been negligible.
Combined Treatments
It has often been in cane growing in pooi, sandy
fields that symptoms attributed to the presence of
nematodes have been noticed. This is paiticularly
true of infestation by Meloidogyne spp. Although soil
fumigation in such fields usually leads to significant
increases in yield, crop responses may be limited by
the nature of the soil. Several trials have therefore been
canied out to discover whethei the presence of fertilizers
and soil ameliorants might enable the plant to
take greater advantage of the period, following soil
fumigation, when nematodes aie least numerous.
Materials tested in this way have included inorganic
fertilisers at various levels, magnesium sulphate,
molasses, compost and filtercake. Some of these,
notably fertilizers and filtercake, led to significant
ciop increases on their own, but no significant interaction
between these and the effects of fumigation was
demonstrated. In two trials on Nkwali Flats, Illovo,
significant reductions in numbers of parasitic worms
followed applications of filtercake (Table III) but in

330 Proceedings of The South African Sugar Technologists' Association—March 1966
other trials yield responses to this material may
have been due to its value as a plant food.
Table III
Nematodes in root samples, 12 months after treatment
Filtercake rate
Nil
20 tons/acre
40 tons/acre
Total
No
Fumigant
1852
1237
905
3994
Other Chemicals
EDB
1631
736
750
3117
Total
3483
1973
1655
7111
In addition to standard soil fumigants, several other
chemicals have been tested for nematocidal properties,
both in the laboratory and in the field. In pot
tests, tomatoes were used as indicator plants and the
numbers of nodules produced on their roots by
Meloidogyne spp. were used to assess the results. A
calcium polysulphide preparation was effective in pot
tests but, possibly on account of method of application,
did not improve the yield of sugarcane in the
field. Similarly, calcium cyanimide, which has been
recommended for nematode control on some other
crops, was effective in pot tests but did not influence
nematode populations or cane yield in field trials in
which EDB led to highly significant results. Ammonium
hydroxide, in several pot tests, caused very
significant reduction in numbers of nodules and noticeably
stimulated the development of tomato plants.
Urea, at equivalent nitrogen rates produced a similar
but less noticeable effect (Table IV). Although it has
400 lb. N.
200 lb. N.
Nil
Table IV
Mean number of nodules on 4 tomato roots
Ammonium
Hydroxide
2.0
16.7
98.8
Urea
15.5
67.1
105.1
not yet been possible to plant field trials with the
specific object of testing the effect of different forms of
nitrogen on nematode populations, advantage has
been taken of an offer by a fertiliser company to
place soil samples from a field trial, in which anhydrous
and aqueous ammonia were included as
treatments, at our disposal for nematode diagnosis.
In these samples (Table V) counts varied so considerably
that no statistical significance was found but,
especially for Meloidogyne, the totals showed differences
suggesting that further investigation might be
justified.
Table V
Numbers of nematodes in soil samples six months
after treatment
Control
Limestone Ammonium Nitrate
Sulphate of Ammonia . . . .
Urea
Anhydrous Ammonia . . . .
Aqueous Ammonia
Meloidogyne
191
165
215
172
31
50
Resistant cover crops
Pratylenchus
173
150
263
282
224
99
Hoplolaimidae
32
69
55
37
43
45
Another form of nematode control which has been
attempted involves the cultivation of resistant cover
crops. Up to the present, only one species has been
investigated under our conditions, namely the Ermelo
strain of Eragrostis curvula. A small trial plot of this
grass was planted on a field near Empangeni, where
heavy infestations especially of Meloidogyne and
Pratylenchus had been observed. After four years,
parasitic nematodes had practically disappeared from
this plot and the grass was ploughed out and sugarcane
replanted. As this was not a replicated trial,
valid yield comparisons could not be made but, when
inspected about a year later, the cane appeared considerably
better grown than that in surrounding
fields. The main disadvantage of this type of control
measure is the lengthy period during which the field
is out of cane production. It is unlikely to be adopted
commercially under normal conditions.
Nematode — Variety Trials
In two trials, the reactions of different varieties of
sugarcane to nematodes has been investigated by
planting them in fumigated and untreated plots.
It was reasoned that varieties which did not respond
to treatment might be assumed to be resistant or, at
least, tolerant. On the other hand, those which
responded to the greatest degree would presumably
prove to be the most susceptible. Results obtained
for the plant cane crop in these trials are given in Tables
VI and VII. Both showed differences due to fumigation
and to variety. The interaction between these
factors was not statistically significant and ranking
N:Co293
N:Co382
Co 331
N:Co 292
N:Co 339
N. 50/211
N:Co 376
N:Co 334
Table VI
EDB
54.36
60.29
40.88
51.68
56.42
56.63
56.85
50.92
Tons cane per acre
Nil
38.71
41.57
25.01
30.50
32.66
32.33
32.23
24.37
Ratio
1.40
1.45
1.63
1.69
1.73
1.75
1.76
2.09

Proceedings of The South African Sugar Technologists' Association — March 1966 331
of the varieties in these trials in order of apparent
tolerance would therefore not be justified. However,
N:Co.382, which shows a small ratio between treated
and untreated plots, gives reasonably good crops in
infested sandy soils while some varieties, at the other
end of the list, do not. Without doubt, investigations
along these lines should continue, since the discovery
and cultivation of tolerant variety would be a far more
satisfactory solution to the nematode problem than
treatment with nematocides.
N :Co 293
N.50/211
N.51/168
N. 52/219
N.10
N.51/539
Table VII
Tons cane per acre
DBCP
69.3
63.7
38.0
55.7
56.0
40.8
Acknowledgments
Nil
60.1
54.7
38.0
46.9
36.5
22.1
Ratio
1.15
1.16
1.00
1.19
1.53
1.85
Indebtedness for co-operation is gratefully acknowledged
to Dr. J. Heyns and Dr. H. Koen of the Plant
Protection Research Institute, Pretoria, for identifications
and advice on technique, to the staff of Kynoch
Ltd., for soil samples from nitrogen trials, and to
Dr. G. Roth of the S.A.S.A. Experiment Station for
preparing pure cultures of the fungus, Mortierella
sp., used in experiments.
80,
70.
60.
5G'
40
30:
20
10-
Monthly Counts Of Eslworms In Cane Roots.
Figures represent means of 12 subsamples
S. O, N D J F M A M J J A S O
Harvested
Summary
The nematodes known to occur in South African
sugarcane fields are listed and their status discussed.
Experiments have shown that Meloidogyne javanica
can affect cane growth, but other species are probably
involved as well. The possible association between
nematodes and organisms causing disease is mentioned
and an instance is quoted in which no interaction
between Meloidogyne javanica and a fungus,
Mortierella sp., could be found. Results of trials with
nematocides are summarised. Although highly significant
responses are often shown, sometimes in ratoons
as well as plant cane, the economic value of soil fumigation
is still in doubt. Attempts at increasing responses
to fumigation by applications of fertilizers
and soil ameliorants gave inconclusive results, although
filter cake produced some effect on nematode
populations. Among other materials tested, ammonia
gave results suggesting that further investigations
would be justified. Eragrostis curvula, as a cover crop,
succeeded in controlling an infestation consisting
mainly of Meloidogyne and Pratylenchus but was too
slow to be of practical value. Variety trials gave results
in which, although interaction between fumigation and
variety was not statistically significant, differences in
behaviour suggested that continued investigation
should be undertaken.
Reference
1. Anon. (1959). Effects of nematodes and fungi, singly and
in combination, on the growth of sugar cane. Rep. Hawaiian
Sug. Plrs. Ass. Exp. Stn. 1959: 18-20.
2. Anon. (1962). Effect of nematodes shown. Ibid., 1962: 51
3. Apt, W. J. and Koike, H. (1962a). Pathogenicity of Helicotylenchus
nannus and its relation with Pythium graminicola
on sugar cane in Hawaii. Phytopathology 52: 798-802.
4. Apt, W. J. and Koike, H. (1962b). Influence of the stubbyroot
nematode on growth of sugar cane in Hawaii. Ibid.,
52: 963-964.
5. Apt, W. J. and Koike, H. (1962c). Pathogenicity of Meloidogyne
incognita acrita and its relation with Pythium
graminicola on sugar cane in Hawaii. Ibid., 52: 1180-1184.
6. Birchfield, W. and Martin, W. J. (1965). Pathogenicity on
sugar cane and host plant studies of a species of Tylenchorhynchus.
Phytopathology 46: 277-280.
7. Dick, J. (1961). Eelworms and sugar cane. Proc. S. Afri.
Sug. Technol. Ass. 35: 110-113.
8. Khan, S. A. (1963). Occurrence and pathogenicity of
Pratylenchus zeae on sugar cane in Louisiana. Proc. Int.
Soc. Sug. Cane Technol. 11: 711-717.
9. Martin, J. P., Wismer, C. A., Koike, H. and Apt, W. J.
(1960). Some biological factors associated with yield decline
of sugar cane varieties in Hawaii. Proc. Int. Soc. Sug. Cane
Technol. 10: 77-84.
10. Vallance, L. G. (1960). Division of Soils and Agronomy.
Rep. Bur. Sug. Exp. Stns. Qd. 61: 16-27.
Mr. du Toit (in the chair): Referring to Table III,
I am surprised that EDB did not have more effect on
the nematodes.
Dr. Dick: These figures were from counts a year
after treatment, and nematode populations had probably
recovered from the effects of fumigation. It
does look as though the response to filter cake has
been more-persistent.

