kasetsart journal natural science

January - March 2006 

Volume 40 Number 1 

http://www.rdi.ku.ac.th

Advisor : Samakkee Boonyawat 

Rangsit Suwanketnikom 

Editor-in-Chief : Ed Sarobol 

The Kasetsart Journal 

Associate Editors : Wanchai Chanprasert, Natural Science 

Suparp Chatraphorn, Social Science 

Editorial Board : Natural Sciences Social Sciences 

Amara Thongpan Suwanna Thuvachote 

Pornsri Chairatanayuth Pongpan Trimongkholkul 

Onanong Naivikul Matrini Ruktanonchai 

Praparat Hormchan Nongnuch Sriussadaporn 

Korchoke Chantawarangul 

Aree Thunyakijjanukij 

Patana Sukprasert 

Manager : Orawan Wongwanich 

Assistant Managers : Surai Suwannarat 

Overseas Members 

G. Baker (Mississippi State University, USA.) 

A. Bruce Bishop (Utah State University, USA.) 

John Hampton (Lincoln University, New Zealand) 

Helen H. Keenan (University of Stathclyde, Scotland) 

Chitochi Miki (Tokyo Institute of Technology, Japan) 

Eiji Nawata (Kyoto University, Japan) 

T. Miyata (Nagoya University, Japan) 

Business Office : Kasetsart University Research and Development Institute (KURDI) 

Kasetsart University, Chatuchak, Bangkok 10900. 

The Kasetsart Journal is a publication of Kasetsart University intended to make available the results 

of technical work in the natural and the social sciences. Articles are contributed by Kasetsart University faculty 

members as well as by those from other institutions. The Kasetsart Journal : Natural Sciences edition is issued 

four times per year in March, June, September and December while The Kasetsart Journal : Social Sciences 

edition is issued twice a year in June and December. 

Exchange publications should be addressed to 

The Librarian, 

Main Library, 

Kasetsart University, 

Bangkok 10900, Thailand.

KASETSART JOURNAL 

NATURAL SCIENCE 

The publication of Kasetsart University 

VOLUME 40 January - March 2006 NUMBER 1 

Tillage System and Fertilizer Rate Effects on Sorghum Productivity in the Central Rift Valley of 

Oromiya, Ethiopia 

........................... Worku Burayu, Sombat Chinawong, Rungsit Suwanketnikom, 

..............................................................Thongchai Mala1 and Sunanta Juntakool 1 

Repeatability, Optimal Sample Size of Measurement and Phenotypic Correlations of 

Quantitative Traits in Guava 

...................................................... Kriengsak Thaipong and Unaroj Boonprakob 11 

Heritability, Heterosis and Correlations of Fruit Characters and Yield in Thai Slicing Melon 

(Cucumis melo L. var. conomon Makino) 

........................................................................Chamnan Iathet and Kasem Piluek 20 

Seed Development and Maturation of Eryngo (Eryngium foetidum L.) 

..........................................................Boonsong Ekpong and Sutevee Sukprakarn 26 

Evaluation for Antibiosis Resistance in Cotton to Helicoverpa armigera Larvae 

.................................................... Praparat Hormchan and Arunee Wongpiyasatid 33 

Influence of Biotic and Chemical Plant Inducers on Resistance of Chilli to Anthracnose 

............................................................. Le Thi Kieu Oanh, Vichai Korpraditskul, 

.................................................. Chainarong Rattanakreetakul and Sirikul Wasee 39 

Incidence of Cymbidium Mosaic Virus and Odontoglossum Ringspot Virus on 

In Vitro Thai Native Orchid Seedlings and Cultivated Orchid Mericlones 

.......................................................... Yuphin Khentry, Ampaiwan Paradornuwat, 

........................... Sureeya Tantiwiwat, Salak Phansiri and Niphone Thaveechai1 49 

Comparative Performances of Holstein-Friesian Cows under Smallholder and Large Scale 

Farmers’ Management in Central Rift Valley, Ethiopia 

........................................ Nega Tolla, Pravee Vijchulata, Pornsri Chairatanayuth 

................................................................................ and Suwapong Swsdiphanich 58 

Biochemical Properties of Nile Tilapia (Oreochromis niloticus) Hemoglobin 

...................................... Kriangkrai Thongsarn, Wanchai Worawattanamateekul, 

........................ Suriyan Tunkijjanukij, Choosri Sribhen and Apassara Choothesa 69 

Screening of Ethiopian Traditional Medicinal Herbs for their Nitrification Inhibiting Ability 

............................................................................. Wassie Haile, Thongchai Mala, 

............................................................... Yongyuth Osotsapar and Visoot Verasan 74 

Seasonal Characteristics of Wood Formation in the Elite Genetic – Based 

Eucalyptus camaldulensis Dehnh. 

..................................................... Teera Veenin, Tadashi Nobuchi, Minoru Fujita 

.................................................................................. and Somkid Siripatanadilok 83

Development of Catalase Gene Nuclear DNA-Based Marker for Population Genetic Analysis 

in Thai Teak (Tectona grandis L.f.) 

............................................................. Jongkon Cheua-ngam and Hugo Volkaert 91 

Effects of Na+, K+ and Ca2+ Accumulation on the Expression of Ca2+-ATPase Gene 

in Rice KDML 105 

.............................................................Wunrada Surach, Mingkwan Mingmuang 

............................................................................................ and Amara Thongpan 99 

Molecular Identification of Cycas by Restriction Fragment Length Polymorphism (RFLP) 

and Random Amplified Polymorphic DNA (RAPD) 

..................................................................... Pattamon Sangin, Amara Thongpan, 

................................................. Anders J. Lindstrom and Mingkwan Mingmuang 107 

Physiological Study and Alcohol Oxidase Gene(s) of Thermotolerant Methylotrophic 

Yeasts Isolated in Thailand 

........................................................................ Nantana Srisuk, Savitree Limtong, 

............................................ Hiroya Yurimoto, Yasuyoshi Sakai and Nobuo Kato 121 

Characterization of Grass Degrading Bacteria Active on β-1,3-1,4-D-glucans from 

Bacillus subtilis GN156 Potential Use for Grass Silage-Making 

........................................................... Jirawan Apiraksakorn, Tonglian Buwjoom 

........................................................................................ and Sunee Nitisinprasert 136 

Thermal Ageing of Thermoplastic Elastomeric Natural Rubber-Low Density 

Polyethylene Blends 

.........................................................Wiwat Suaysom and Wirunya Keawwattana 148 

Morphology and Haemolymph Composition Changes in Red Sternum Mud Crab (Scylla serrata) 

.............................................................Jintana Salaenoi, Anchanee Sangcharoen, 

..................................................... Amara Thongpan and Mingkwan Mingmuang 158 

A Comparison of Rearing Pacific White Shrimp (Liptopenaeus vannamei Boone, 1931) 

in Earthen Ponds and in Ponds Lined with Polyethylene 

........................................................ Onanong Prawitwilaikul, Chalor Limsuwan, 

..................................................................... Wara Taparhudee and Niti Chuchird 167 

Application of Near Infrared Spectroscopy to Predict Crude Protein in Shrimp Feed 

.................................................... Jirawan Maneerot, Anupun Terdwongworakul, 

........................................................ Warunee Tanaphase and Nunthiya Unprasert 172 

The Effect of Peptidoglycan on Immune Response in Black Tiger Shrimp 

(Penaeus monodon Fabricius) 

..................................................... Watchariya Purivirojkul, Nontawith Areechon 

................................................................................... and Prapansak Srisapoome 181 

Distribution and Early-life Development of Thai River Sprat Clupeichthys aesarnensis 

Wongratana, Larvae, in Pasak Jolasid Reservoir, Lop Buri Province, Thailand 

............................................................................................... Santi Poungcharean 188 

Gonadal Development and Sex Inversion in Saddleback Anemonefish Amphiprion polymnus 

Linnaeus (1758) 

.....................................................Sukjai Rattanayuvakorn, Pisut Mungkornkarn, 

................................................ Amara Thongpan and Kannika Chatchavalvanich 196

In sacco Degradation Characteristics of Crop Residues and Selected Roughages in 

Brahman-Thai Native Crossbred Steers 

........................................................ Songsak Chumpawadee, Kritapon Sommart, 

.......................................................... Thevin Vongpralub and Virote Pattarajinda 204 

Comparative Efficiency of KU and ISO Plungers in Mixing Composite Bulk Raw Milk 

......................................................................... Jigme Wangdi, Pravee Vijchulata, 

........................................ Pornsri Chairatanayuth and Suwapong Swasdiphanich 215 

Influences of Physicochemical Characteristics of Rice Flour and Cassava Starch on the 

Gelation of Calcium-Induced Egg Albumen-Flour Composite 

................................................... Parichat Hongsprabhas and Kamolwan Israkarn 223 

The Product Design of Puffed Snacks by Using Quality Function Deployment (QFD) 

and Reverse Engineering (RE) Techniques 

................... Wiwat Wangcharoen, Tipvanna Ngarmsak and Brian H. Wilkinson 232 

Effect of Coating on Doughnut Cake Preference using R-index 

............................. Tunyaporn Sirilert, Anuvat Jangchud, Phaisan Wuttijumnong 

....................................................................................... and Kamolwan Jangchud 240 

Application of Artificial Neural Networks for Reservoir Inflow Forecasting 

....................................................... Varawoot Vudhivanich, Santi Thongpumnak, 

.......................Nimit Cherdchanpipat, Areeya Rittima and Nattaphun Kasempun 247 

Jordan Derivations on Rings 

..................................................... Orapin Wootijiruttikal and Utsanee Leerawat 260 

Geoinformatic Public Domain System Model “ SWAT “ in Thailand 

............................................................................................... Hansa Vathananukij 264

Kasetsart J. (Nat. Sci.) 40 : 1 - 10 (2006) 

Tillage System and Fertilizer Rate Effects on Sorghum Productivity 

in the Central Rift Valley of Oromiya, Ethiopia 

Worku Burayu 1, *, Sombat Chinawong 1 , Rungsit Suwanketnikom 2 , 

Thongchai Mala 1 and Sunanta Juntakool 2 

ABSTRACT 

Soil moisture and soil nutrient are the most sorghum yield limiting factors in semi-arid Oromiya. 

Hence, to identify appropriate crop management system for sorghum productivity, the field experiment 

was conducted in 2004 cropping season using factorial combination of four levels of tillage systems and 

four rates of fertilizer in spilt plot design at two locations. It was found that the stand count and height 

of sorghum varied significantly between locations, and lower stand count recorded at Wolenchity (55738 

plant ha -1 ) than at Malkassa (61548 plant ha -1 ) while greater plant height was obtained at Wolenchity. 

Grain yield was significantly affected by location and fertilizer rate. Significantly (P≤0.05) higher grain 

yield was obtained at Wolenchity (2381 kg ha -1 ) than that at Malkassa (1747 kg ha -1 ). Grain yield 

achieved at the highest fertilizer rate of 57.4-64.4 kg N-P 2O 5 ha -1 was significantly (P≤0.01) higher than 

that at the current rate of 41-46 kg N-P 2O 5 ha -1 . The highest grain yield was recorded from tie-ridge plot 

but varied with fertilizer rate for each location. Harvest Index (HI) followed the same pattern as the 

grain yield. However, significantly (P ≤ 0.05) higher stover and biomass yield of sorghum were obtained 

at Malkassa. These findings indicated that applications of fertilizer beyond 49.2-55.2 kg N-P 2O 5 ha -1 

could give no significant yield advantage and thus, would not be economically feasible. The tie-ridge 

and reduced tillage tied furrow were encouraging but need further investigation to incorporate in sorghum 

cropping system. 

Key words: harvest index, no-tillage, soil nutrient, soil moisture, tied-ridge 

INTRODUCTION 

Cereal crops account for over 86% of the 

area planted with food crops each year in Ethiopia 

(CSA, 2001). Sorghum (Sorghum bicolor) is one 

of the food crops that occupy 20% of the cultivated 

land under cereals (CSA, 2001). It is a staple food 

for a significant proportion of the lowland rural 

population. Known as the most drought tolerant 

crop, sorghum is grown as one of the major 

1 Faculty of Agriculture, Kamphaeng Saen, Kasetsart University, Kamphaeng Saen Campus, Nakhonpathom 73140, Thailand. 

2 Kasasart University, Bangkean Campus, Bangkok 10900, Thailand; 

* Corresponding author, e-mail: workuburayu@yahoo.com 

multipurpose cereals in the semi-arid areas. Close 

to one million hectares is developed and about 1.2 

million tones are produced each year (Central 

Statistical Authority, CSA, 2001). 

Despite the importance of sorghum crop, 

the productivity of sorghum is far below the 

genetic potential of the crop; the national mean 

yield has been estimated at about 1.3 t ha -1 on 

peasant farms (CSA, 2003). However, research 

results have shown that using improved varieties 

Received date : 25/07/05 Accepted date : 13/12/05

2 

and management practices a sorghum grain yield 

of 4-5 t ha -1 can be achieved (Kidane et al., 2001). 