332 Proceedings of The South African Sugar Technologists' Association—March 1966
Mr. Gilfillan: Dr. Dick said there are difficulties in
applying fumigants to the soil but I do not see why.
For example, liquid ammonia can be applied in the
field.
We have recently applied fumigants to the soil at a
cost of RIO per acre and an increased yield of 5 tons
per acre.
He also says the application of fumigants might
stop the nitrification process in the soil but surely
increased fertilization will overcome this?
Dr. Dick: What was meant by difficulty was not
merely the mechanical problem of getting the nematocides
into the soil. For the best results, soil temperature,
moisture status and tilth must satisfy somewhat
circumscribed requirements.
Inhibition of nitrification is not mentioned in the
paper. The action on micro-organisms is quoted
merely to demonstrate that nematode control is not
the only effect of nematocides.
Mr. R. A. Wood: The fumigant will affect the total
amount of nitrogen that is mineralised probably due
to a partial sterilisation eifect on certain microorganisms
in the soil. It has been clearly shown where
fumigants have been used on tobacco in Rhodesia
that nitrification in the soil was reduced.
In sandy soils the ammonium form of nitrogen is
retained longer than the nitrate form and a larger
response on the soil can be expected.
Mr. Johnson: In a trial carried out at Hippo Valley
using methyl-bromide a definite suppression of
nitrification occurred with a reduction in germination
and cane yield. The nematocides, as such, had no
effect on cane yield.
It is all rather confusing and I think Dr. Dick is
right when he says that until we know which of these
nematocides affect the cane roots adversely we cannot
make much progress.
Mr. Aucock: Does Mr. Gilfillan's price of RIO per
acre include chemicals and application?
Mr. Gilfillan: It was for chemicals only—for one
gallon of active ingredient.

Proceedings of The South African Sugar Technologists' Association—March 1966 333
THE PRODUCTION OF TRASH AND ITS EFFECTS AS A
MULCH ON THE SOIL AND ON SUGARCANE NUTRITION*
Introduction
Trash conservation is practised primarily to obtain
a reliable means of weed control in ratooa crops of
sugarcane, and in Natal it is done with the knowledge
that a crop yield response may also be obtained, due
largely to moisture conservation in an area where
rainfall is severely limiting to crop production.
However, in considering the conditions which might
be associated with a response to mulching compared
with burning, it is essential to recognise that each
response might be the sum of several positive effects,
and even the net result of both positive and negative
effects. The production and composition of trash have
therefore been studied, and long-term burning v.v.
trashing experiments have provided a means for
measuring the effects of trash on the soil and crop
nutrition.
Trash Production
The progressive accumulation of trash was measured
in a growth analysis experiment where variety
N:Co.376 was harvested from six replications at
weekly intervals. Additional plots, harvested at
monthly intervals, were used to study the effects of
other treatments which included varieties N:Co.310
and N:Co.382, variety N:Co.376 planted at 1 ft.
6 in. row spacing, variety N:Co.376 irrigated, and
variety N:Co.376 without nitrogen fertilisation. The
trash consisted of all dead leaves and the immature
cane tops above the point of attachment of the sixth
sheath on the stalk, and these two components were
weighed and sampled separately at each time of
harvest.
It is usual to refer to the amount of trash produced
per acre in terms of the amount of harvested stalk.
This convention has the practical convenience that
the weight of trash can readily be estimated from the
sugarcane yield, but the value of the information in
the field is invariably limited, since the moisture content
of the trash material is not known. The moisture
contents of the representative samples of trash were
therefore determined, and the amounts of oven-dry
trash produced per acre at various stages of crop
development in the different treatments are shown in
Table I.
Table I
by G. D. THOMPSON
For variety N:Co.376, produced in six replications
under standard dryland conditions, the amount of
oven-dry trash increased from 3.9 tons per acre at
5 months of age to over 8 tons per acre at 20 months
of age. The amount of millable cane increased from
10.6 tons at 5 months to 57.3 tons at 20 months.
Approximately a two-fold increase in the amount of
trash produced was thus accompanied by a five-fold
increase in millable cane.
A rapid increase of the millable cane, oven-dry trash
ratio over the period from 5 months' to 10 months
of age was common to all varieties and treatments,
but the absolute values of the ratios appeared to vary
significantly, and to maintain different levels during
the period of gradual increase after the crop was 10
months old. Varieties N:Co.376 and N:Co.382 were
fairly similar, as shown in Figure 1, but for variety
N:Co.310
. . N:Co.382
. « N:Co.3?6
8 .
5 ' 6 ' 7 ' 8 ' 9 'ID 'll'ia'13 ' lV 15' l6'W 'l8 'W 20'
Age of crop in months
FIGURE I : Ratio of millable cane to oven dry trash for three
varieties of sugarcane.
N:Co.310 the ratio was low throughout the duration
of the crop. The omission of nitrogen fertilisation did
not affect the cane/trash ratio, both vegetative components
of the crop simply being produced in lesser
quantities. Irrigation, however, increased the cane
trash ratio whilst closer spacing of the cane rows
caused the ratio to be lower.
It is clear that no single ratio can be used to estimate
dry trash from the amount of harvested cane since
variety, age and cultural treatment all affect it. However,
for variety N:Co.376 under dryland conditions
at 4 ft. 6 in. spacing after 12 months of age, a ratio
of 6 : 1 may be reasonably accurate.
Nutrient Content of Trash
In comparing the two most common field practices
This paper contains mainly material used for post-graduate work burning in the Department and trash of blanketing, Crop Science, for University
the nutation of
the succeeding ratoon crop, it should be appreciated

334
that burning probably causes only a loss of nitrogen,
the mineral constituents of the trash remaining largely
in the ash left on the land after burning. On the other
hand there may be some advantage to be gained from
the gradual release of nutrients from decomposing
trash in contrast to the rapid return to the soil which
is effected by burning. If the singed tops remain on the
land after burning there may still be a significant
conservation of nitrogen which is held in the apical
meristem in high concentrations.
Translocation of nutrients from the ageing leaf
tissue to the vigorously growing parts of the sugarcane
plant is known to occur, so that the dead leaf
tissue is not as rich in nutrients as the green leaves.
Furthermore, nutrients are leached from the dead
leaves and re-enter the soil even before the crop is
harvested. The composition of the trash left in the
field after harvesting therefore does not represent
either the total return of nutrients to the soil from the
above-ground parts of the previous crop, nor the
entire difference in nutritional effects between burning
and trashing. It is of interest primarily as an indication
of the approximate amounts of nutrients conserved
in this manner, and as a measure of the potential
loss of nitrogen when burning is complete. The
amounts of nutrients contained in the trash residue
from 40 tons of cane per acre in the growth analysis
experiment are shown in Table II.
Table II
Nutrients in trash from 40 tons of cane per acre for various
treatments
Treatment
N:Co.310
N:Co.382
N:Co.376
N:Co.376, irrigated . . . .
N:Co.376, 1 ft. 6 in. spacing .
N:Co.376, no nitrogen
N
lb
85
71
69
81
100
45
P
lb
4
8
8
9
12
7
K
lb
83
99
76
149
136
70
Proceedings of The South African Sugar Technologists' Association—March 1966
Mg Ca
lb lb
37 45
28 35
27 38
34 41
29 52
26 26
Both nitrogen and potassium were returned to the
soil in considerable quantities, and thorough burning
would have resulted in a loss of nitrogen which could
have affected the nutrition of the succeeding crop.
The mean yield from six replications of N:Co.376
at 19 months of age was 55.0 tons of cane per acre,
and the amounts of nutrients in the total trash per
ton of cane were as follows:
The results from the growth analysis experiment
are compared in Table III with data obtained by Golden
and Ricaud (1963) in Louisiana in particular, and
average results from other parts of the world. The
local results were characterised mainly by the low
amounts of P in the trash, despite a standard fertilisation
of 800 lb. superphosphate per acre in the furrow
at the time of planting.
Effects of mulching on the soil and on crop nutrition
These effects were studied mainly in thiee burning
v.v. trashing experiments.
Experiment I: Four replications of a split plot
experiment were planted in October, 1939 on a
Rydalvale clay loam at Mount Edgecombe to compare
four treatments:
(i) plots burnt, mixed NPK fertiliser applied
(ii) plots burnt, no fertiliser applied
(iii) plots trashed, mixed NPK fertiliser applied
(iv) plots trashed, no fertiliser applied.
The experiment was continued uninterruptedly
through three successive cycles, each cycle consisting
of plant cane and three ratoon crops.
Experiment II: Planted in October, 1947, this
experiment was designed to compared three trash
treatments in six replications:
(i) trash blanket
(ii) trash burnt, tops removed from plots
(iii) trash burnt, tops spread on plots.
Sub-plot treatments consiste of three fertiliser
treatments:
(i) P with one level of N (NP)
(ii) P and K with one level of N (NPK)
(iii) P and K with two levels of N (NNPK).
The site was on a Waldene fine sandy loam at Chaka's
Kraal and the experiment was conducted through a
cycle of a plant crop and two ratoons, followed by a
plant crop and five ratoons.
Experiment HI: This experiment was designed to
study the effects of different amounts of trash on
sugarcane yields and did not include any burning
treatments. Trashing was therefore made a sub-plot
factor in a split plot design with two levels of N
(120 lb. and 240 lb. per acre) as the whole plot treatments,
and there were six replications. The trash
treatments were 0, 7, 16 and 25 tons of trash per acre.
The experiment was planted in July, 1955 on a Waldene
fine sandy loam at Chaka's Kraal, and was continued
through four ratoon crops.
Soil Nutrients
Experiments I, II and III were soil-sampled intensively,
plot by plot, during 1962. The samples
were taken by 3-inch strata down to 12 inches and
were analysed for available P, K, Ca and Mg. The
results for the 0-3 inch stratum are shown in Table IV.

Proceedings of The South African Sugar Technologists' Association—March 1966 335
Tabic IV
Available nutrients in the surface 3 inches of soil in Experiments I,
II and III
Even after 23 years of treatment in Experiment I
the soils in burnt plots showed no depletion of major
mineral elements compared with the soils from trashed
plots. The effects of P and K fertilisation were apparent,
and the much heavier crops removed from the
fertilised plots presumably caused the significant
depletion of Mg in the soils in these plots compared
with those in unfertilised plots.
No marked differences in available soil nutrients
were apparent in Experiment II after 15 years of treatment,
except for an apparent depletion of magnesium
where the burnt tops were removed from the plots.
In Experiment III the increasing amounts of trash did
not cause significant differences in the amounts of
available P, K and Ca in the soil, but the amounts of
available Mg increased with successive trash increments.
The analyses of soil samples from the 3-6 in.,
6-9 in., and 9-12 in. strata were similar to those of the
0-3 in. stratum samples, but the variations in available
Mg content tended to decrease with increasing soil
depth.
Total exchange capacity, total cations and pH
The soil samples from 3-inch strata down to 12
inches in Experiment 1 were analysed for total exchange
capacity, total cations and pH. The results
are shown in Table V and. although differences were
not always significant, it was apparent that trashing
increased the total exchange capacity of the soil
compared with burning. Trashing also caused an
Table V
Mean data for total exchange capacity, per cent base saturation
and pH in successive soil strata in Experiment I
increase in acidity, reflected both in lower pH's
and lower per cent base saturation of the soils from
trashed plots. The data for similar samples from
Experiments II and III showed the same consistent
effect of trash in lowering soil pH.
The increase in total exchange capacity of the soil
in Experiment I tended to be greater in the presence
than in the absence of fertiliser. This effect probably
derived mainly from the greater amounts of trash in
the fertilised plots, since the results for Experiment
HI showed that the total exchange capacity of the soil
increased progressively with increasing amounts of
trash.
Experiments II and III were both sited on Waldene
fine sandy loam soils at Chaka's Kraal. A third experiment
involving burning and trashing treatments
was planted on this same soil series in 1959, and soil
samples from the 0-1 in. stratum were taken from all
three experiments in 1964. The mean data for the
total soil exchange capacity, comparing treatments
without trash and treatments with the maximum
trash blanket, were as follows:
Experiment II: an increase from 7.05 to
9.14
= 2.09 meq/100g soil after
17 years.
Experiment III: an increase from 6.98 to
8.55
= 1.57 meq/100g soil after
9 years.
Additional experiment: an increase from 6.53 to
6.95
= 0.42 meq/lOOg soil after
5 yeais.
There was thus evidence that the effect of trash increased
with time, which was possibly a fuither
expression of the effect of amount of trash.
Available soil moisture
Undisturbed soil cores, 4 in. in diameter and 3 in.
deep, were taken from the surface soil of all plots in
Experiments II and III in July, 1964. Field capacity
was determined at 1/3 atmosphere tension in the laboratory
on these samples and wilting point was determined
at 15 atmospheres tension on disturbed samples taken
from the same soil depth. The results are shown in
Table VI.

336 Proceedings of The South African Sugar Technologists Association—March 1966
The effects of trash mulching on both field capacity
and wilting point were significant in Experiment II,
but in these Waldene series soils there was no apparent
change in available water held between these limits.
Tension curves were not established for any of these
soils and it is therefore not possible to comment on
the effects of trash treatments on the relative availability
of water held between 1/3 and 15 atmospheres
tension.
Undisturbed cores and disturbed samples were also
taken from the surface 3 in. of a Rydalvale clay loam
soil in plots adjacent to Experiment I in 1963
and 1964. In the absence of any crop, eight plots had
been trash-covered and eight plots had been kept bare
for 30 months. The mean results of laboratory determinations
of field capacity at 3 different tensions and
wilting point at 15 atmospheres tension, after 15
months and after 30 months of treatment, are shown in
Table VII. The effects of treatment were highly signi-
ficant on all three moisture characteristics at both
times of sampling. The large differences in mean available
water content on the two occasions was difficult to
explain, and it is of interest that Salter and Williams
(1963) found similar effects, due to season, in soils
with and without farmyard manure. From December,
1960 to August, 1961, the available water per unit
volume of soil had changed from 1.05 to 0.59 in the
untreated soil, and from 1.32 to 0.89 in the treated
soil. These authors also determined that the additional
available water in the manure-treated soil was advantageously
held at low tensions.
Soil organic matter
The variations in organic matter content in 3-inch
strata of soil down to 12 inches depth in Experiment
II are shown in Figure 2. The difference between burning
and trashing treatments was significant in the 0-3
in. stratum, but not at other depths. The patterns of
variation in Experiments I and III were similar to
those in Experiment II, the organic matter contents
of the soils decreasing with depth, and the trashmulched
soils containing more organic matter than
the unmulched soils.
A summary of the average organic matter contents
of soil samples taken from Experiments II and III to a
depth of only 1 inch in 1964 is shown in Table VIII.
As was to be expected, the effects of trash were much
more marked in samples from the shallow surface
layer of soil, and it may be noted that the effects in
Experiment III were highly significantly linear over
the range from 0 to 25 tons of trash per acre.
Soil Crumb Structure
The water stability of soil crumbs greater than 0.5
mm. in diameter was determined by the method of
Beater (1962) in samples from all 3-inch strata down
to 12 inches in Experiment I and to 6 inches in Experiment
III. The results for Experiment I are shown
in Table IX, and it is apparent that, whilst water
stability of soil crumbs was always greater for trashed
plot soils than for burnt plot soils, the condition was
most exaggerated in the 0-3 in. depth. The marked
increase from the 0-3 in. stratum to the 3-6 in. stratum
under burnt conditions, in contrast to the slight decrease
over the same depths under trash, may be
construed as evidence that the effect of the mulch
treatment was to protect the water stability of crumbs
more than to create water stability. The differences
at 6-9 in. and 9-12 in. depths may have been due to
improved conditions under trash in the current cycle,
but the effects of soil inversion at previous ploughouts
cannot be ignored.
The results of the analyses for the Waldene soil
in Experiment III are shown in Table X. The vast
inferiority of the Waldene series compared with the
Rydalvale series in terms of water stability of crumb
structure was shown in the values for total percentage
of water stable aggregates greater than 0.5 mm. in