Such low yields and the production shortfall of 

sorghum cropping system in semi-arid areas are 

attributable to several factors. Among these, soil 

moisture stress, poor soil fertility, and pests are 

the most limiting factors (Kidane and Abuhay, 

1997). Water stress is one of the major causes for 

low yields and total crop failure of cereals in the 

semi-arid areas of the country and soil fertility is 

the next constraint generally faced in such regions 

(Kidane et al., 2001). Hence, moisture stress and 

nutrient deficiency is critical in such soils and 

regions. Of all nutrients, nitrogen and phosphorus 

are the most crop growths and yield limiting 

factors in the country (Kidane et al., 2001). The 

study conducted in the central rift valley of 

Oromiya also emphasized that the principal 

constraints to increase crop production in semiarid 

regions were the minimal combination of 

technologies for water availability, soil fertility, 

and new genetic material (Kidane et al., 2001). 

These insist the necessity of further studies in 

combined technologies for soil moisture 

conservation and fertilizer requirements for crop 

production of this region. Furthermore, the studies 

on combined influences of conservation tillage 

practices and soil fertility management have been 

minimal particularly in the dryland central rift 

valley of the country. Hence, the need for in-depth 

research on the combination of moisture 

conservation techniques and fertility management 

is unprecedented. This experiment was, therefore, 

initiated to undertake comprehensive studies with 

the objectives to identify the appropriate tillage 

systems and fertilizer rates for productivity of 

sorghum in the semi-arid, central rift valley of 

Ethiopia. 

MATERIAL AND METHODS 

The experiment was initiated in 2004 at 

Melkassa Agricultural Research Center (MARC) 

Kasetsart J. (Nat. Sci.) 40(1) 

and Wolenchity research sub-station with the 

hypothesis that implementation of various 

conservation-tillage systems with different rates 

of fertilizer would result in better sorghum crop 

yield as compared to the conventional tillage. 

MARC and Wolenchity are located in the semiarid 

central rift valley of Oromiya. The MARC is 

located at 8° 24’N latitude and 39° 12’E longitude 

of 1550 m above mean sea level and Wolenchity 

at 8° 40’N latitude and 39° 26’E longitude of 1450 

m above mean sea level. The soil of experimental 

sites are loam soil with 41% sand, 37% silt, 22% 

clay content and a pH of 6.41 for MARC, and 

sandy loam with 46% sand, 34% silt and 20% clay 

content with a pH of 6.64 for Wolenchity. 

The combined effects of tillage systems 

and fertilizer rates on sorghum productivity trial 

consisted of 16 treatments comprising the factorial 

combinations of four levels of tillage management, 

i.e., Conventional Tillage (CT), Reduced Tillage 

(RT), No-Tillage (NT) and Tie Ridge (TR), and 

four levels of fertilizer, i.e., 0-0 (F 0), 41-46 (F 1), 

49.2-55.2 (F 2), 57.4-64.4 (F 3) kg N-P 2O 5 ha -1 . The 

experiment was laid out in a 4 × 4 spilt plot design 

with three replications at two locations. The fourtillage 

systems were main plots of 14m × 16.5m 

(231m 2 ), and fertilizer rates placed in sub plots of 

14m × 3.75m (52.5m 2 ). Pathways of 0.5m, 0.75m 

and 1m were placed between subplots, main plots 

and replications, respectively. A row spacing of 

0.75m was used. The conventional tillage system 

consisted of four plowings with traditional oxen 

plow ‘Maresha’ (farmer’s practice) to a depth of 

first pass approximately 8 cm and the other two 

passes perpendicular to the previous path with a 

final one at 20-cm depth prior to planting. In 

conventional tillage the first hand weeding was 

done 20-29 days after crop emergence (DAE) and 

the second hand weeding was done 40-50 DAE. 

In the no-tillage treatment, no soil disturbance was 

made except for seeding and fertilizer application. 

In the reduced tillage tied furrow, it was designed 

to use the ridger only after one pass with the ox-

plow, then furrow ties were made during planting 

at 5 m interval. Both no-tillage and reduced tillage 

tied furrow plots were sprayed with glyphosate at 

a rate of 3 l ha -1 as preplanting herbicides. In tie 

ridge treatments, after three plowings with 

traditional ox plow, 35 cm high ridges were 

constructed 75 cm apart and cross-tied with soil 

bunds across the ridges with tie ridger at about 

every 5 m ridge length. 

The fertilizer sources were urea (46% N) 

and diammonium phosphate (18% N and 46% 

P 2O 5). All of the P fertilizer and half of the N 

fertilizer were banded 5 cm below and 5 cm away 

from the rows as a basal application during 

planting. The rest half of the N fertilizer was 

applied when crop reached a knee height. 

The improved sorghum variety, 

Meko-1, an early maturity type (60-70 days to 

anthesis) was used and the seeds were placed in 

rows and sorghum seedlings were thinned to one 

plant per hill 15 days after emergence to ensure 

the targeted population. 

Data on various crop parameters were 

collected throughout the cropping season. Stand 

count at harvesting was recorded by counting the 

actual numbers of plant in the subplot area and 

expressed on hectare basis. Plant height for a 

randomly selected six plants (two plants within a 

36 m segment of three rows) per sub-plot was 

determined. Sorghum heads and stover were 

harvested at the base of the lowest grain branch 

and at the ground surface level, respectively from 

areas of 13.5m 2 (6m × 2.25m) 105 DAE. Then 

sorghum head height was determined, sun-dried 

and weighed. Counting 250 grains in duplicates 

and weighing them on two decimal electronic 

balances, thousand seed weight was determined. 

The weights thus obtained were added and 

multiplied by two to reach 1000-seed fresh weight. 

They were then oven-dried at 55-60 °C and 

weighed again to determine moisture contents and 

to obtain 1000-seed dry weight. Grain yield and 

above ground biomass were determined from areas 

Kasetsart J. (Nat. Sci.) 40(1) 3 

of 13.5 m 2 . Grain yield was adjusted to 12.5% 

moisture content. A total above ground biomass, 

which included stover and whole panicles, was 

used to obtain biomass yield. Harvest index (HI) 

values were computed as the ratio of the mass of 

grain yield to total biomass (grain + stover). 

Soil moisture at 0-15 depth was 

determined gravimetrically for each plot in the 

central row in two replications using a core 

sampler. Soil water data were recorded at various 

growth stages from planting until the physiological 

maturity of sorghum crop. Gravimetric water 

content was converted to a volumetric basis using 

bulk densities of soil core taken from each depth 

(Lopez et al., 1996). 

Data were subject to General Linear 

Models (GLM) Procedure using SAS Statistical 

Software (SAS, 1989). Duncan’s Multiple Range 

Test (DMRT) and Least Significant Differences 

(LSD) were used for mean separation at the 0.05 

or 0.01 probability levels. 

RESULTS AND DISCUSSION 

Stand count, plant height, head height and 

1000-grain weight 

Stand count of sorghum varied 

significantly between locations and different 

tillage systems but no significant differences 

among fertilizer rates could be detected. The data 

in Table 1 revealed that an estimated mean stand 

count of sorghum at Wolenchity was significantly 

lower than that observed at Malkassa. When the 

data for different fertilizer rates and both locations 

were combined the stand counts from conventional 

tillage and the reduced tillage tied furrow were 

significantly higher than those obtained from either 

no-tillage or tie-ridge tillage systems. 

Slight difference was observed in plant 

height between locations, among different rates 

of fertilizer application and the interaction between 

location and fertilizer, and between tillage and 

fertilizer rates. Unlike the stand count of sorghum,

4 

the greater plant height was obtained at Wolenchity 

as compared to that obtained at Malkassa. The 

unfertilized plot had significantly lower plant 

height those that obtained from fertilized plots 

(Table 2). 

Almost equal head height was observed 

at Wolenchity and Malkassa (Table 3). It was only 

reduced tillage that was varying across locations 

and had significantly higher head height at MARC 

than the corresponding tillage at Wolenchity. 


The 1000-grain weight of sorghum was 

significantly affected by location (P ≤ 0.01), and 

significantly higher 1000 seed weight was obtained 

at Wolenchity as compared to that obtained at 

Malkassa (Table 4). 

The highest seed weight was observed 

on the tied ridge treatment with the highest rate of 

fertilizer application at Wolenchity and no-tillage 

treatment of the same rate of fertilizer at Malkassa. 

Sorghum grain weight reflects the crop growing 

Table 1 Influences of tillage system on stand count (plant ha -1 ) with varied locations. 

Tillage system Location 

Wolenchity Malkassa Mean 1 

CT 64143 70952 67548 a 

RT 63214 70863 67039 a 

NT 52333 57292 54813 b 

TR 43262 47083 45173 c 

Mean 55738 B 61548 A 

1 Means followed by common lowercase letters within a column do not differ significantly at 5% probability level of significance, 

and means followed by different uppercase letters within row differ significantly at 5% probability level of significance. 

Table 2 Influences of fertilizer on plant height (cm) with varied locations. 

Fertilizer rate Location 

Wolenchity Malkassa Mean 

F 0 148.73 137.13 142.93b 

F 1 149.40 146.88 148.13a 

F 2 151.08 145.55 148.31a 

F 3 155.58 141.38 148.48a 

Mean 151.19 142.73 

1 Means followed by common letters within a column do not differ significantly at 5% probability level of significance. 

Table 3 Influences of tillage systems on head height (cm) with varied locations. 

Tillage system** Location 

Wolenchity MARC Mean 

CT 23.25AB* 23.25AB 23.25 

RT 20.13B 24.25A 22.19 

NT 23.38AB 22.88AB 23.13 

TR 23.25AB 22.25AB 22.75 

Mean 22.50 23.16 

* Means followed by the same common letters are not significantly different at 5% probability; ** CT = Conventional tillage, 

RT = Reduced tillage, NT = No-tillage, TR = Tie ridge

conditions during the grain filling period. The 

greater seed weight at Wolenchity than at Malkassa 

might be found due to mild water deficit during 

grain filling at the latter location. 

Grain yield and harvest index of sorghum 

Grain yield of sorghum were 

significantly affected by locations (Table 5). The 

sorghum grain yield obtained at Wolenchity was 


significantly higher (P

6 

at Wolenchity (Figure 3) that led to higher grain 

yield. The difference in mean grain yield among 

fertilizer rate was highly significant (P

significant yield advantage. 

In the study, the highest sorghum grain 

yield was recorded due to tie-ridge tillage but 

varied with fertilizer rate for each location (2762 

kg ha -1 grain at 49.2-55.2 kg N-P 2O 5 ha -1 for 

Wolenchity; and 2143 kg ha -1 at 57.4-64.4 kg N- 

P 2O 5 ha -1 for Malkassa). The yield obtained due 

to tie-ridge and reduced tillage tied furrow was 

equal at Malkassa. There were many other results 

which validated these findings, as it was evident 

from the extensive published data on tillage 

affecting crop yield that differed with soil 

conditions and environments (Lal, 1986; Triplett, 

1986; Arnon, 1992; Dao, 1993; Radford et al., 

1995). 

Harvest index of sorghum varied 

significantly with location (P

8 

Stover and aboveground biomass 

Contrary to the grain yield and HI, 

significantly higher (P

CONCLUSION 

The stand count, plant height, grain yield, 

HI, stover and biomass yield of sorghum differed 

significantly between locations and some among 

fertilizer rates but no significant differences among 

tillage systems could be detected. Significantly 

greater stand count, stover and biomass yield of 

sorghum were obtained at Malkassa. Plant height, 

grain yield and harvest index of sorghum were 

significantly higher at Wolenchity. Grain yield and 

HI achieved at the highest fertilizer rate was 

significantly (P

10 

Radford, B.J., A.J. Dry, L.N. Robertson and B.A. 

Thomas. 1995. Conservation tillage increases 

soil water storage, soil animal population. 

J. Soil Water Consv. 40: 466-470. 

SAS, 1989. SAS Institute Inc., Cary, NC, USA. 

SAS software release 6.12. Unpublished. 


Triplett, G.B. 1986. Crop management practices 

for surface-tillage systems, pp. 149-182. In 

M.A. Sprague and G.B. Triplett (eds.). The 

tillage revolution. Zero-tillage and Surfacetillage 

Agriculture. John Wiley, New York.


Repeatability, Optimal Sample Size of Measurement and 

Phenotypic Correlations of Quantitative Traits in Guava 

ABSTRACT 

Kriengsak Thaipong and Unaroj Boonprakob* 

Five fruits from each of 11 guava genotypes were evaluated in dry and early rainy seasons 

under Thailand conditions for fruit weight, flesh thickness, flesh weight, seed cavity (central pulp) 

weight, fruit firmness, total soluble solids, titratable acidity, juice acidity, and ascorbic acid to estimate 

repeatability (R), phenotypic correlations (r), and to predict the optimal sample size. The repeatability 

of the fruit weight, flesh thickness, flesh weight, seed cavity weight, titratable acidity, juice acidity, and 

ascorbic acid were relatively high (R ≥ 0.60). The flesh thickness, titratable acidity, juice acidity, and 

ascorbic acid were the traits with highest estimates, 0.85, 0.85, 0.87, 0.76 and 0.85, 0.83, 0.84, 0.80 in 

dry and early rainy seasons, respectively. Based on a threshold of 10% increase in relative efficiency, a 

sample of three fruits was sufficient for evaluating guava fruit traits. Most physical traits (fruit weight, 

flesh thickness, flesh weight, and seed cavity weight) had weak negative correlations (-0.25 ≤ r ≤ –0.38) 

with chemical traits (total soluble solids, titratable acidity, and ascorbic acid). Fruit firmness had no 

correlation with all other fruit traits. There were strong positive correlations between fruit weight and 

flesh thickness (r = 0.81), flesh weight (r = 0.99), and seed cavity weight (r = 0.88). Therefore, fruit 

weight could be used as an indirect selection for flesh thickness, flesh weight, and seed cavity weight. 