Proceedings of The South African Sugar Technologists' Association- -March 1966 337
diameter in the surface 3 in. of the two soils. The
Waldene average was 4.5 per cent and the Rydalvale
39.0 per cent. The results for the soils from Experiment
III showed further that the water stability of
Soil Bulk Density and Porosity
The effect of trash mulching on the bulk density
of the soil was studied on five separate occasions.
Two-inch diameter undisturbed cores were taken from
the surface soil of the interrow of each plot of Experiment
I in 1961. The mean bulk densities for
treatments were 1.14 g/cc in the trashed plots and
1.16 g/cc in the burnt plots, the difference not being
significant. The corresponding porosities were 53.9
and 53.1 per cent respectively.
Four-inch diameter cores were taken from Experiments
II and III during 1964 and the bulk density
results are shown in Table XI. Therewereno significant
differences due to treatments and no trends were
apparent.
Bulk densities were also determined on the undisturbed
core samples taken from the plots adjacent
to Experiment I on a Rydalvale clay loam in 1963
and 1964. No significant differences due to trash
mulching were apparent, the mean bulk densities for
the bare soils being 1.06 and 1.04 g/cc and for the
mulched soils 1.09 and 1.07 g/cc in. 1963 and 1964
respectively. The average porosities of the bare and
aggregates increased consistently with amount of
trash, and also tended to confirm the observation that
mulching protected water stability of crumbs.
mulched soils in 1963 were 53 per cent and 54 per
cent respectively.
Soil Temperature
The effects of a trash mulch on soil temperature in
the presence of a crop were studied in a ratoon of
N:Co.310 on a Rydalvale clay loam. Of the area
selected for this work, half was harvested in July,
1962, and the remainder left unharvested when the

338 Proceedings of The South African Sugar Technologists' Association—March .1966
standing crop was approximately a year old. The two
halves were then further sub-divided, one quarter
being treated with a trash blanket after harvesting,
a second quarter being cleared of trash after harvesting,
a third quarter being given a fresh trash blanket
although unharvested, and the fourth quarter being
cleared of all trash although unharvested. Soil
temperatures were read daily from thermometers
placed midway between the centre of the row and the
centre of the interrow at 1 in., 4 in. and 9 in. depths.
The results for the 1 in. depth at 8 a.m. in terms of
five point moving averages of weekly means are presented
in Figure 3.
FIGURE 3 : Five point moving averages of mean weekly soil
temperatures at the I in. depth at 8 a.m. for four treatments
The most marked, effect by far was the high soil
temperature in the absence of trash where the cane
had been harvested. Differences between bare and
mulched soil temperatures were maintained at mean
values of about 4° C during the spring period when
the cane canopy was developing in the harvested
plots, but from January onwards when the canopy
was well developed, the differences became small.
After February, in fact, the mulched soils had a
higher temperature at the 1 in. depth at 8 a.m. than
the unmulched soils, due indubitably to the combined
effects of mulch and canopy in reducing nocturnal
longwave back radiation. The same effect was apparent
from the data for unharvested plots, although
to a much smaller degree. Only for the five-month
period from October through March was daily
insolation sufficiently great on the unmulched plots
for the 8 a.m. temperatures at the 1 in. depth to exceed
those of the mulched soils. Thus in winter the
effect of the mulch generally was to cause early
morning temperatures to be slightly higher than those
in plots with bare soil.
The 3 p.m. data for the same plots at the 1 in. depth
are shown in Figure 4. Mean differences in temperature
between mulched and unmulched soils at this time of
day in the harvested plots reached more than 9° C
in spring and summer, and the differences were apparent
throughout the period of measurement, although
a sharp decline developed in February.
Distinctly higher soil temperatures in the unharvested
FIGURE 4 : Five point moving averages of mean weekly soil
temperatures at the I in. depth at 3 p.m. for four treatments
plot without trash compared with those in plots with
trash indicated, that soil insolation occurred even
through a well-developed cane canopy. The presence of
the canopy, however, reduced the temperature difference
between treatments to about 2" C at 3 p.m.
An additional feature apparent in Figures 3 and 4
was that the trash mulch tended to smooth out the
intraseasonal soil temperature variations both at
8 a.m. and at 3 p.m.
The patterns of soil temperature change at the 9 in.
depth were similar to those at the 1 in. depth, but
the temperatures at the 9 in. depth at 8 a.m. were
higher than those at the 1 in. depth except for the
harvested treatment without trash for the 3-month
period from September to December, when the
temperatures at the 1 in. depth were very slightly
higher than those at the 9 in. depth. The 3 p.m. data
for the 9 in. depth were much lower than the comparable
1 in. depth data throughout, particularly for the
harvested plots without trash. The important feature
of these results was that even at 3 p.m. the soil temperatures
at the 9 in. depth in the trash mulched plots
exceeded those in the unmulched plots from early in
March onwards.
The mulch treatments reduced diurnal soil temperature
variations markedly. The mean daily range of
temperature in the unharvested plots without a trash
mulch was reasonably close to 2° C from July through
to April, and this was regarded as the effect of a fairly
complete leaf canopy alone. In contrast, the mean
range varied from 4" C to more than 6° C on the
harvested plots without trash during the period of
canopy formation, and then dropped to a comparable
2 o C by the end of the summer. In the presence of
trash, however, the mean daily variation tended to
be less than 1° C whether or not a cane leaf canopy
existed. Thus the unmulched soils always received
more energy than the mulched soils during the day,
but the amount of back radiation from these same
exposed soils was so great in March and April that
the maximum soil temperatures during the day were
lower than those in the mulched soils.
The effects of different amounts of trash on soil

Proceedings of The South African Sugar Technologists' Association—March 1966 339
temperatures in the absence of a crop were studied on
a Waldene series soil where small plots were covered
with nil, 1 ton, 2 tons, 4 tons, 8 tons and 16 tons of
trash per acre. On 29th September, 1964, the weather
was cloudy and. cool and the results of temperature
measurements at the 5 cm. depth are shown for only
three treatments in Figure 5.
The effective protection against nocturnal back
radiation by 16 tons of trash per acre was apparent
since the temperature of the soil in this treatment was
higher than those in plots with no trash and 2 tons
of trash per acre at 8 a.m. Within the ensuing hour,
however, insolation on the bare soil had reversed the
situation, but the reversal did not occur until 11 a.m.
between the 2 tons and 16 tons of trash per acre
treatments. The results at the 10 cm. depth shown in
Figure 6 indicated that even the 2 tons trash per acre
treatment conserved heat at this depth in excess of
that conserved in the plot without trash.
It must be acknowledged that soil moisture conditions
under the different treatments were inevitably
different and that the addilional heat capacity of the
moister soils under trash precluded any quantitative
evaluation of the results in terms of an energy balance.
Generalisations based on the data, even in terms of
temperature effects, are therefore made with this
reservation in mind.
The weather on 7th October, 1964, was warm with
only intermittent cloud, and the clearly defined effects
of five levels of trash mulch, compared with a control
treatment without trash, under these conditions are
shown in Figure 7 for the 5 cm. depth and in Figure
8 for the 10 cm. depth.
The collective results from this site on a Waldene
series soil showed that the post meridiem decline in
soil temperatures at shallow depths occurred earlier
as the total insolation increased and the mean temperature
rose, and that the decline was postponed by
increasing amounts of trash mulch. In general it can
be said that in the absence of a crop, the soil temperature
in bare plots always rose above that in mulched
plots during the day, even in midwinter, at the 5 cm.
and 10 cm. depths. This was in contrast to data

340
obtained from Experiments 1 and III, on Rydalvale
and Waldene soils respectively, where the presence of
a crop canopy sometimes caused the reverse to be
true.
Crop Nutrition
If a trash mulch increases crop yield by contributing
to the nutritional status of the crop, an interaction
between fertiliser and trash effects should be found.
More particularly, the reponse to trash mulching
compared with burning or removing the trash should
be greater in the absence than in the presence of
fertilisation. Further, if a nutrient derived from a
trash mulch contributed to a positive response to the
treatment, an increase in the amount of the particular
nutrient in a plant indicator tissue might be expected.
The yield data for three crop cycles in Experiment I
have been reported previously (Thompson, 1965).
The mean response to trash mulching compared with
burning over nine ratoon crops was 4.61 tons of
cane per acre per annum in the absence of fertiliser and
4.06 tons of cane per acre per annum in the presence
of fertiliser. Whilst these mean results might be construed
as evidence of a slight nutritional effect of
trash, the single instance in which a significant interaction
was obtained, occurred under unusual conditions
in the second ratoon stage of the third cycle.
In this crop a severe drought caused excessive stalk
mortality in the trashed, fertilised plots without a
similar effect occurring in the trashed, unfertilised
plots. This phenomenon was undoubtedly due to the
much higher consumptive use of water in the fertilised
treatment compared with the unfertilised treatment,
and the results cannot therefore be regarded as direct
nutritional effects. In fact, when the results for this
crop are excluded from the mean data for Experiment
I, the response to trash treatment appears to be
slightly greater in the presence of fertilisation than in
its absence.
In Experiment II the differences in yield between
trash treatments (i) and (iii) represent the response to
a trash blanket compared with burning and leaving
the tops spread on the plots. The mean yield differences
over seven ratoon crops for the three different
fertiliser treatments were:
NP : 0.31 tons cane per acre per crop.
NPK : 1.18 tons cane per acre per crop.
NNPK : 0.61 tons cane per acre per crop.
The only occasion on which the interaction between
trash and fertiliser treatments was significant was in
the first ratoon stage of the second cycle, and the
effect was the reverse of that required to indicate a
nutritional contribution from trash. The comparisons
of average data indicate, if anything, that trash
improved the efficiency of potassium top dressings
but not necessarily the nitrogen top dressings.
The design of the fertiliser treatments in Experiment
III was based on the assumption that trash,
being highly carbonaceous, would cause high C/N
ratios to develop in the soil and consequently nitrogen
deficiencies. No significant interactions between
fertiliser and trash treatments occurred, but it was
possible that the lower nitrogen level of 120 lb. N
Proceedings of The South African Sugar Technologists" Association—March 1966
per acre was sufficient alone to prevent any serious
deficiency from developing. Table XII shows the mean
responses to 7 tons of trash per acre compared with
no trash (T2-TI), and also the mean responses to
25 tons of trash per acre compared with 7 tons per
acre (T4-T2).
The consistently greater response to trash at the
higher level of N (240 lb. per acre) when, comparing
7 tons of trash with no trash may be some indication
that nitrogen nutrition was inadequate at 120 lb.
N per acre to overcome the deleterious effect of a high
C/N ratio in the shallow surface stratum of soil.
The average results, when comparing the effects of
25 tons and 7 tons of trash per acre apparently
warrant a different explanation since the trend was in
the opposite direction. A depression in yield due to
the higher amount of trash was on the average
greater at the N2 level than at the Nl level. If the C/N
ratio was a real consideration, therefore, it must have
been relatively independent of the amount of trash
mulch.
Leaf and sheath samples were taken from all of the
plots in Experiment I at monthly intervals during
the second ratoon crop of the third cycle. Sheath
moisture levels for all treatments were in excess of
80 per cent on 9th April, 1962, and 11th March, 1963
when the crop was 9 months and 20 months old.
On these dates it was found that third leaf nitrogen
in samples from the trashed plots always exceeded
that in burnt plot samples. In all instances but one
the higher nitrogen levels in trashed treatments
compared with corresponding burnt treatments were
associated with higher tissue moisture contents.
There must therefore be at least a suspicion that the
apparent nitrogen effects were only indirect, being
due not necessarily to better nutrition but to the better
moisture conditions undei trash.
No significant effects of trashing compared with
burning were established for third leaf contents of
P, K, Ca, Mg, Cu, Mn or Zn.
Discussion
It is generally agreed that adequate weed control
in a ratoon can be effected with the trash from a crop
of 40 tons of cane per acre. For varieties N:Co.376
and N:Co.382 this amounts to 6 or 7 tons of dry
matter in trash per acre, and it is possible that the

Proceedings of The South African Sugar Technologists" Association—March 1966 341
same amount of trash could be obtained from as
little as 30 tons of N:Co.310 per acre. The yield
response to such a trash blanket, compared with burning,
may amount to 4 tons of cane per acre per annum
and it is due primarily to moisture conservation by a
trash blanket (Thompson, 1965). In comparison with
the considerable economic importance of ratoon
weed control and soil moisture conservation, the
effects of trash on the other soil properties studied
and the nutrition of the succeeding crop must be of
minor importance in the light of the experimental
results reported here.
Soil and tissue analyses have confirmed that mineral
elements are returned to the soil equally well
whether burning or trashing is practised. The more
gradual release from trash of nitrogen and other
nutrients, contained in an average trash layer in
quantities of practical significance, does not appear to
cause any significant improvement of crop nutrition
and yield. If nitrogen volatilisation is practically
complete in a well-burned crop, it is difficult to understand
how the 70 to 100 lb. of N per acre contained
in the trash could fail to improve the yield of the
succeeding crop if the trash were conserved. In this
respect, the controlled burning of experimental plots
may result in less N volatilisation than would occur
in normal field-scale fires. In any event, the raking
of the singed cane tops into a limited number of
interrows, where they control weeds without interfering
with the cultivation of the major part of the
field, can be recommended in preference to the piling
and reburning which is sometimes practised.
There is evidence to show that a trash blanket
increases the efficiency of nitrogen and potassium
top-dressings when compared with burning. This may
be an indirect effect due to better moisture relationships
in trashed areas, particularly as far as nitrogen
is concerned.
The significant increase in the exchange capacity of
soils under a trash blanket compared with those where
cane is burnt derives almost certainly from organic
colloids associated with the increased organic matter
content. No advantage apparently accrues, however,
in terms of additional adsorbed nutrient cations since
the increase in exchange capacity is devoted entirely
to hydrogen ions, causing a consistent decrease in the
pH of soils permanently under a trash mulch.
Whilst the effects of trash on surface soil moisture
characteristics were often statistically significant,
these were quantitatively so small that they were
unlikely to have been of importance in contributing
to the moisture supply to the crop. On the Waldene
series soil there could only have been a contribution if
the soil moisture tension characteristics were improved
by trash, since the increases in. field capacity
were always accompanied by increases in wilting
point. Once again, this effect of trash was most probably
associated with the increase in organic matter
content of the soil.
Increases in the organic matter content of the soil
have generally been found when the effects of organic
mulches have been studied. Samuels et a!. (1952)
compared the soil from plots which had been burned
with that from plots where the trash had either been
lined or buried. There were highly significant increases
in soil organic matter content due to treatment with
unburnt trash, the mean analytical results being 1.42,
1.82 and 1.70 per cent organic matter for the three
treatments respectively.
Any practice which tends to increase the organic
matter content of a soil must usually be advantageous
but need not necessarily be economically warranted.
In many areas of the cane belt burning is regularly
practised, and it is therefore of interest to note the
organic matter status of soils where the crops have
been burnt foi many years. In a Rydalvale clay loam
the mean organic matter content in the surface foot
of soil after 23 years of treatment was 7.65 per cent
where the crop had been trashed continuously and
7.46 per cent where the crop had been burnt. In a
Waldene fine sandy loam the mean results after 15
years were 2.67 per cent for the trashing treatment
and 2.45 per cent for the burning treatment. A significant
decrease in soil organic matter content due to
burning therefore seems unlikely to have occurred,
and an increase may well have occurred due to the
large amounts of underground organic residues from
each crop.
Although it is possible that the additional organic
matter content of the soil under trash contributed to
its greater water-stability of crumbs, the major
factor in the surface 3 in. of soil was probably the
physically protective action of a trash layer in preserving
water-stability.
The effects of a mulch on soil temperature may
obviously be complex, at times causing conditions to
be cooler than in a bare soil, and at times causing
conditions to be warmer. Under local conditions the
main interest lies in the extent to which a trash layer
causes soil temperatures in the root zone of sugarcane
in winter to be lower than those in bare soil. It is
possible that soil temperatures in winter become
limiting to root growth and crop development, and
it is an observed fact that a trash layer severely inhibits
the ratooning of some sugarcane varieties in the
high altitude areas.
According to Burr et al. (1957), root temperatures
below 21° C were severely limiting to sugarcane
growth in culture solutions. If these results can be
applied to field soil conditions they may constitute
a guide to the possibility of a mulch being deleterious
to crop growth. Referring to Figure 4 it will be seen
that at the 1 in. depth on the harvested plots, without
a mulch, the soil temperature was always above 21° C
at 3 p.m., but on the mulched plot soils this temperature
was exceeded only from November onwards. In
the unharvested plots, the critical temperature level
was passed in early October in. the unmulched soil, but
not until mid-November in the mulched soil. The
8 a.m. results in Figure 3 show a lag of two months,
from September to November, between the date on
which the harvested, unmulched soil temperature
exceeded 21" C and the date on which the harvested,
mulched soil temperature did so. The comparable
lag was only about one week in late November for
the unharvested plot soil temperatures. Even at the
9 in. depth there were lags of about 2 months in the