Key words: Psidium guajava L., breeding, quantitative trait analysis, fruit qualities 

INTRODUCTION 

Guava (Psidium guajava L.) is native to 

tropical America and presently found distributing 

in several tropical and subtropical regions (Cobley, 

1976) such as India, South Africa, Brazil, Cuba, 

Venezuela, New Zealand, the Philippines, Hawaii, 

Florida, and California (Yadava, 1996), Vietnam 

(Le et al., 1998), and Thailand (Tate, 2000). In 

part because it is a highly variable species for many 

morphological and horticultural traits, tolerant to 

environmental stress such as salinity (Nakasone 

and Paull, 1998), and its fruit has a high nutritional 

value; especially ascorbic acid, dietary fibers and 

some antioxidant compounds (Jimenez-Escrig et 

al., 2001). 

In Thailand, major guava production 

areas of nearly 8,000 ha are located in the Central 

and Western parts of the country, especially 

Nakhon Pathom, Samut Sakhon, and Ratchaburi 

provinces; however, a guava plant can grow and 

produce fruits well in most regions in Thailand 

throughout the year. Prominent commercial 

cultivars are ‘Paen Seethong’, ‘Klom Salee’, and 

‘Yen Song’. These white flesh cultivars account 

for more than 90% of fresh guava consumption. 

Department of Horticulture, Kasetsart University, Kamphaeng Saen campus, Nakhon Pathom 73140, Thailand. 

* Corresponding author, email: unaroj.b@ku.ac.th 


12 

At present, new cultivars with high nutritional 

value, excellent flavor, tolerant to biotic and abiotic 

stresses are increasingly important. 

Major fruit qualities are quantitative traits 

and the phenotypic expression is complex. 

Knowledge of genetic and environmental factors 

that influence their phenotypic expressions is 

fundamental for a successful breeding program. 

The phenotypic variance can be partitioned into 

variances within and between individuals when a 

trait is repeatedly measured on each individual. 

Repeatability is a ratio of the between individual 

variance to the phenotypic variance. Repeatability 

estimates are useful for making predictions on 

progress in measurement, determining an upper 

limit of heritability, and predicting future 

performance from past records (Becker, 1984; 

Falconer and Mackay, 1996). Knowledge of the 

repeatability of quantitative traits helps in selecting 

efficient breeding strategies, including optimal 

sample size and evaluation methods. Several fruit 

breeding programs such as persimmon (Yamada 

et al., 1993), strawberry (Sacks and Shaw, 1994), 

apricot (Akca and Sen, 1995), and peach (De 

Souza et al., 1998) used the benefits of 

repeatability. 

In the present research, the repeatability, 

optimal sample size, and phenotypic correlations 

of guava fruit traits were estimated to provide 

quantitative genetic information for guava 

breeding programs. 

MATERIALS AND METHODS 

Experimental materials 

Eleven randomly selected guava clones 

consisted of six white flesh dessert types (‘Klom 

Salee’, ‘Khoa Um-porn’, ‘Yen Song’, ‘Paen Yak’, 

‘Paen Seethong’ and ‘Na Suan’), one pink flesh 

dessert type (‘Keynok Daeng’), two maroon flesh 

dessert types (‘Daeng Siam’ and ‘Philippines’), 

and two pink flesh processing types (‘MCL-326- 

S’ and ‘PC 12-102’) from the guava germplasm 


collection of the Department of Horticulture, 

Kasetsart University, Kamphaeng Saen campus, 

Nakhon Pathom, Thailand were used. Guava trees 

were randomly planted in an experimental field 

(14°01′N lat., 99°58′E lon.) in December 1999, at 

a 4.0 m × 4.0 m spacing. The environmental 

conditions in the dry season (November to 

February) and the early rainy season (March to 

June) in 2001 had daily average max/min air 

temperature of 31.8/20.8°C and 33.6/24.5°C, daily 

average max/min RH of 95/50% and 95/58%, total 

precipitation of 9 mm and 265 mm, and daily 

average saturated light duration of 7.5 h d -1 and 

6.5 h d -1 , respectively. 

Sampling methods 

Fruit thinning by leaving one fruit per 

shoot was done in order to minimize the effects of 

over-cropping on fruit qualities such as size and 

sugar contents. Five fruits were randomly sampled 

from the same tree of each genotype in dry and 

early rainy season when the trees were 14 and 18 

months old, respectively. In general, guava trees 

propagated by air-layering or cutting begin to set 

fruits in two to three months after planting but most 

growers do not allow trees to set fruits until six to 

eight months old. The changing in skin color was 

used as harvesting indicator. White flesh fruits 

were harvested when their skin color changed from 

dark green to light green, maroon flesh fruits were 

harvested when their skin color changed from dark 

maroon to light maroon, and processing types were 

harvested when their skin color changed from dark 

green to yellow green. 

Fruit quality measurements 

Five physical fruit traits: fruit weight 

(FW), flesh thickness (FLT), flesh weight (FLW), 

seed cavity (central pulp) weight (SCW), fruit 

firmness (FF), and four chemical fruit traits: total 

soluble solids (TSS), titratable acidity (TA), juice 

acidity (pH), and ascorbic acid (AA) were 

evaluated. FW (g) and SCW (g) were measured

y digital balance (SK-5001, A&D, Japan). FLT 

(cm) was measured at equatorial plane with a 

caliper. FLW (g) was calculated by subtracting FW 

with SCW. FF (Newton; N) was determined on 

one side of fruit with fruit hardness tester (N.O.W., 

Japan) using 0.5 cm diameter probe after 0.3 cm 

skin was sliced off. Extracted juice from a flesh 

portion was used for determining the chemical 

traits. TSS was measured as °Brix with a 

temperature compensated hand refractometer 

(ATC-1E, Atago, Japan). TA (%) was determined 

by titration with 1.0 N NaOH and 1% 

phenolphthalein as an indicator using a digital 

burette (Burette digital III, Brand, Germany). The 

pH was determined using pH meter (pHScan 2, 

Eutech, Singapore). AA (mg) was estimated with 

oxalo-acetic acid solution and titration with 2, 6dichlorophenolindophenol-dye 

solution 

(A.O.A.C., 1990). 

Statistical analysis 

Data from each season was analyzed as 

a completely randomized design. An appropriate 

statistical model for expressing the phenotypic 

value of a trait is P ij = µ + g i + f ij (Becker, 1984). 

Where P ij is the phenotypic value of the j th fruit of 

the i th genotype, µ is the overall mean, g i is the 

random effect of the i th genotype, and f ij is the 

random effect of j th fruit in the i th genotype. The 

repeatability of the guava fruit traits was estimated 

using one-way analysis of variance procedure 

(Becker, 1984). The formula is written as 

2 

σ 

Repeatability = B 

2 2 

σB + σE 

where 

σ2 B is the between genotypic variance and σ2 E is 

the within genotypic variance. 

with standard error of repeatability 

S.E. = 

2 2 

[ 21 ( − R) ][ 1+ ( k − 1) 

R] 

kk ( −1)( n−1) 

Where k is the number of measurements (fruits) 


per genotype, n is the number of genotypes, and 

R is the repeatability value. 

The relative efficiency of measurements 

was estimated to obtain the optimal sample size 

for evaluating guava fruit traits. The formula is 

k 

Relative efficiency = 

1+ ( k − 1) 

R 

Where k is the number of measurements (fruits) 

and R is the repeatability value. 

In this research, optimal sample size was 

selected when the relative efficiency increased by 

less than 10% with an additional measurement. 

The phenotypic correlations among traits were 

estimated on a cultivar mean basis from two 

seasons using Pearson’s correlation coefficient (r) 

analysis. 


Variance components 

The phenotypic variance (σ 2 P) of guava 

fruit traits in the dry and the early rainy seasons 

was different (Table 1), indicating that seasonal 

environmental conditions influenced the 

phenotypic expression of guava fruit qualities. The 

combined analysis of variance (ANOVA) over 

seasons confirmed that several traits, especially 

the chemical traits, were affected by seasons (Table 

2). Therefore, it could be concluded that genetic 

expressions of chemical traits were highly 

sensitive to the changing of seasonal environments 

probably temperature and precipitation because 

these were clearly different between the two 

seasons as previously described in materials and 

methods. Rathore (1976) has reported that guava 

fruits harvested in spring, rainy and winter seasons 

in India had different levels of several chemical 

traits with rainy season fruits showing the lowest 

levels due to the fruits having the highest moisture 

contents. Effects of temperature on chemical 

compounds were also reported in several fruit 

crops such as apple (Hauagge and Cummins,

14 

2000), grapevine (Lavee, 2000), peach (George 

and Erez, 2000). 

Cultivars were different in most traits 

(Table 3) reflecting their difference in genetic 

background. The between genotypic variance 

(σ 2 B) in most traits was higher than the within 

genotypic variance (σ 2 E) in both seasons. The σ 2 B 

consists of genotypic variance (V G) + general 

environmental variance (V Eg) whereas the σ 2 E is 

the specific environmental variance (V Es) or 

sampling error associating with fruit set on 

different dates of the same plant. In this case, V Eg 

referred to the seasonal environmental conditions 

such as temperature, precipitation, relative 

humidity, and light duration, while V Es referred to 

the position and maturity stage of each fruit 

on the plant. Thus, guava fruit traits were 

influenced more by the seasonal environmental 

conditions than the fruit position or fruit maturity. 

In addition, σ 2 B was higher than σ 2 E in part due to 

the diverse guava cultivars used in this experiment 

(Table 4). Based on the ANOVA from Table 3, 

major part of σ 2 B of FW, FLT, FLW, and SCW 

could be the effect of V G, whereas of FF, TSS, 

pH, and AA could be the effects of V Eg. 


Repeatability 

The repeatability of guava fruit traits in 

the dry and the early rainy seasons for most traits 

were relatively high (Table 1). Guava had higher 

repeatability than apricot for FW and TSS (Akca 

and Sen, 1995), peach for FW, TSS, and TA (De 

Souza et al., 1998). This high repeatability in guava 

cultivars could be in part due to the diverse nature 

of their genetic background. To test this hypothesis, 

six commercial cultivars only from the white flesh 

dessert type (‘Klom Salee’, ‘Khoa Um-porn’, ‘Yen 

Song’, ‘Paen Yak’, ‘Paen Seethong’, and ‘Na 

Suan’) were used for repeatability estimation. This 

analysis yielded lower repeatability estimates than 

the 11 cultivars analysis (data not presented). The 

lower repeatability estimates for fruit traits from 

the commercial white flesh dessert type indicated 

that the genetic variance among these cultivars 

was small and consequently, guava breeding 

programs should include guava cultivars from 

other types such as processing cultivars and native 

cultivars to increase genetic variation of the 

breeding materials and to increase genetic gain in 

breeding program. Repeatability also establishes 

the upper limits of heritability (Becker 1984; 

Falconer and Mackay, 1996). Therefore, the 

Table 1 Variance components, repeatability (R), and standard error (S.E.) of repeatability of guava 

fruit traits in dry and early rainy seasons. 

σ 2 P σ 2 B σ 2 E R ± S.E. 

Trait 1 Dry Early rainy Dry Early rainy Dry Early rainy Dry Early rainy 

FW 68,027 65,260 47,214 52,684 20,813 12,575 0.69 ± 0.11 0.81 ± 0.08 

FLT 0.39 0.59 0.33 0.50 0.06 0.09 0.85 ± 0.07 0.85 ± 0.07 

FLW 48,247 45,996 33,305 37,152 14,942 8,844 0.69 ± 0.12 0.81 ± 0.08 

SCW 1,961 2,617 1,282 1,944 678 673 0.65 ± 0.12 0.74 ± 0.10 

FF 53.5 81.7 2.4 33.1 51.2 48.6 0.04 ± 0.11 0.40 ± 0.15 

TSS 2.34 1.66 1.22 0.82 1.12 0.84 0.52 ± 0.15 0.49 ± 0.15 

TA 0.33 0.12 0.28 0.10 0.05 0.02 0.85 ± 0.07 0.83 ± 0.07 

pH 0.15 0.19 0.13 0.16 0.02 0.03 0.87 ± 0.06 0.84 ± 0.07 

AA 2,745 1,420 2,099 1,141 645 280 0.76 ± 0.09 0.80 ± 0.08 

1 FW = fruit weight, FLT = flesh thickness, FLW = flesh weight, SCW = seed cavity weight, FF = fruit firmness, TSS = total 

soluble solids, TA = titratable acidity, pH = juice acidity, and AA = ascorbic acid.


Table 2 Analysis of variance showing mean squares, and probabilities of test statistics for guava fruit 

traits. 

Designation df Mean square Probability 

Fruit weight 

Cultivar 10 495,380

16 

repeatability of FW, FLT, FLW, SCW, TA, pH, 

and AA was relatively high for both seasons; 

actual heritability estimates for these traits 

would be expected to be relatively high. Similarly, 

the repeatability of FF and TSS was small for 

both seasons indicating that heritability estimates 

for these traits would also be relatively low. 

Based on the estimates of heritability, improving 

FF and TSS through selective breeding would 

be harder than for FW, FLT, FLW, SCW, TA, pH, 

and AA. 