342 Proceedings of The South African Sugar Technologists'' Association—March 1966
8 a.m. and 3 p.m. data for the harvested plots, but
only very small lags in the unharvested plots.
It is apparent that the possible effects of trash
mulching on crop production through adverse soil
temperature conditions take place early in the development
of the crop in winter and early spring, and
this is reflected in crop performance in these seasons
in the high altitude areas.
Conclusions
Sugarcane varieties differ in the amounts of oven
dry trash produced per ton of millable cane, and the
ratio may also be affected by cultural treatment.
Millable cane : oven dry trash ratios vary generally
under local conditions between 5 and 7, but may rise
even higher under irrigated conditions. The nutrient
content of trash from 40 tons of N:Co.376 per acre
is approximately 69 lb. N, 8 lb. P, 76 lb. K, 27 lb.
Mg and 38 lb. Ca.
The nutritional contribution of a trash blanket to
the succeeding crop is of no greater commercial
importance than that following burning of the trash.
Trash mulching affects both the field capacity and
the wilting point of the soil but does not necessarily
increase the total available moisture holding capacity
of the soil. Increasing amounts of trash increase the
amounts of organic matter in the soil progressively,
and trash mulching increases the degree of waterstability
in the soil crumb fraction greater than 0.5
mm. in diameter. A mulch does not affect soil bulk
density or porosity significantly.
Trash tends to lower soil temperatures and to
reduce diurnal soil temperature variations. Increasing
amounts of trash cause greater effects, and in winter
trash may cause higher soil temperatures, particularly
in the early mornings, and to a greater extent in the
presence of a crop leaf canopy.
Summary
The effects of trash mulching compared with burning
on soil properties and crop nutrition were studied
in three long-term experiments and several additional
experiments. Trash production and the nutrient
content of trash were measured in a growth analysis
experiment.
For normally harvesttable crops of cane the ratio
of millable cane : oven dry trash lay between 5 :1
and 7 : 1 and the nutrient content of the trash was
approximately 69 lb. N, 8 lb. P and 76 lb. K per acie.
There were no detectable differences in available
soil nutrients between samples from burnt plots and
trashed plots even after 33 years of treatment. Tissue
analysis confirmed that the nutritional effects were
the same whether the trash was burned or conserved
as an organic mulch. However, trash caused significant
increases in soil organic matter, total exchange
capacity, field capacity, wilting point and waterstability
of soil crumbs greater than 0.5 mm. in
diameter. Trash did not change soil bulk density or
porosity.
Soil temperatures were sometimes lower under
trash than in a bare soil and sometimes higher,
depending upon season, time of day, soil depth and the
degree of crop canopy.
Acknowledgments
Thanks are due to Mr. J. M. Gosnell who conducted
the growth analysis experiment; Mr. R. T. Bishop who
analysed the soils and plant tissues; and Dr. R. R. Maud
and Mr. E. von der Meden who determined the soil
physical characteristics.
References
Beater, B. E. 1962. The sampling and analysis of field soils.
S.A. Sugar Assoc. Expt. Stn., Natal, 2nd ed.
Burr, G. O., Hartt, C. E., Brodie, H. W., Tanimoto, T., Kortschak,
H. P., Takahashi, D., Ashton, F. M., and Coleman,
P. E., 1957. The sugarcane plant. Am. Rev. Plant Phys. 8,
275-308.
Golden, L. E. and Ricaud, R. 1963. The nitrogen, phosphorus
and potassium contents of sugarcane in Louisiana. L.S.U.
Agr. Expt. Stn. Bull. No. 574.
Salter, P. J. and Williams, i. B. 1963. The effect of farm-yard
manure on the moisture characteristics of a sandy loam soil.
J. Soil Sci. 14, 73-81.
Samuels, G., Landrau, P. Jr., and Lugo-Lopez, M. A. 1952.
Handling of sugarcane trash: its effects on yield and soil.
Sugar 47, 47-49.
Thompson, G. D. 1965. The effects of trash conservation on
soil moisture and the sugar cane crop in Natal. Proc. 39th
Cong. S.A. Sugar Tech. Ass., pp. 143-156.
Mr. du Toit (in the chair): With reference to Table
IV I think that nitrification may be responsible for
the drop in calcium and magnesium.
Mr. Hill: I would like to question Dr. Thompson
on two points. Firstly, he has demonstrated very
clearly that trash conservation increases the organic
matter content of the soil and thereby the waterstability
of the topsoil crumbs. Both of these characteristics
should result in a decrease in soil bulk density,
yet this is not shown to be the case. Did Dr. Thompson
take the precaution of sampling for soil density
at similar moisture contents? Soils containing appreciable
quantities of 2:1 lattice clays change bulk densities
with moisture content. Secondly, trashing is
seen to increase the exchange capacity but decrease
the percentage base saturation of a soil. Did Dr.
Thompson notice any interaction with fertiliser in
this respect?
Dr. Thompson: The base saturation data are available
and 1 shall pass them on to you. The bulk
density samples were probably equally wet at the
time of sampling as it is our standard procedure to
water a site prior to taking an undisturbed core.