Optimal sample size 

The repeatability of a trait is used to 

estimate the relative efficiency of measurement 

to determine the optimal sample size (Becker, 

1984). The relative efficiency of trait 

measurements with high repeatability was low, 

while those with low repeatability were high 

(Table 4). According to the formula, the relative 

efficiency with one fruit was 100%. The optimal 

sample size of measurement was determined when 

the increase in relative efficiency of measurement 

Table 3 Mean and standard error of fruit traits 1 in 11 guavas. 

Name Type FW FLT FLW SCW FF 

(g) (cm) (g) (g) (N) 

Klom Salee Dessert 585 ± 79 2.8 ± 0.1 502 ± 67 83.1 ± 14.3 26.8 ± 2.9 

Khoa Um-porn Dessert 674 ± 30 3.1 ± 0.2 584 ± 28 89.6 ± 9.5 33.5 ± 2.6 

Yen Song Dessert 746 ± 39 2.6 ± 0.1 592 ± 31 153.4 ± 10.7 33.0 ± 1.8 

Paen Yak Dessert 640 ± 75 2.4 ± 0.1 523 ± 62 117.5 ± 14.5 32.8 ± 2.1 

Paen Seethong Dessert 716 ± 69 2.6 ± 0.1 582 ± 61 131.8 ± 12.8 31.6 ± 1.6 

Na Suan Dessert 526 ± 27 2.6 ± 0.1 445 ± 25 81.6 ± 5.3 33.7 ± 2.2 

Keynok Daeng Dessert 142 ± 10 1.3 ± 0.0 113 ± 8 29.2 ± 2.3 28.6 ± 4.2 

Daeng Siam Dessert 388 ± 29 2.1 ± 0.1 323 ± 25 64.9 ± 6.3 27.6 ± 2.2 

Philippines Dessert 304 ± 26 1.7 ± 0.0 239 ± 20 65.7 ± 7.1 24.9 ± 1.9 

MCL-326-S Processing 381 ± 18 2.0 ± 0.1 316 ± 16 65.0 ± 6.3 20.5 ± 2.8 

PC 12-102 Processing 118 ± 8 1.2 ± 0.0 94 ± 6 23.2 ± 2.4 25.4 ± 4.5 

1 FW = fruit weight, FLT = flesh thickness, FLW = flesh weight, SCW = seed cavity weight, FF = fruit firmness. 

Table 3 Mean and standard error of fruit traits 1 in 11 guavas (continued). 

Name Type TSS TA pH AA 

(°Brix) (%) (mg) 

Klom Salee Dessert 6.8 ± 0.2 0.31 ± 0.05 4.3 ± 0.1 133 ± 9 

Khoa Um-porn Dessert 6.5 ± 0.4 0.38 ± 0.07 4.3 ± 0.1 120 ± 11 

Yen Song Dessert 7.4 ± 0.8 0.39 ± 0.07 4.3 ± 0.1 115 ± 12 

Paen Yak Dessert 7.1 ± 0.4 0.31 ± 0.03 4.3 ± 0.1 70 ± 8 

Paen Seethong Dessert 7.2 ± 0.5 0.39 ± 0.04 4.2 ± 0.1 68 ± 8 

Na Suan Dessert 7.0 ± 0.3 0.30 ± 0.02 4.4 ± 0.1 87 ± 7 

Keynok Daeng Dessert 9.4 ± 0.4 0.84 ± 0.07 3.8 ± 0.0 56 ± 2 

Daeng Siam Dessert 6.8 ± 0.3 0.40 ± 0.03 4.2 ± 0.1 126 ± 6 

Philippines Dessert 7.5 ± 0.4 0.37 ± 0.05 4.5 ± 0.1 82 ± 7 

MCL-326-S Processing 7.0 ± 0.4 1.75 ± 0.20 3.1 ± 0.1 119 ± 9 

PC 12-102 Processing 8.6 ± 0.5 0.56 ± 0.05 4.0 ± 0.1 184 ± 17 

1 TSS = total soluble solids, TA = titratable acidity, pH = juice acidity, and AA = ascorbic acid.

was less than 10% when an additional 

measurement was done. The optimal sample size 

for most traits in both seasons was about three 

fruits except for FF. The FF needed 51 and six 

fruits per genotype in the dry and the early rainy 

seasons, respectively to reach the same accuracy. 

Phenotypic correlations 

Phenotypic correlations as determined by 

the Pearson’s correlation coefficient (r) analysis 

between two traits may result from genetic 

associations due to linkage or pleiotropy (Falconer 

and Mackay, 1996). The four physical fruit traits; 


FW, FLT, FLW, and SCW were strongly positively 

correlated (r ≥ 0.80) among themselves except for 

FLT with SCW which was moderately positively 

correlated (r = 0.50) (Table 5). Therefore, using 

FW as a guide to screen for FLT, FLW, and SCW 

was a possibility since selection for higher FW 

should result in an increase in FLT, FLW, and SCW. 

However, selection for larger fruit may increase 

SCW and FW equally because the correlations of 

FW with SCW (r = 0.88) and with FLT (r = 0.81) 

were very similar. The correlations between all 

fruit traits associated with fruit size (FW, FLT, 

FLW, and SCW) with TSS and TA were negative 

Table 4 Relative efficiency of measurements and optimal sample size of guava fruit traits in dry and 

early rainy seasons. 

Relative efficiency Relative efficiency Optimal sample size 

with two fruits with three fruits 

Trait 1 Dry Early rainy Dry Early rainy Dry Early rainy 

FW 1.18 1.11 1.26 1.14 3 3 

FLT 1.08 1.08 1.11 1.11 2 2 

FLW 1.18 1.11 1.26 1.14 3 3 

SCW 1.21 1.15 1.30 1.21 3 3 

FF 1.92 1.42 2.78 1.67 51 6 

TSS 1.32 1.34 1.47 1.52 5 5 

TA 1.08 1.09 1.11 1.13 2 2 

pH 1.07 1.09 1.09 1.12 2 2 

AA 1.13 1.11 1.19 1.15 3 3 


soluble solids, TA = titratable acidity, pH = juice acidity, and AA = ascorbic acid. 

Table 5 Phenotypic correlations among guava fruit traits based on 11 cultivars in two seasons. 

Trait 1 FW FLT FLW SCW FF TSS TA pH 

FLT 0.81** 2 

FLW 0.99** 0.85** 

SCW 0.88** 0.50** 0.83** 

FF 0.18 ns 0.11 ns 0.18 ns 0.17 ns 

TSS -0.32** -0.38** -0.32** -0.28** 0.08 ns 

TA -0.27** -0.26** -0.27** -0.27** -0.05 ns 0.27** 

pH 0.32** 0.34** 0.32** 0.32** 0.00 ns -0.36** -0.84** 

AA -0.14 ns -0.01 ns -0.11 ns -0.25** 0.01 ns 0.22* 0.20* -0.20* 


soluble solids, TA = titratable acidity, pH = juice acidity, and AA = ascorbic acid. 

2 ns , *, ** are non significant and significant at p ≤ 0.05 and 0.01, respectively.

18 

(Table 5) indicating that selection for fruit size 

might reduce TSS or TA. Therefore, improving of 

fruit size and TAA or TA may be carried out in 

separate crossing plan and combining these traits 

later. However, the correlations between all fruit 

traits associated with fruit size with TSS and TA 

were quite low (-0.26 ≤ r ≤ -0.38), thus probably 

not of much practical importance. Three chemical 

fruit traits; TSS, TA, and AA were weakly 

positively correlated (0.20 ≤ r ≤ 0.27) among 

themselves, while these three traits were negatively 

correlated with pH. Most of physical traits, 

especially FW had no correlation with AA. 

One objective of this guava-breeding program 

is to develop new cultivars with larger fruit and 

high ascorbic acid. These results indicated that 

selection for large fruit with high ascorbic acid 

was feasible. 

CONCLUSION 

Repeatability estimates for FW, FLT, 

FLW, SCW, TA, pH, and AA were relatively high, 

indicating that response to selection for these traits 

would be realized in breeding program. Generally, 

three fruits per genotype provided sufficient 

efficiency for evaluating guava fruit traits. Most 

chemical traits had weak positive or negative 

correlation with fruit size, suggesting that early 

screening for chemical traits could be assayed 

indirectly using FW. 

ACKNOWLEDGEMENTS 

This research was financially supported 

by The Kasetsart University Research and 

Development Institute, a grant from The Graduate 

School of Kasetsart University, and the Thailand 

Research Fund. 

LITERATURE CITED 

Akca, Y. and S.M. Sen. 1995. Repeatability in the 


Prunus armeniaca L. and the importance 

repeatability in breeding selection. Acta Hort. 

384: 215-218. 

A.O.A.C.: Association of office analytical 

chemists. 1990. Official method of analysis. 

15 th ed. George Banta, Washington, DC. 

Becker, W.A. 1984. Manual of quantitative 

genetics. 4 th ed. Academic Enterprises, 

Washington, DC. 

Cobley, L.S. 1976. An introduction to the botany 

of tropical crops. 2 nd ed. (revised by W.M. 

Steel). Longman, New York. 

De Souza, V.A.B., D.H. Byrne and J.F. Taylor. 

1998. Heritability, genetic and phenotypic 

correlations and predicted selection response 

of quantitative traits in peach: II. An analysis 

of several fruit traits. J. Amer. Soc. Hort. Sci. 

123: 604-611. 

Falconer, D.S. and T.F.C. Mackay. 1996. 

Introduction to quantitative genetics. 4 th ed. 

Longman, New York. 

George, A.P. and A. Erez. 2000. Stone fruit species 

under warm subtropical and tropical climates, 

pp. 231-265. In A. Erez, ed. Temperate fruit 

crops in warm climates. Kluwer Acad. Publ., 

Dordrecht. 

Hauagge, R. and J.N. Cummins. 2000. Pome fruit 

genetic pool for production in warm climates, 

pp. 267-304. In A. Erez (ed.). Temperate fruit 

crops in warm climates. Kluwer Acad. Publ., 

Dordrecht. 

Jimenez-Escrig, A., M. Rincon, R. Pulido and F. 

Saura-Calixto. 2001. Guava fruit (Psidium 

guajava L.) as a new source of antioxidant 

dietary fiber. J. Agri. Food Chem. 49: 5489- 

5493. 

Lavee, S. 2000. Grapevine (Vitis vinifera) growth 

and performance in warm climates, pp. 343- 

366. In A. Erez (ed.). Temperate fruit crops 

in warm climates. Kluwer Acad. Publ., 

Dordrecht. 

Le, H.T., J. Hancock and T.T. Trinh. 1998. The 

fruit crop of Vietnam: introduced species and

their native relatives. Fruit Var. J. 52: 158- 

168. 

Nakasone, H.Y. and R.E. Paull. 1998. Tropical 

Fruits. CAB International, Wallingford. 

Rathore, D.S. 1976. Effect of season on the growth 

and chemical composition of guava (Psidium 

guajava L.) fruits. J. Hort. Sci. 51: 41-47. 

Sacks, E.J. and D.V. Shaw. 1994. Optimum 

allocation of objective color measurements for 

evaluating fresh strawberries. J. Amer. Soc. 

Hort. Sci. 119: 330-334. 


Tate, D. 2000. Tropical fruit of Thailand. Asia 

Books Co., Ltd., Bangkok. 

Yadava, U.L. 1996. Guava (Psidium guajava L.): 

an exotic tree fruit with potential in the south 

eastern United States. Hort Science 31: 789- 

794. 

Yamada, M., H. Yamane, K. Yoshinaga and Y. 

Ukai. 1993. Optimal spatial and temporal 

measurement repetition for selection in 

Japanese persimmon breeding. Hort Science 

28: 838-841.


Heritability, Heterosis and Correlations of Fruit Characters and 

Yield in Thai Slicing Melon (Cucumis melo L. var. conomon Makino) 

ABSTRACT 

Chamnan Iathet* and Kasem Piluek 

Two inbred lines (RM1 and LM2) of slicing melon (C. melo L. var. conomon markino) and 

their related progenies were determined for their quantitative inheritance and heterosis of fruit width, 

length, shape index, weight, fruit number per plant and yield and correlations among these traits. The 

results indicated that heritability based on fruit width, fruit length, fruit shape index, and fruit weight 

were relatively high at 0.60, 0.68, 0.55, and 0.71 respectively. The heritability as considered from fruit 

number per plant and yield were also high at 0.60 and 0.61. Heterobeltiosis was recorded to be 12.71% 

for fruit number per plant while total yield per plant showed 8.20% heterosis. The performance of fruit 

characters of F 1 hybrid was not exceeding over that of the better parent or equal to the mid-parent 

values. The width of marketable immature fruit showed negative correlation to fruit length and fruit 

shape. Fruit shape and size were not related to fruit number per plant and yield while fruit number per 

plant had highly positive correlation to yield per plant. 

Key words: Thai slicing melon, heritability, heterosis, correlation 

INTRODUCTION 

Thai slicing melon (Cucumis melo L. var. 

conomon makino) is simply called as melon 

cucumber. The vernacular name in Thailand is 

Taeng-Thai (Paje and Vossen, 1993). Thai people 

use immature fruit of slicing melon for 

consumption in the same way as cucumber. 

Cultivating areas of this melon in 

Thailand have not been formally reported. 

However, this fruit vegetable can be commonly 

found in wholesale markets and local markets 

especially in central regions. In the future, it is 

expected to become popular for consumption 

because the flesh is crispier than cucumber. For 

cultivation, it has good advantages on short crop 

duration, tolerance to disease and well-adapted to 

various conditions (Tindall, 1983). 

Since slicing melon is classified as a 

cross-pollinated crop, genetic and fruit character 

variability could be high among its population such 

as fruit shape, fruit skin color, and fresh color. This 

is due to freely cross pollinating between cultivars, 

so there are many intermediate types (George, 

1999). 

Although Thai slicing melon has an 

advantage in plant growth characters, it gives low 

yield and unattractive fruit characters resulting to 

lower price when compared to cucumber. The 

F 1-hybrid variety should solve these problems. 

This study was conducted to find the inheritance 

and correlation of fruit characters and yields that 

might be helpful in the breeding programs for 

improving marketable yield and fruit shape. 

Department of Horticulture, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 

Corresponding author, e-mail: chamnan@thaiseed.com 



Inbred lines selection 

Two inbred lines of round (RM1 as P 1) 

and cylindrical (LM2 as P 2) fruit types were 

crossed to make an F 1 hybrid (without reciprocal) 

and were followed by related progenies as F 2 

and back-crossed to both parents. Forty plants of 

P 1, P 2, F 1, F 2, BCP 1 and BCP 2 were grown in 

upper trailing system. Individual plant in each 

population was observed separately for all 

characters. The fruit characters were observed on 

immature fruits. Female flowers were marked 

with colored wire at blooming date and measured 

7 days after anthesis. All the observations were 

recorded as a mean of three fruits of each plant on 

immature fruit length, width, shape index (L/W 

ratio) and weight. 

Fruit collection and characterization 

The yield and fruit numbers per plant 

were recorded on individual plant of all 

populations. To assure that the immature fruits 

were harvested at the same stage, the female 

flowers were marked with colored wire at 

blooming date with the certain color designated 

for that day. The immature fruits were harvested 

at the stage of seven days after anthesis. Fruit 

harvesting was done daily for 20 days. The data 

of immature fruit weight per plant and fruit number 

per plant were recorded. 

To verify correlations of immature fruit 

width, length, shape and yield, the data of 40 plants 

of the F 2 population which assumed complete 

segregation of various fruit types were designed 

to analyze correlation values between fruit width, 

length, shape index, weight, fruit number per plant 

and marketable yield per plant. 

Data analysis 

The Microsoft Excel program was 

employed for the analysis of variance and 

correlation for all traits recorded. 


Narrow-sense heritability of fruit 

characters, yield and fruit number per plant 

was analyzed by Warner’s method (Warner, 1952). 

Heritability (h2 ( 1/ 2)D 

)= 

V F 

2 

(1/2) D = the additive genetic component 

of variance of F 2 

and V F2 = total within variance of F 2 

and (1/2) D = 2( V F2 ) – ( V B1 + V B2 ) 

Where VB1 and VB2 are the total within 

variance of the backcrosses of the F1 to the 

respective parents. 

Heterosis and heterobeltiosis (better 

parent heterosis) were calculated using the 

following formulae. 

Heterosis = ((F1 – MP)/MP) ×100 

Heterobeltiosis = ((F1 – HP)/HP) ×100 

Where MP = mid-parent mean and 

HP = better-parent mean. 

Favorable heterosis was assigned if F1 mean was significantly different from its midparent 

mean and favorable heterobeltiosis was 

given if F1 mean was significantly different from 

its better-parent mean. 


Heritability of fruit characters and yields 

Narrow sense heritability (h 2 ) of six 

characters was estimated from the data of 

populations derived from crossing between two 

lines P 1 (RM1) and P 2 (LM2). Generation means 

and heritability are presented in Table 1. 

Fruit size Heritability based on fruit 

width and length were as high as 0.60 and 

0.68 respectively. It signified the possibility of 

improving these characters as desired. The genetic 

gain of conducting mass selection at ten 

percent for these characters was estimated 

at 0.46 for width and 2.83 for fruit length. 

Fruit shape The shape index was

22 

measured to define fruit shape in terms of 

quantitative character. It was obvious from the 

generation mean comparison that P 1 and P 2 had 

the lowest and the highest index values of 1.2 

and 4.20, respectively. The narrow sense 

heritability determined for this character was as 

high as 0.55. The genetic gain of conducting mass 

selection at ten percent for this character was 

estimated at 0.71. 

Fruit weight The marketable fruit of 

cylindrical shape in P 2 parent made the highest 

fruit weight of 102.03 g. While the fruit of round 

shape in P 1 parent produced lower fruit weight at 

70.13 g and it was not statistically different from 

its F 1, F 2, and BC 1. The narrow sense heritability 

determined for fruit weight was as high as 0.71. 

The genetic gain of conducting mass selection at 

ten percent for this character was estimated at 

16.74 (Table 1). 

Fruit number per plant F 1 generation 

had the best performance considered from fruit 

number per plant, followed by that of BC 2, whereas 

the poorest performance was P 1. The narrow sense 

heritability examined for this character was as high 

as 0.60. The genetic gain when conducting mass 


selection at ten percent was estimated at 9.06 fruit 

per plant. 

Yield per plant P 2 parent gave the best 

performance for yield per plant (2.62 kg), followed 

by that of first filial (2.19 kg), whereas the P 1 

parent yield was the lowest (1.42 kg).The narrow 

sense heritability evaluated for this character was 

as high as 0.61. Thus the genetic gain of 

conducting mass selection at ten percents was 

estimated at 0.75 (Table 1). 

The narrow sense heritability estimated 

of fruit width, fruit length, fruit shape, and fruit 

weight were high (55-71%). This study gave the 

similar results of Lippert and Hall (1982) who 

suggested that heritabilities of fruit diameter, fruit 

length, fruit shape index, and fruit weight in 

muskmelon were relatively high (53-71%). 

When, Kalb and Davis (1984) estimated moderate 

heritability for fruit weight and shape index in 

bush muskmelon (23 and 36%, respectively). 

Pornsuriya (2005) discovered that the cylindrical 

fruit shape was incompletely dominant to 

round fruit and this character was governed by a 

single gene. 

Table 1 Heritability of marketable immature fruit characters and yields of slicing melon. 

Fruit characters Z Yield Z 

Width (cm) Length (cm) Shape index Weight (g) Fruit no. Total 

/plant. yield/plant.(kg) 

P1 4.95a 5.04e 1.02e 70.13c 22.90d 1.42d P2 3.39f 14.22a 4.20a 102.03a 27.15bc 2.62a F1 4.06d 7.19c 1.78c 71.22c 30.60a 2.19b F2 4.35c 7.49c 1.78c 71.55c 26.18bc 1.89c BCP1 4.56b 6.19d 1.39d 66.58c 25.28cd 1.71c BCP2 3.80e 9.47b 2.52b 79.55b 28.48ab 2.16b CV. (%) 7.87 18.82 22.73 14.59 24.50 26.7 

F-test ** ** ** ** ** ** 

h2 0.60 0.68 0.55 0.71 0.60 0.61 

∆G 0.46 2.83 0.71 16.74 9.06 0.75 

Z Means in a column followed by the same letter are not statistically different at 5% according to Duncan’s new multiple range 

test (DMRT) 

** Significant at P < 0.01 

∆G Genetic gain estimated for ten percent mass selection

The inheritance of fruit number per plant 

and yield were reported previously in muskmelon 

by Lippert and Hall (1982) that could be 

considered to be low (9-12%). This study revealed 

the contrast results for the characters which were 

estimated as high as 60 and 61%. It might be due 

to the fact that the studies by Lippert and Hall were 

observed in the mature fruit of muskmelon, 

whereas the immature fruit characters were 

determined in this study. 

In general, the narrow-sense heritability 

of fruit characters and yield were high, indicating 

that the genetic variance was highly expressed 

in the phenotype and that superior genotype 

might be identified efficiently through the 

evaluation of phenotype. The breeding program 

for incorporation of these traits could be 

manageable and signified the high potential for 

improving these characters through breeding 

method. Pure lines can be attained by direct 

phenotypic selection of all traits. 

Heterosis and mean performance of F 1 hybrid 

The F 1 hybrid giving the mean values of 

fruit width, length, weight, and shape index were 

not exceeding over that of the better parent or equal 

to the mid-parent values. These might infer to the 

role of incomplete dominant effect or possibly 

epistatic gene action. 

Favorable heterosis were indicated for 

fruit number per plant, yield per plant and, to a 

lesser extent, for fruit width (22.28%, 8.20% and 

-2.64%, respectively). The negative heterosis of 

fruit width, which made the fruit shape more 

slender and impressive for consumers, was the 

favorable character. Undesirable heterosis was 

found in fruit length, weight and fruit shape 

index (-25.28, -31.89, and -17.26%, respectively) 

(Table 2). The positive heterosis of yield per plant 

might be the reflection of heterobeltiosis of fruit 

number per plant whereas the negative 

heterobeltiosis was obtained from negative 

heterosis of fruit width, fruit length, fruit shape 


index, and fruit weight. 

Favorable heterobeltiosis was marked on 

fruit number per plant as 12.71 % (Table 2). This 

result suggested the successive way to improve 

yield of F 1 hybrid through fruit number per plant. 

Pornsuriya (2005) also reported positive heterosis 

and heterobeltiosis on this trait in many crosses of 

slicing melon. However, the improvement of fruit 

shape and fruit weight, fruit length of both inbred 

lines should be aimed for long and cylindrical fruit 

in order to meet the outstanding hybrid which will 

give higher yield and good fruit characters. 

Correlations of immature fruit sizes and 

marketable yield in F 2 population 

The correlations among fruit characters 

and yield are presented in Table 3. The fruit width 

had highly negative correlation to fruit length and 

shape (r = -0.74 and -0.83, respectively). It implied 

that the big diameter fruit character was inherited 

together with short fruit and the small diameter 

was inherited together with cylindrical fruit 

trait. The result of no relationship between fruit 

width and fruit weight might be due to the 

harvesting stage of marketable fruit given 7 days 

after anthesis and signified little variation in fruit 

diameter. 

The length of marketable immature fruit 

expressed highly positive correlation with fruit 

shape and fruit weight (r = 0.99 and 0.79, 

respectively). Increasing of fruit length gave the 

higher fruit shape index and fruit weight that 

offered positive correlation between fruit shape 

and weight (r = 0.70). Therefore, selection for 

longer fruit could be considered to improve 

immature fruit shape and weight. 

The marketable fruit width, length, and 

weight of this population revealed no correlation 

with fruit number per plant (r = -0.06, 0.17 and 

0.15, respectively). Total marketable yield per 

plant also had low correlation with immature 

fruit width, length and weight (r = -0.05, 0.29 

and 0.25, respectively). It indicated that breeding

24 

for higher fruit number and yield could be 

done with disregard of fruit sizes at the immature 

stage. 

Highly positive correlation coefficients 

were recorded for fruit number per plant and 

marketable yield that gave 0.88 for correlation 

coefficients (Table 3). The results indicated 

that fruit number per plant had a close association 

with the yield in this population. Components 

of fruit number per plant should be considered 

for yield improvement in Thai slicing melon 

breeding programs. 

CONCLUSION 

1. Inheritance as expressed by 

heritability of immature fruit sizes (fruit width, 

length and weight), shape, fruit number per plant 

and yield were high. 

2. Fruit number per plant of F 1 hybrid 

was higher than better parent. Other F 1 characters 


Table 2 Mean performance of F 1, mid-parents (MP) and better parent (BP), heterosis and heterobeltiosis 

of the studied characters. 

Characters F 1 MP BP Heterosis (%) Heterobeltiosis (%) 

Fruit width (cm) 4.06 4.17 4.95 -2.64 * -18.00 ** 

Fruit length (cm) 7.19 9.63 14.22 -25.28 ** -49.40 ** 

Fruit shape index 1.78 2.61 4.20 -31.89 ** -57.70 ** 

Fruit weight (g) 71.22 86.08 102.03 -17.26 ** -30.19 ** 

Fruit No./plant 30.60 25.03 27.15 22.28 ** 12.71 ** 

Yield/plant (kg) 2.19 2.02 2.62 8.20 * -16.50 ** 

* , ** Significant at P < 0.05 and P < 0.01, respectively 

Table 3 Correlation coefficients (r) of immature fruit characters and yields in F 2 populations. 

Immature characters Fruit length Fruit weight Shape index Fruit no./plant Y Yield (kg/plant) Z 

Fruit width X -0.74** -0.27 -0.83** -0.06 -0.05 

Fruit length X 0.79** 0.99** 0.17 0.29 

Fruit weight X 0.70** 0.15 0.25 

Shape index X 0.22 0.26 

Fruit no./plant 0.88** 

X Immature fruit characters measured at marketable stage (7 days after anthesis) 

Y Total marketable immature fruits from 20 harvests 

Z Total weight of marketable immature fruits from 20 harvests 

** Correlation is significant at 0.01 level 

were recorded in the range between P1 and P2. 

Fruit sizes, fruit shape, fruit weight, and yield were 

expressed under quantitative inheritance. 

3. Fruit number per plant was only one 

character that had high relationship to total yield 

of plant. The marketable fruit width, fruit length, 

fruit shape, and fruit weight did not affect the fruit 

number and yield per plant. 