Proceedings of The South African Sugar Technologists' Association—March 1966 343
STUDIES OF THE EFFECT ON SUGARCANE
OF DAMAGE CAUSED BY FROST AND ASSOCIATED
MICRO-ORGANISMS
Introduction
In the winter of 1964, sugarcane suffered severely
from frost damage in the higher altitude areas of
Natal. Damage also occurred in scattered areas at
lower elevations, where frost had seldom been known
previously. In these instances, damage was relatively
serious, particularly where frost pockets formed.
The injury caused by frost is characterised by browning
of the canopy and the leaf sheaths, but the
growing point may also be severely damaged by
destruction of some or all of the meristematic tissue.
Experience gained in previous years has served to
indicate that frosted cane does not in all cases deteriorate
with the same rapidity. Thus, when rain falls on
cane damaged by frost, or conditions become humid,
then deterioration is rapid. However, deterioration of
the cane is also influenced by the degree of frost injury
suffered, and this must be taken into account when
judging whether a frost-damaged crop should be cut
immediately, or left.
To clarify the nature and scope of the damage
caused by frost, the injuries it causes have been studied
in some detail. The investigation embraces:
(a) A study of the role of micro-organisms involved
in the deterioration of cane, when this has been
damaged by frost.
{b) Cataloguing of symptoms associated with frost
damage at different levels of severity, so that
the grower can determine the significance of the
damage caused, and decide whether and when
to harvest.
(c) Determining the effect of frost on the quality
of cane juice.
Data used were obtained from cane damaged by
frost both in the field and in cooling chambers, where
frost conditions were simulated. The nature of this
damage, the regeneration of the cane, the relative
susceptibility of different varieties, the effect of frost
and the microflora associated with the damaged tissue
on purity, brix and sucrose percentage, were all
examined in the course of this study.
Field Observations
During the wiater of 1964, cane in the field, which
was exposed to frost, displayed varying degrees of
injury. These ranged from slight symptoms of damage
on the canopy, to total destruction of both the terminal
growing point and all the visible lateral buds.
Symptoms in the field compared closely with the
standards established in 1960 by J. Wilson, Director
of the Experiment Station of the South African Sugar
By G. ROTH
Association. Four categories of frost damage were
listed at that time and details are quoted in the South
African Sugar Journal, 44, pp. 837-839. These categories
are:
"(a) Where the fully developed or exposed parts of
leaves are killed, but the innermost and covered
parts of the spindle leaves are uninjured and
remain green.
(b) Where the innermost leaves of the spindle are
killed, but the apical growing point and the
basal parts of other leaves in the spindle are
unaffected.
(c) Where the apical growing point is killed in
addition to (a) and (b).
(d) Where the lateral buds on the main stem are
also killed, in addition to (a), (b) and (c)."
In 1964 it was found that slight damage caused by
frost may result in wilting of the leaf tips or parts of
the youngest leaves soon after exposure to freezing
temperatures, and that these symptoms may in turn
extend slowly over the whole canopy. These symptoms
are usually found following exposure to temperatures
down to — 2° C. for a short period, but in no case
for longer than about one hour. In all save a very few
cases, the susceptible growing point is not affected
(Fig. 1), and no damage was caused to the protected
rolled leaves of the spindle.
Yet another phenomenon which appears in cane
exposed to mild frosts for a short period, is the
evidence of damage found on leaves within the rolled
spindle. These symptoms are illustrated in Fig. 2 and
in this case they occurred at temperatures of not lower
than - 1° C. They are invariably associated with the
presence of small quantities of free water within the
rolled spindle, and the damage is presumably caused
by the formation of ice crystals. Extension of the
damaged areas depends on the population of microbes
present and on environmental conditions following
frost. High humidity favours the development of
micro-organisms within the damaged plant tissue, and
a typical spindle rot may develop.
Damage to the spindle is influenced by groups
of micro-organisms. In cases where saprophytic microbes
alone are present, damage may be localised,
but in other cases spindle rot can develop rapidly,
reaching parts of the growing point. When this happens,
the leaves of the spindle rot and the youngest
leaves wilt. At the time, however, the canopy as a
whole does not show any obvious symptoms of frost
injury and, provided the growing point has not been
completely killed by micro-organisms, the plants may
still recover without serious ill effects.

344 Proceedings of The South African Sugar Technologists' Association—March 1966

Proceedings of The South African Sugar Technologists' Association—March 1966
345

346 Proceedings of The South African Sugar Technologists' Association—March 1966
The damage caused by frost can usually be seen on
cane for several months after it has been affected.
Tissues in the injured lesions decompose and dry up,
and then healthy tissue regenerates around these
lesions (Fig. 3). Where the terminal growing point has
been so badly damaged that it dies, then side shoots
will develop (Fig. 4). However, where living cells
remain and the growing point recovers, typical black
rings or bands will be seen around the upper nodes
for many weeks after the frost.
Cane does not normally deteriorate when the
weather remains dry, even when the terminal point
has been killed. However, when high humidity or
rain, accompanied by warm weather follow frosting,
the damaged cane will deteriorate rapidly. It has been
found that where the damage extends downwards from
the growing point for more than 4 in., most of the
lateral buds will also have been injured and cane will
have reached the category (d) denned by Wilson. At
this stage, serious deterioration will occur when conditions
favour growth of micro-organisms. Various
categories of frost damage are illustrated in Fig. 5,
6 and 8.
Studies of Frost Damage
Field Cane
Frosted cane in fields in the Midlands and on the
North Coast of Natal was examined during the winter
of 1964. Samples of this cane were taken at intervals
of four weeks and analysed to determine their microbe
content. The samples used consisted of whole sticks
taken from randomly selected sites. These sticks were
examined in detail from the lowest bud to the terminal
growing point. Well known micro-biological techniques
were used to isolate and cultivate microorganisms,
the media utilised being:
(1) for bacteria: I/3-D agar, pH 7.0,
(2) for bacteria and fungi: potato dextrose agar,
pH 5.8 and 7.5,
(3) for yeasts: Sucrose-malt extract agar, pH 6.0;
and yeast morphology agar B 393 (see DIFCO
Manual, 9th Edition) pH 4.5.
A large number of micro-organisms have been
isolated, a few of which were found in most of the
samples examined, while the others were found only
occasionally and have therefore been classified as
casual infections. Samples were taken from the
growing points of 50 sticks of cane collected in
different fields. These revealed that more than 45
genera of bacteria were present, including at least 8
genera of yeasts and 11 genera of other fungi (Fig.
10-13). Longitudinal sections of the cane tops and
growing points revealed a pinkish coloured, cottonwool-like
mycelium and the presence of a characteristic
odour, symptoms which are typical of infection
by Fusarium poae. It was this fungus, together with
Fusarium moniliforme which proved to be the only
organisms closely associated with spindle rot in the
field (Fig. 10) and it is they which harm the frost
damaged growing point. All other fungi found on
frosted cane behave as saprophytes, and have been
ignored in this study.
Artificial Frosting
Experiments were carried out to study the effect of
frost on eleven varieties of cane, namely Co.331;
N:Co.293; N:Co.310; N:Co.376; N.50/211; N.50/805;
N.5I/168; N.51/539; N.53/216; C.B.28/22; and C.B.
36/14. The cane was subjected to low temperatures in
a freezing chamber, which measured 6 ft. by 11 ft. by
15 ft. 3 in. Temperature treatments were: -2° C.
Individual cane samples exposed to this temperature
for periods varying from 1 to 10 hours, in increments
of 1 hour; to -4° C. for periods varying from half an
hour to 6 hours, in increments of half an hour; and to
-6° C. for periods varying from half an hour to four
hours, in intervals of half an hour.
The cane used in these experiments was cut not
more than 4 hours before it was treated. The sticks
were severed at the surface of the soil, using shears
made for this purpose, and after cutting they were
packed in bundles for treatment.
Assessing Frost Damage
To assess the damage caused by frost, the five uppermost
nodes, including the terminal growing point,
were sectioned longitudinally. Four categories of
damage were defined and this classification was used
to record frost injury. The germination capacity of
the frosted cane was also compared with unfrosted
samples. Two methods were used to establish germination
capacity. In the first, sticks were cut to provide
single bud setts, which were put into plastic tubes of
2.5 in. diameter and then kept in the dark for four
weeks. The second method employed, involved placing
whole sticks in water, which was kept fresh and shaded
for a period of four weeks.
Studies of Microbes
Analysis of the microbe populations in cane were
carried out on material treated, or to be treated, in
the freezing chamber. These analyses were restricted
to the variety N:Co.376 and they were carried out
both immediately before subjecting cane to low temperatures
— but about four hours after cutting — and
one week and four weeks after treatment. The treatments
involved in this case were:
Cane kept for 2 hr., 5 hr. and 10 hr. at a temperature
of-2° C.
Cane kept for 4 hr. at a temperature of - 4° C.
Cane kept for 4 hr. at - 6° C.
Samples were taken from the terminal growing
point and the penultimate node at the base of each
harvested stick, and these were analysed. A sterile cork
borer was used for sampling and this yielded cores
which were identical in size. In each case the core was
0.5 by 1.0 cm. in size, and before it was used it was
first homogenized in 10 cc. of 0.7 sterile salt solution
and then gradually diluted using a standard dilution
technique. The number of micro-organisms present
was, in each case, determined by averaging the counts
from 4 samples, using the agar plate technique. The
data obtained is shown in Table II and they indicate
that the greater the damage caused by frost, the more
quickly the various micro-organisms responsible for
cane deterioration develop.