George, R.A.T. 1999. Vegetable Seed 

Production. CABI International Publishing, 

Wallingford. 328 p. 

Lippert, L.F. and M.O. Hall. 1982. Heritabilities 

and correlations in muskmelon from parentoffspring 

regression analyses. J. Amer. Soc. 

Hort. Sci. 107(2): 217-221. 

Paje, M.M. and H.A.M. van der Vassen 1993. 

Cucumis melo L., pp.153-157. In J.S. 

Seimonsma and K. Piluek (eds.).Plant

Resources of South-East Asia No.8 : 

Vegetables. Prosea Foundation, Boger, 

Indonesia. 

Pornsuriya, P. 2005. Genetic Studies and 

Inheritance of Fruit Characters in Slicing 

Melon. Ph.D. Thesis, Kasetsart University, 

Bangkok. 

Tindall, H.D. 1983. Vegetables in the Tropics. 


The Macmillan Press Ltd., London. 533 p. 

Tomas J. Kalb and D.W. Davis. 1984. Evaluation 

of Combining ability, Heterosis, and Genetic 

Variance for Fruit Quality Characteristics in 

Bush Muskmelon. J. Amer. Soc. Hort. Sci. 

109(3): 411-415 

Warner, J.N. 1952. A method for estimating 

heritability. Agr. J. 44: 427-430.


Seed Development and Maturation of Eryngo (Eryngium foetidum L.) 

ABSTRACT 

Boonsong Ekpong 1 and Sutevee Sukprakarn 2 * 

Flowering pattern and seed development were studied in two eryngo accessions (EF006 and 

EF007) to determine the optimum time of harvest for seed production. Flowering peak was 117 days 

after transplanting (35.6 umbel heads) and 121 days after transplanting (38.7 umbel heads) in EF006 

and EF007 accessions, respectively. Seed in both accessions reached their physiological maturity 40 

days after anthesis, with 93.75-95.75% germination when the seed head had just begun to turn brownish 

black in color. Shattering, however, commenced 65 days after anthesis. 

Key word: seed development, optimum harvest time, seed germination 

INTRODUCTION 

Seed crops must be harvested when seed 

quality is at maximum. Some physiological 

markers have been used in order to identify the 

time of maximum seed quality. Harrington (1972) 

proposed the hypothesis that maximum seed 

quality was achieved when seeds reached 

maximum dry weight at the end of the seed filling 

period. This hypothesis has been supported by 

results in Zinnia violacea Cav. (Miyajima, 1997), 

Brassica napus L. (Still and Bradford, 1998) and 

Daucus carota L. (Nascimento et al., 2003). 

Seed maturity effects on seed quality are 

particularly evident in indeterminate crops. For 

example, flowering in carrots progresses from the 

1 st umbel to the 2 nd and 3 rd umbels during 

reproductive development (Oliva et al., 1988). In 

these cases, the effects of maturity on seed quality 

are exacerbated by the fact that these crops exhibit 

shattering (or shedding) of seeds as they mature. 

Thus, delaying harvest to allow later developing 

seeds to mature risks losing mature seeds to 

shattering. On the other hand, early harvesting 

results in more poor quality immature seeds that 

may be difficult to remove by standard cleaning 

and grading techniques (Bradford, 2004). 

Eryngo (Eryngium foetidum L.) is known 

to be a native of Central and Latin America, from 

southern Mexico to Panama through Brazil and 

from Cuba to Trinidad. It was introduced into 

South-East Asia by the Chinese as a substitute for 

coriander. It is an aromatic plant which is usually 

grown as a leafy vegetable used as a seasoning 

and for medicinal purpose in various countries 

such as Vietnam, India and in the Amazon region 

(De Guzmao et al., 2002). 

The effect of umbel head maturity at 

harvest on seed quality has not been reported for 

eryngo. This information is needed to determine 

the stage of harvest attaining maximum 

germination and vigor. Therefore, the objective of 

this work was to determine the physiological 

changes during seed development and correlated 

them with visual indicators to determine optimum 

harvest time. 

1 Faculty of Agriculture, Ubon Ratchathani University, Ubon Ratchathani 34190, Thailand. 

2 Department of Horticulture, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: agrsts@ku.a.c.th 

Received date : 0709/05 Accepted date : 30/01/06


Study sites 

The experiment was conducted at the 

Tropical Vegetable Research Center (TVRC), 

Kasetsart University, Khampaeng Saen Campus, 

Nakhon Pathom Province, Thailand, from August 

2003 to March 2004. 

Plant materials 

Seeds of two eryngo accessions (EF006 

and EF007) maintained at TVRC were sown in a 

peat medium in plastic trays containing 104 

inverted cone cells with a depth of 5.7cm and a 

volume of 20 cm 3 . Plants were thinned to one per 

cell 4 weeks after seeding and transplanted to 6inch 

round pots in a rain shelter 49 days after 

seeding. During flowering period, from August 

2003 to March 2004, all heads of the 3 rd umbel 

order were tagged individually and the ages of 

seeds monitored and expressed as days after 

anthesis. Seeds from 5 to 70 days after anthesis 

(at intervals of 5 days) were collected and 

analyzed. 

Flowering pattern 

Each week, newly-opened flowers of 

each accession were labeled with different colored 

nylon yarn. The days to reach anthesis for the first 

umbel to the tenth umbel, numbers flower/plant/ 

week (from 20 plants) and numbers of seed/umbel 

head (from 20 tertiary umbel heads) were recorded. 

Thereafter, days to peak flowering were 

determined for both accessions. 

Seed weight and moisture content 

Seed moisture content was measured by 

the hot-air oven method (ISTA, 2003). Preweighed 

fresh seeds were dried in an oven at 103° 

C for 16 hours, and dry weight was recorded. Seed 

moisture content was calculated using the 

following formula: 

% moisture content (w.b.) = 


Fresh weighr of seed - Dry weight of seed 

×100 

Fresh weight of seed 

Germination (GERM) and germination index 

(GI) 

The harvested seeds were transferred to 

a nylon net bag and dried in ambient conditions 

for 5 days. Seeds were then placed in a controlled 

room (20°C, 30% relative humidity) to minimize 

the loss of seed moisture content (SMC). When 

the seed lot weight remained constant (around 7% 

SMC), a germination test (four 100 seed samples) 

was carried out on two layers of germination paper 

in rectangular transparent plastic boxes (14 × 9 × 

5 mm) incubated at 20-30°C (16 hours and 8 hours, 

respectively). Normal seedlings were evaluated 

daily as described by International Seed Testing 

Association (ISTA, 2003) and germination index 

(GI) was calculated by the following formula as 

proposed by AOSA (1983). 

GI = (N i / D i) 

Where N i = number of normal seedlings 

counted at i th date. 

D i = Number of days required to the i th 

germination. 


Flowering pattern 

Eryngo flowers in a compound 

inflorescence starting from the first umbel to the 

second umbel and so on during reproductive 

development. Within each umbel head, the flowers 

open from the outside ring of florets to the central 

ring which is a characteristic of the Umbelliferae 

plants. 

Anthesis of the 1 st umbel head occurred 

73 and 77 days after transplanting in EF006 and 

EF007, respectively. The peak flowering was 

reached at 117 days after transplanting in EF006 

and 121 days after transplanting in EF007 

(Table 1). The flowering interval between the 1 st 

and the 2 nd umbels was about 5 days, and up to 

7-8 days between the 9 th and the 10 th umbel orders

28 

(Table 2). The numbers of flowering umbel head 

at differing umbel orders fitted the quadratic curves 

as indicated by the high coefficients of 

determination (r 2 ). According to these regression 

equations, the number of umbel heads rapidly 

increased up to 35.2 and 38.7 heads at the eighth 

umbel order in EF006 and EF007 accessions, 

respectively, and then there was a sharp decline in 

higher umbel orders (Figure 1). Thus, both 

accessions had a similar flowering pattern with 

only one peak. 


Because both accessions had only one 

peak flowering, determination of optimum harvest 

time was relatively easy. However, EF007 

accession had more umbel heads at peak flowering 

and seeds per umbel than EF006 accession. 

Therefore, for seed production at TVRC, it was 

possible that EF007 accession could have a higher 

yield than EF006, due to the fact that plants with 

high flower usually being correlated with high seed 

yield (Soffer and Smith, 1974). 

Table 1 Days to peak flowering, number of umbel heads at peak flowering and number of seeds per 

head at the 3 rd umbel order. 

Accessions Days to peak Number of heads/plant seeds/head 2/ 

flowering 1/ at flowering peak 

EF006 117 35.2 60.0 

EF007 121 38.7 73.2 

1/ Days after transplanting 

2/ Obtained from 3 rd umbel order 

Table 2 Days to anthesis 1/ at different umbel orders after transplanting. 

Accessions Umbel orders 

1 st 2 nd 3 rd 4 th 5 th 6 th 7 th 8 th 9 th 10 th 

EF006 73 78 84 91 98 104 111 118 125 133 

EF007 77 82 88 95 102 108 115 121 127 134 

1/ (50% of floret blooming) 

Figure 1 Relationship between number of umbel heads and umbel order of EF006 (A) and EF007 (B). 

Coefficients of determination (r 2 ) are significant at P ≤ 0.01.

Seed weight and moisture content 

The fresh weight of eryngo seeds differed 

significantly among the collection periods (Figure 

2). Initially, seed fresh weight increased from 0.55 

g in EF007 and 0.62 g in EF006 10 days after 

anthesis to a maximum of 0.66 g 20 days and 25 

days after anthesis, respectively. Seed dry weight 

increased from 0.10 g in EF007 and 0.12 g in 

EF006 10 days to the respective maximum of 0.40 

g and 0.41 g 40 days after anthesis. This indicated 

that seed of both accessions reached physiological 

maturity (PM) 40 days after anthesis. A similar 

pattern has been reported in other plant species 

e.g., Hibiscus esculentus L. (Demir, 1994), 

Brassica napus L. (Still and Bradford, 1998) and 

Daucus carota L. (Nascimento et al., 2003). 

However, at physiological maturity seed moisture 

content was still too high (23.7-25.9%) for 

harvesting. Thus, harvest date should be later than 

physiological maturity but before seed shattering. 

In the present study, it was found that 

seed shattering started 65 days after anthesis in 

both accessions. Therefore, there was at least a 

25-day window (days from PM to shattering) 

which mature seed could be harvested before seeds 

% 

100 

80 

60 

40 

20 

0 

A 

SMC 

FW 

DW 

10 15 20 25 30 35 40 45 50 55 60 65 70 

Days after anthesis 

SMC FW DW 


0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 

0 

gm/1000seeds 

began to shatter. 

Seed germination and germination index 

5 and 10 days after anthesis, the seed was 

still immature and showed no germination in both 

accessions. Seed started to germinate 15 days after 

anthesis or 25 days before it reached physiological 

maturity (Figure 3). Seeds harvested during 

15-25 days after anthesis were able to germinate, 

but the germination was lower than 80%. 

Maximum germination was obtained 40 days after 

anthesis (95.75% in EF006 accession and 93.75% 

in EF007 accession). In the last harvest, however, 

germination was obtained at 90% in both 

accessions. Germination index followed the same 

trend as germination in both accessions (Figure 

3). 

Changes in umbel head and seed color 

Determination of seed age by days after 

anthesis alone is considered an inaccurate marker 

of seed development as the environment and 

micro-environment in which the plants are grown 

affect the rate of seed development and maturation 

(Yang et al., 2004). Seed maturity must be related 

% 

100 

80 

60 

40 

20 

0 

B 

10 15 20 25 30 35 40 45 50 55 60 65 70 


SMC FW DW 

Figure 2 Changes in fresh weight (FW), dry weight (DW) and seed moisture content (SMC) during 

seed development (A) EF006 and (B) EF007 

SMC 

FW 

DW 

0.7 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 

0 

gm/1000seeds

30 

to the color of fruit and seed coat, fresh and dry 

weight and seed moisture content (Thomsen, 

2000). 

Changes in color of umbel head and seed 

during their development are illustrated in Table 

3. The umbel head exhibited a series of color 

changes from green 5-15 days after anthesis to 

yellowish brown 25-30 days after anthesis, and 

finally to brownish black at PM (40 DAA). Seed 

Germination (%) 

100 

80 

60 

40 

20 

0 

A 

GERM 

10 20 30 40 50 60 70 


GERM GI 

GI 


7 

6 

5 

4 

3 

2 

1 

0 

GI 

color followed the same trend as umbel head color 

and was brownish black at PM. At this stage, both 

umbel head and seed color clearly indicated that 

seeds were mature. Therefore, it is useful to know 

the morphological and physical changes that occur 

during the maturation of seed, as such information 

may determine the proper time of seed harvest 

(Yang et al., 2004). Similar color changes with 

maturity have been shown in many plants such as 

Figure 3 Changes in germination percentage (GERM) and germination index (GI) during seed 

development of (A) EF006 and (B) EF007 

Table 3 Changes in umbel head and seed color at different development stages of eryngo . 

Days to anthesis Umbel head color Seed color 

5 green light green 

10 green light green 

15 green yellowish green 

20 light green greenish yellow 

25 yellowish green yellowish brown 

30 yellowish brown light brown 

35 brown brown 

40 brownish black brownish black 







Germination (%) 

100 

80 

60 

40 

20 

0 

B 

GERM 

10 20 30 40 50 60 70 


GERM GI 

GI 

7 

6 

5 

4 

3 

2 

1 

0 

GI

light brown seed integument in Atriplex 

cordobensis (Aiazzi et al., 1998), brown seed in 

carrot (Rubatzky et al., 1999) and yellow pod in 

common vetch (Samarah et al., 2004). Therefore, 

the changes of umbel head and seed color would 

be a good indicator for predicting eryngo seed 

maturation. 

Overall, the optimum time for harvesting 

should take into account. Firstly, days to peak 

flowering and secondly days from anthesis to PM. 

From the present work (conducted from August 

to March), the optimum times for seed harvest in 

EF006 and EF007 accessions were 157 and 161 

days (days to peak flowering + days from anthesis 

to physiological maturity) after transplanting or 

204 and 208 days after seeding, respectively. 

However, if transplanting date was not in the same 

period as the present experiment, harvest date 

might be different. Thus, seed color should be used 

to predict the optimum harvesting time for eryngo 

seed production. 

CONCLUSION 

Eryngo reached peak flowering at the 8 th 

umbel orders or 117 days after transplanting in 

EF006 accession (35.6 umbel heads) and 121 after 

transplanting in EF007 accession (38.7 umbel 

heads). Seed of two eryngo reached physiological 

maturity (PM) 40 days after anthesis. Seed began 

to germinate 15 days after anthesis and maximum 

when reached physiological maturity. Changes in 

umbel head and seed color from brown to brownish 

black clearly indicated that the physiological 

maturity and germination maturity were reached. 

The optimum times for seed harvest in EF006 and 

EF007 accessions were 157 days and 161 days 

after transplanting, respectively. 


We would like to thank the 

Commissioner on Higher Education under the 


Academic Staff Development Program for 

providing financial support and the Tropical 

Vegetable Research Center (TVRC) for providing 

all facilities used in this study. 


Aiazzi, M.T., J.A. Arguello and J. A. Di. Rienzo. 

1998. Physiological maturity in seeds of 

Atriplex cordobensis (Gandoger et Stuckert): 

correlation with visual indicators. Seed Sci. 

& Tech. 26: 405-411. 

AOSA. 1983. Seed Vigor Testing Handbook. 

Association of Official Seed Analysts. 

Handbook. No 32. 

Bradford, K.J. 2004. Seed production and quality. 

Available source: http://veghome.ucdavid. 

edu/classes/Spring2004/AMR118, April 3, 

2005. 

De Guzman, C.C. and J.S. Siemonsma. 1999. 

Plant resources of South-East Asia No 13: 

Spices. Backhuys Publishers, Leiden 400 p. 

Demir, I.1994. Development of seed quality during 

seed development in okra. Acta Hort. 362: 

125-131. 

Harrington, J.F. 1972. Seed storage and longevity. 

In T.T Koslowski (ed.). Seed biology, Vol III. 

Academic Press, New York. 

ISTA. 2003. International Rules for Seed 

Testing. P.O. BOX 308, 8303 Bassersdorf, 

CH-Switzerland. 

Miyajima, D. 1997. Zinnia seed harvest time 

affects germination and plant growth. Hort. 

Science. 32(4): 687-689. 

Nascimento, W.M., J.V. Vieira and M.C. Alvares. 

2003. Physiological maturity of carrot seeds 

cv. alvorada under tropical conditions. Acta 

Hort. 607: 49-51. 

Oliva, R.N.,T.Tissaoui and K.J. Bradford. 1988. 

Relationships of plant density and harvest 

index to seed yield and quality in carrot. J. 

Amer. Soc. Hort. Sci. 113: 532-537. 

Rubatzky, V. E., C. F. Quiros and P. W. Simon.

32 

1999. Carrots and related vegetable 

Umbelliferae. Crop Production Science in 

Horticulture Series 10. CABI Publishing. 

New York. 

Samarah, N., N. Allataifeh, M.A. Turk and A.M. 

Tawaha. 2004. Seed germination and 

dormancy of fresh and air-dried seed of 

common vetch (Vicia sativa L.) harvested at 

different stages of maturity. Seed Sci. & Tech. 

32: 11-19. 

Soffer, H. and O.E. Smith, 1974. Studies on lettuce 

seed quality. III. Relation between flowering 

pattern, seed yield and quality. J. Amer. Soc. 

Hort. Sci. 99: 114-117. 

Still, D.W. and K.J. Bradford.1998. Using 

hydrotime and ABA-time models to quantify 


seed quality of brassicas during development. 

J. Amer. Soc. Hort. Sci. 123: 692-699 

Thomsen, K. 2000. Handling of desiccation and 

temperature sensitive tree seed. DSFC series 

of technical notes. TN56, Danida forest seed 

center, Humleback, Denmark. Cites. 

Q.H.Yang, S.H. Yin, S.Q. Song and W.H. Ye. 

2004. Development of desiccation tolerance 

and germination physiology of Crotalaria 

pallida Ait. seeds. Seed Sci. & Tech. 32: 99- 

111. 

Yang, Q.H., S.H. Yin, S.Q. Song and W.H. Ye. 

2004. Development of desiccation tolerance 

and germination physiology of Crotalaria 

pallida Ait. seeds. Seed Sci. & Tech. 32: 99- 

111.


Evaluation for Antibiosis Resistance in Cotton 

to Helicoverpa armigera Larvae 

Praparat Hormchan 1* and Arunee Wongpiyasatid 2 

ABSTRACT 

The antibiotic effect of the new cotton lines, NR1, NR2, NR3, NR4 and NR5 and the control, 

SR3, on selected developmental parameters of the 2 nd instar Helicoverpa armigera larvae was 

investigated. The larvae were fed on four diet types of each variety/line: fresh leaf, fresh square, artificial 

diet mixed with lyophilized powder of leaf and artificial diet mixed with lyophilized powder of square. 

RCB design was employed with four replicates. Each replicate had 10 diet cups of each variety/line 

with one larva. The results showed that the larval and pupal weights and the adult longevity of larvae 

reared on all diet types from NR1 were significantly lower and the larval period was significantly longer 

than those from SR3. The larvae produced on 4 diets from the other tested lines were found to have only 

some parameters significantly differed from the control. Fresh leaves/squares had better antibiotic 

effects on the bollworm development than the mixed diet with dried parts. 

Key words: antibiosis, cotton, Helicoverpa armigera, bollworm 

INTRODUCTION 

Cotton cultivation was abandoned in 

many areas in Thailand due to the increase of pest 

damage which severely reduces the yields and 

fiber qualities. One of the major insect pests of 

cotton is the american bollworm, Helicoverpa 

armigera. The primary reason for the importance 

of this insect is the difficulty the cotton growers 

have in controlling it once an economic level is 

reached plus no effective insecticides are available. 

Plant breeders and entomologists have devoted 

attention to find source of resistance as the 

control measure for this bollworm in the past 

several years. 

In the cotton plants, small though 

conspicuous, pigment glands are distributed 

throughout all portions of the plants. The principal 

constituent of these glands is gossypol (Lukefahr 

et al., 1966). Gossypol has been shown to confer 

some resistance to insects in cotton (Bottger et al., 

1964). The glandless cotton was developed by 

plant breeders and was proved to be susceptible 

to the insect pest, Heliothis spp. It has also been 

suggested that certain related substances also 

found in and around the pigment glands with high 

activity against cotton pest as well. The 

biochemical variation in plants may differentially 

affect the biology of insect feeding on them 

including insect growth and survival. The purpose 

of the study was, therefore, to determine the effect 

of several promising cotton lines for source of 

antibiosis on larval growth and developmental 

times of H. armigera. 

1 Department of Entomology, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Applied Radiation and Isotopes, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: agrprh@ku.ac.th 


34 


Laboratory test with H. armigera larvae 

Five cotton lines namely NR1, NR2, 

NR3, NR4 and NR5 derived from The Cotton 

Breeding Program of Dr. Ngarmchuen 

Ratanadilok, Department of Agronomy, Faculty 

of Agriculture, Kampaengsan Campus, Kasetsart 

University were tested for antibiosis against the 

2 nd instar bollworm larvae. SR3, one commercially 

grown variety was used for comparison in the 

following experiments. All variety/lines were 

glanded cotton. 

1) Feeding with fresh young leaves and squares 

(flower buds) 

Each cotton variety/line was planted in 

pot kept in the greenhouse. Young leaves and 

squares of 3-4 cm and 6 mm in diameter, 

respectively, from each line and the control were 

placed in 1 oz plastic cup (replicated 10 times per 

variety/line). They were washed in diluted sodium 

hypochlorite for 1 min to remove pathogen 

contamination. Then the leaves/squares were 

rinsed with tap water, spreaded on paper towels 

and allowed to dry. Moistened paper towel with 

distilled water was placed in the bottom of each 

cup to maintain approximately 70% RH. After 

that, one 2 nd instar bollworm larva was placed in 

each cup. The cups were kept at room temperature 

and were arranged in RCB design with 4 replicates. 

Seven days after feeding, the larval were 

individually weighed and were kept on feeding 

with fresh leaves/squares until they pupated and 

emerged into adult stage. The following data were 

recorded for each larva: larval weight after feeding 

on leaves/squares for 7 days, larval period, pupal 

weight and adult longevity. 

2) Feeding with lyophilized powders of leaf and 

square incorporated in artificial diets 

Artificial diet used in mixing with 

lyophilized powder of leaf/square was the 

modified formula prepared for the bollworm by 


The Insect Mass Rearing of Insect Pathology Lab, 

Department of Entomology, Ministry of 

Agriculture. Leaves/squares, of the same sizes as 

used in the preceding experiment, from the control 

and 5 tested lines were collected, washed and 

lyophilized, and reconstituted in the media using 

40 cc of distilled water to 10 g of powder. A 

small piece of 1 cm 3 mixed diet with leaves/ 

squares of each variety/line was placed in a small 

cup with one 2 nd instar bollworm larva. The same 

procedure and data recording as the preceding test 

were employed. 

RESULTS 

1. Feeding with fresh young leaves and squares 

(flower buds) 

Table 1 shows the mean larval and pupal 

weights, the mean larval period and the mean adult 

longevity of the 2 nd instar H. armigera larvae 

feeding on fresh leaves of SR3 and the tested lines, 

NR1, NR2, NR3, NR4 and NR5 for 7 days. It 

was found that both mean weights of the larvae 

and pupae on the control were significantly greater 

than those of the larvae reared on all new lines. 

The mean adult longevity of larvae on the tested 

lines were significantly shorter than that of the 

larvae feeding on SR3. Significant difference 

between NR1 and SR3 was also observed in the 

larval period. 

The mean weights of larva and pupa on 

fresh squares of the 5 lines were found to be less 

than those of the control (Table 2). Adults reared 

on squares of SR3 also lived significantly longer 

than those on the squares of the tested lines. Both 

results were similar to those on fresh leaves. As 

for the larval period of the larvae reared on NR3 

and NR5, they were found not to be significantly 

different from that of the larvae feeding on the 

control (Table 2). 

2. Feeding with lyophilized powders of leaf and 

square incorporated in artificial diets 

The effects of mixed diet with dried

leaves of SR3 and the tested lines on the 

development of bollworm larvae are shown in 

Table 3. The mean larval weight of the insects 

reared on NR3 was not significantly different from 

that on SR3 while only pupal weight of H. 

armigera produced on NR1 was significantly 

lower than that on the control. No significant 

difference was found between the adult longevity 

of the larvae reared on NR1 and NR3 but was 

significantly different from those of the larvae on 

the rest tested lines including SR3. Only the larval 

period of larvae produced on NR1 was 

significantly shorter than the control. 

Table 4 presents the effects of the control and 

tested lines on the developmental parameters of 

the 2 nd instar H. armigera larvae produced on 

artificial diet mixed with lyophilized square 


powder. It was found that only mean larval weight 

of larvae reared on NR3 did not significantly differ 

from that of larvae on SR3. The pupae of insects 

feeding on NR1 and NR5 were significantly 

smaller than those of the control and the tested 

lines. 

With the exception of mean larval period 

exhibited on NR3, those reared on all tested lines 

were significantly longer than that on SR3. The 

mean adult longevities of the bollworm on NR1, 

NR2 and NR3 were also observed to be 

significantly shorter than that of NR3 whereas 

only that of larvae produced on NR1 was 

significantly shorter than all. 

NR1 as made in all diet types was found 

to give the least development in all parameters to 

H. armigera. 

Table 1 Antibiotic effects of the control and tested cotton lines on development of second instar 

Helicoverpa armigera larvae reared on fresh leaves. 

Var/Line Means1/ Larval wt. Pupal wt. Larval period Adult longevity 

(g) (g) (day) (day) 

NR1 0.0200 a 0.1521 a 25.2500 c 11.2825 ab 

NR2 0.0215 a 0.1530 a 24.3125 bc 9.8250 a 

NR3 0.0202 a 0.1525 a 23.5200 ab 11.2325 ab 

NR4 0.0340 a 0.1526 a 23.9500 ab 13.8200 c 

NR5 0.0407 ab 0.1685 ab 24.4250 bc 12.4200 bc 

SR3 (control) 0.0793 c 0.2025 c 21.9325 a 16.0900 d 

1/ Means within column not followed by a common letter are significantly different at 0.05% level by DMRT. 


Helicoverpa armigera larvae reared on fresh squares. 