Proceedings of The South African Sugar Technologists' Association—March 1966 347

348
Experiments
Two detailed experiments were put down to determine
the effect on individual tissues, of exposure to
frosts of different severity for varying lengths of time.
In the first of these, eleven varieties of sugarcane were
exposed to artificially created temperatures of- 2° C,
- 4° C. and - 6° C. Injury to the uppermost 5 nodes
including the growing point, and to buds, starting
with the 6th and extending to the bottom of the stick,
was recorded by scoring from. 0 to 5, 0 being interpreted
as representing undamaged cane or cane with
only the faintest trace of damage, while 5 represents
a complete kill. The results of treatments at - 2" C.
and -4° C. are summarised in Fig. 14A.
It can be seen from Fig. 14A that both nodes and
buds are damaged within half an hour, when the
temperature is kept down to -4" C, and that this
damage increases with the length of time of exposure.
After 3 hours at this temperature, damage to the nodes
and death of the buds exceeds an estimated 80 per
cent. When placed in plastic tubes, single bud setts of
cane which had been subjected to this low temperature
were, after 4 weeks, so inundated with micro-organisms
that the frost-damaged buds were completely watersoaked.
Cane in one replicate in this experiment was kept
for 4 weeks with its lower nodes immersed in fresh
water to a depth of 9 in. — the water being changed
every second day. Four sticks of each of the eleven
varieties were used for juice quality tests and the data
obtained are summarised in Table I.
No significant differences in the frost resistance of
different varieties was found in the first experiment.
Indeed, differences between individual sticks of the
same variety were greater than those found between
varieties. A second trial was therefore put down, involving
only seven varieties of sugarcane and exposure
to only - 2° C. for periods of 1 to 8 hours, in increments
of 1 hour. Once again, variations in frost
hardiness between sticks of the same variety was so
great that no statistically significant varietal differences
were found. The results from this trial are shown
graphically in Fig. 14B.
Proceedings of The South African Sugar Technologists' 1 Association—March 1966
Results and Conclusions
(1) Damage caused by frost increases as the temperature
drops, and as the period of exposure lengthens.
Thus, a light frost in which the temperature
does not fall below - 2° C. and does not continue
longer than 1 hour, causes only superficial damage
to the canopy, or slight injury to the leaves of the
spindle. However, where frost of a similar intensity
lasts for more than 1 hour, but less than 2
hours, it may cause minor damage to the canopy
and the spindle, and eventually to the growing
point. The extent of this injury will vary according
to the conditions under which cane has been
grown and its age. Young cane is more susceptible
to frost than older cane.
(2) At temperatures below - 2 o C. cane suffers more
noticeable damage. Severe damage is caused by
exposure to temperatures of -4° C. and - 6° C.
for half an hour.
(3) The terminal growing point is more easily damaged
by frost than any other part of the plant.
Greatest frost resistance is found at the internodes,
and at the nodes and the buds towards the
bottom of the stick.
(4) Frost injury starts where cells have been separated
by rifts along the middle lamellae, the cells
becoming isolated from the neighbouring tissues
and then dying. Complete recovery may occur
when only a relatively small peninsula of uninjured
cells remains in the growing point or in the
axillary buds, provided that these are not isolated
from uninjured sections of the vascular bundles.
(5) Deterioration of tissue, both at the growing point
and at the nodes following damage by frost, is
associated with a tremendous increase in the
population of microbes in these tissues.
(6) The most harmful fungi attacking, the frost damaged
growing point are Fusariwn poae and F.
moniliforme. These are responsible for "heart
rot" of cane.
(7) Juice quality of harvested cane, which has been
stored standing in fresh water in the shade, is
not affected by frost when the temperatures at
harvest drop to a mere - 2° C, even when this
low temperature is maintained for up to 8 hours.
However, when cane has been seriously damaged
by frost, as is the case when it is exposed to
temperatures of - 4° C. and - 6° C, then there
is a very marked drop in sucrose percentage and
purity. This drop is associated with an increase
in the population of micro-organisms during
storage.
(8) Eleven varieties were examined for their resistance
to frost, but no sound indications of resistance
were found.
(9) The data obtained in these studies should help
the cane grower to determine the extent of frost
damage suffered by his crop. It should also enable
him to decide whether to cut his cane to avoid

350
loss or if he should allow it to regenerate. Alternatively,
he will be able to judge how quickly
frost-damaged cane will deteriorate.
Summary
In the winter of 1964, sugarcane in several parts of
South Africa suffered serious damage as a result of
unusually harsh frosts. Studies of the effects of frost
damage were carried out and, in this paper, the
nature of the damage is described. Investigations
include the effect of frost on the quality of cane juice,
the relative susceptibility of different varieties, the
identification of micro-organisms associated with
damaged tissue and their influence on purity, brix and
sucrose percentage.
To enable this investigation to be carried out, it was
necessary to supplement studies of frost damage in
the field, by inducing frost damage in cooling chambers.
Details are given of these studies. Damage to cane was
negligible at temperatures down to - 2° C, provided
the exposure period did not exceed 1 hour. Longer
Proceedings of The South African Sugar Technologists' Association—March 1966
periods of exposure, or a further lowering of temperature,
have a steadily greater impact. The terminal
growing point has proved to be the tissue which is
most susceptible to frost damage. The uppermost
nodes are also liable to suffer damage quite easily,
but the lower down the stem they are located, the
more frost hardy they are.
The most important micro-organisms affecting cane
are two semi-parasitic fungi Fusarium poae and F.
moniliforme. No differences could be found in the
susceptibility of varieties to frost damage. The data
obtained from this investigation should, however,
help the grower to decide on the action he should take
following a frost.
Acknowledgment
The author wishes to thank Mr. C. Whitehead for
his assistance in editing this paper and preparing it
for publication.
References
Wilson, J. "Some observations on the effect of frost on sugarcane
in the cane belt during 1960." S.A. Sugar J. 44, 837-839.

Proceedings of The South African Sugar Technologists' Association—March 1965
SNSTRUCTIONS TO AUTHORS
1. All papers for the Congress must be in the hands of the Technical Secretary
twenty-eight days before the meeting but authors are requested to
submit papers earlier to assist the printers.
2. Manuscripts are to be typewritten if possible.
3. Titles should be short, e.g. "A report on preliminary investigations into
possible new methods of cane sampling" could be reduced to "New methods
of cane sampling — preliminary investigations".
4. Illustrations and diagrams accompanying the papers must be carefully drawn
on smooth white drawing paper in Indian ink to ensure good repro­
duction. Lettering on margins should be in pencil so as to allow of easy
replacement by printer's type. The size of the largest diagram, curve etc.,
should not be greater than 10 by 15 inches. Curves should be drawn on
black graph paper, which can be obtained from the Technical Secretary.
A careful correction should be made by authors of typographical and other
errors.
5. If a statement is copied word for word from any publication whatsoever, it
must be placed in inverted commas and due acknowledgment be made of
its source. Sketches, diagrams or illustrations so copied must also be ack­
nowledged.
6. A summary of the paper should be given at the end before the
REFERENCES. It should consist of a brief recapitulation of the whole work
with its divisions and sub-divisions in their original order from introduction
to conclusion.
7. References at the end of the paper should be arranged in alphabetical order
and should be as follows:
Name and initials of author, title of article, title of periodical, volume
number, date, first and last pages,
Example:
Brown, S. J. Cane sampling methods. Sugar Milling Research Institute Quarterly
Bulletin, No. IS, January 1963, f>p. 12-15.
8. Photographs, diagrams and graphs must be given consecutive 'Figure Nos.'
in Arabic Numerals. Tables must be numbered successively in Roman
Numerals.
9. Authors may have 25 copies of their papers on application to the Technical
Secretary. A charge will be made for any additional copies.
10. Papers read at the Congress will not necessarily be published in the
Proceedings.
351