NR1 0.0440 a 0.1649 ab 23.4975 bc 11.8450 ab 

NR2 0.0759 abc 0.1461 a 22.1125 bc 11.2325 a 

NR3 0.0896 bc 0.1886 bc 21.4750 ab 14.5750 bc 

NR4 0.0734 ab 0.1924 c 22.2700 bc 13.4600 b 

NR5 0.0865 bc 0.1868 bc 21.4225 ab 15.6450 c 

SR3 (control) 0.1185 d 0.2288 d 19.7850 a 17.4375 d 

1/ Means within column not followed by a common letter are significantly different at 0.05% level by DMRT.

36 

DISCUSSION 

Extensive studies have been made on 

antibiosis as a source of resistance to Heliothis 

spp. One practical way to identify antibiosis 

resistance in cotton population was through 

estimates of the insect development . Because of 

the poor survival normally exhibited by the first 

instar larvae, studies on the effects of cotton line 

on Helicoverpa larval development were executed 

using the 2 nd instar larvae previously fed on 

artificial diet. Comparison of the experimental 

results revealed substantial differences in most 

parameters measured from the tested cotton lines, 

NR1, NR2, NR3, NR4 and NR5 with SR3, the 

control. However, it was found that only the larval 

and pupal weights and the adult longevity of the 



Helicoverpa armigera larvae reared on artificial diet mixed with lyophilized powder of leaf. 



NR1 0.0466 a 0.2244 a 20.225 a 10.2700 a 

NR2 0.0663 ab 0.2487 b 17.2150 b 14.2575 b 

NR3 0.1249 c 0.2311 ab 16.5800 ab 13.1500 b 

NR4 0.1956 d 0.2729 c 14.3600 c 11.7900 a 

NR5 0.1716 d 0.2466 b 14.7300 b 15.5075 bc 

SR3 (control) 0.1381 c 0.2644 bc 14.6650 bc 14.4900 b 



Helicoverpa armigera larvae reared on artificial diet mixed with lyophilized powder of square. 



NR1 0.1404 a 0.2433 a 17.0375 c 11.2150 a 

NR2 0.1490 a 0.2694 b 16.1175 bc 13.1187 b 

NR3 0.1934 c 0.2636 b 14.6550 ab 13.3975 b 

NR4 0.1694 b 0.2600 b 15.4862 bc 14.8825 bc 

NR5 0.2323 d 0.2403 a 15.2075 bc 16.6450 c 

SR3 (control) 0.1977 c 0.2659 b 13.3650 a 17.9025 c 


2 nd instar H.armigera larvae fed on both fresh 

leaves/squares and lyophilized leaf/square 

powders of NR1 incorporated in artificial diets 

were less and the larval period was longer than 

those observed in SR3. Larvae produced on some 

diets from the other tested lines had only some 

developmental characters better than the control. 

This should be because there were higher amounts 

of harmful chemicals in NR1 than in the rest as 

supported by Lukefahr et al. (1966) who indicated 

the difference in % gossypol in cotton lines to have 

different effects on tobacco budworm and 

bollworm. 

The fresh young leaves of all variety/ 

lines were observed to be more toxic to the larvae 

than the other feeds according to the development 

data acquired in this study. Also, the lyophilized

powders in artificial diets were found to have less 

effect to the insects than the fresh ones. The results 

were actually uncomparable in terms of effects to 

the insects because young leaves and squares, as 

for fresh feed as well as for lyophilization, were 

collected at different times of the day. This was 

supported by McKay (1974) who stated that the 

concentration of secondary compounds in most 

plants varied diurnally. Furthermore, the quantities 

of the compounds are likely to be altered by 

climatic and edaphic factors. However, it could 

be that young leaves had chemical activity greater 

than the other plant parts since they lack the tough 

tissues used as the defense against the insects. 

Maxwell et al. (1965 ) and Lukefahr et 

al. (1966) reported that the glandless cotton was 

more susceptible to attack by several insect species 

than their glanded counterparts which should be 

retained and utilized as a natural protective 

mechanism against cotton insects. The tested lines 

also had pigment glands in every part which should 

be good in insect protection as indicated by those 

authors. Lukefahr et al. (1966) demonstrated that 

the larval growth of the cotton bollworm, Heliothis 

zea and Heliothis virescens was related to the 

content of gossypol. Shaver and Parrot (1970) 

showed inhibited growth, larval development 

period and high mortality of Heliothis spp. on high 

gossypol diet. It was reported by Chan et al. (1978) 

that condense tannins were toxic to the bollworm 

by laboratory feeding tests, yet, the inheritance of 

condense tannins in cotton was not documented. 

Gossypol and tannin showed negatively correlated 

with weight gains of tobacco budworm larvae fed 

in the field on plant terminals ( Hedin et al., 1983). 

He also reported the contents of chrysanthemin 

and gossypol showing negatively correlated with 

tobacco budworm larval growth in the field while 

tannins were slightly positively correlated. It was 

observed that the flower buds from certain wild 

and primitive cottons showed more insecticidal 

activity than could be accounted for gossypol. The 

additional activity was ascribed for “X” factors 


which were identified as sesquiterpenoid quinone, 

hemigossypol and heliocides (Gray et al., 1976; 

Bell and Stipanovic, 1977). In accordance, it 

appeared that chemical resistance in cotton to the 

bollworm insect was due to multiple chemicals. 

According to the results, it was not 

known about the real nature of resistance in the 

tested lines, but some apparently possessed certain 

level of antibiosis as shown by the reduced growth 

for bollworm larvae as compared to the control, 

especially NR1. Yet, which chemicals produced 

such effects, further analysis must be pursued. Or 

other resistance mechanisms may be further 

studied. If the plant has antibiotic characteristics, 

the probability of finding a resistant factor will be 

greater if the plant part on which the insect feeds 

during the early life stage is investigated (Parrot 

et al., 1978). 

CONCLUSION 

Of all lines tested for antibiosis 

resistance, NR1 was the only one that had all 

parameters tested, the mean larval and pupal 

weights, the larval period and the adult longevity, 

significantly better than those of SR3, the control. 

Therefore, further investigation on the chemicals 

responsible for the cause of antibiosis should be 

pursued. 


Bell, A.A. and R.D. Stipanovic. 1977. The 

chemical composition, biological activity and 

genetics of pigment glands in cotton. Proc. 

Beltwide Cotton Production Res. Conf., Jan 

10-12. 

Bottger, G.T., E.T. Sheehan and M.J. Lukefahr. 

1964. Relation of gossypol content of cotton 

plants to insect resistance. J. Econ. Entomol. 

57: 283-285. 

Chan, B.C., A.C. Waiss, Jr. and M.J. Lukefahr. 

1978. Condense tannin, an antibiotic chemical

38 

from Gossypium hirsutum. J. Insect Physiol. 

24: 113-118. 

Gray, J.R., T.J. Mabry, A.A. Bell and R.D. 

Stipanovic. 1976. Para-hemi gossypolone: a 

sesquiterpenoid aldehyde quinone from 

Gossypium hirsutum. J. Chem. Soc. Chem. 

Commun: 109-110. 

Hedin, P.A., J.N. Jenkins, D.H. Collum, W.H. 

White and W.L. Parrott. 1983. Multiple factors 

in cotton contributing to resistance to the 

tobacco budworm, Heliothis virescens F. pp. 

375. In P.A. Hedin (ed.,) Plant Resistance 

to Insects, USDA, Syposium Series 208, 

Amer. Chem. Soc., Washington, D.C. 

Lukefahr, M.J., L.W. Nable and J.E. Houghtaling. 

1966. Growth and infestation of bollworms 


and other insects on glanded and glandless 

strains of cotton. J. Econ. Entomol. 59: 817- 

820. 

Maxwell, F.G., H.N. LaFever and J.N. Jenkins. 

1965. Blister beetles on glandless cotton. 

J. Econ. Entomol. 58(4): 792-3. 

McKay, D. 1974. Adaptive patterns in alkaloid 

physiology. Am. Nat. 108: 305-320. 

Parrott, W.L., J.N. Jenkins, J.C. McCarthy and L. 

Lambert. 1978. A procedure to evaluate for 

antibiosis in cotton to the tobacco budworm 

J. Econ. Entomol 71(2): 310-312. 

Shaver, T.N. and W.L. Parrott. 1970. Relationship 

at larval age to toxicity of gossypol to 

bollworms, tobacco budworms and pink 

pollworms. J. Econ. Entomol. 63: 1802-4.


Influence of Biotic and Chemical Plant Inducers on Resistance 

of Chilli to Anthracnose 

Le Thi Kieu Oanh 1 , Vichai Korpraditskul 1 , 

Chainarong Rattanakreetakul 1 * and Sirikul Wasee 2 

ABSTRACT 

The potential of biotic and chemical plant inducers to trigger the resistant in plant against 

disease was investigated. A biotic inducer as Trichoderma harzianum and chemical inducers as salicylic 

acid and Bion ® were tested for their induced resistance ability to protect anthracnose disease on two 

tested chilli cvs. Mae Ping and Man Dum in both seedling and mature stages. The necrotic symptom 

development on seedling of both tested chilli varieties after treated with Bion ® 0.05 mg a.i./ml was 

significantly lower than the others. The potential of anthracnose management on chilli cv. Man Dum 

was also treated by T. harzianum at 10 8 spore/ml and salicylic acid at 1 mM. Treatments with any plant 

inducer did not lead to an increase activity of polyphenol oxidase in chilli cotyledon 1, 3 and 5 day after 

challenge inoculation. In field experiment, the result showed that the foliar sprays with any of plant 

inducer reduced the infection percentage of anthracnose on chilli. Bion ® also showed the best percentage 

of disease reduction compared to the untreated treatment at 76.3% and 75.9% on both chilli varieties, 

whereas in treatment of 10 8 spore/ml T. harzianum and 2 mM salicylic acid were reduced 24.2, 57.4% 

and 19.8, 36.6% respectively. However, fruit yield of Bion ® treatment was very low and significantly 

different to control. The phytotoxicity occurred at high concentration. T. harzianum and salicylic acid 

treatment showed the fruit yield to be significantly increased in comparison with the control on chilli cv. 

Man Dum, but not on cv. Mae Ping. All plant inducers did not affect to fruit fly infestation on both 

varieties under field conditions. 

Key words: plant inducers, disease resistance, anthracnose, chilli 

INTRODUCTION 

Anthracnose disease, caused by 

Colletotrichum spp., is an important disease of 

chilli (Capsicum annuum L.) and many other kinds 

of plants in tropical areas worldwide. It occurs as 

a pre-harvest or post-harvest fruit rot, causing 

extensive losses in chilli grown during the warm, 

wet season in tropical and subtropical climates 

(AVRDC, 1998). Management strategies for this 

disease include the use of disease-free seeds and 

transplants, resistant varieties, and chemical 

sprays. The pathogenic fungi, C. capsici (Syd.- 

1913) E. J Butler & Bisby and C. gloeosporioides 

(Penz) Penz. & Sacc. in Penz., were reported as a 

causal agent of chilli anthracnose in Thailand 

(Sangchote, 1999). They are also the main causal 

agent in the tropical Asia (Kim et al., 1989). 

1 Department of Plant Pathology, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand. 

2 Tropical Vegetable Research Center, KURDI, Kasetsart University, Kamphaeng Saen, Nakhon Pathom 73140, Thailand. 

* Corresponding author, e-mail: aqrcnr@ku.ac.th 


40 

The natural resistance of plants to 

pathogens is based on constitutive and inducible 

barriers. Induced resistance may be expressed 

locally at the infection site as well as systemically 

that so-called Systemic Acquired Resistance 

(SAR) (Sticher et al., 1997). SAR is a broadspectrum 

resistance to pathogen includes viruses, 

bacteria and fungi that occurs when the plant is 

challenged by either a pathogen or chemical 

compounds such as salicylic acid (SA), jasmonic 

acid (JA), 2,6 dichloroisonicotinic acid (INA) and 

DL-3-amino-n-butanoic acid (BABA) (Kessmann 

et al., 1994; Oostendorp et al., 2001). Many studies 

showed that SA acts as an endogenous signal 

responsible for inducing SAR in plant (Malamy 

et al., 1990; Gaffney et al., 1993) and exogenous 

applied SA is also induced the defence in some 

crops (Manandhar et al., 1998; Murphy et al., 

2000). A new product, Benzo (1,2,3) thiadiazole- 

7-carbothioic acid S-methyl ester (BTH), has trade 

name as Bion ® in Europe or Actigard in the 

United States. Bion ® is the first synthetic plant 

activator that triggers plant to elicit the identical 

response as biological induction of SAR (Tally et 

al., 1999). At low application rates it activates 

resistance in many crops against a broad spectrum 

of diseases, including fungi, bacteria and viruses. 

Exogenous application of BTH induce resistance 

to disease on a range of commercial crops (Cole, 

1999; Godard et al., 1999; Romero et al., 2001). 

However, there is no report of the activity of BTH 

against anthracnose disease of the Thai chilli. 

Apart from the above mentioned 

chemicals, some biotic agents also can induce 

disease resistance in plants. Few micro-organisms 

have been fully commercialized for the control of 

foliar plant pathogens. One of the most studied 

commercial biocontrol agents is Trichoderma spp. 

which can be regarded as a model to demonstrate 

a control of pathogenic pathogens under 

commercial conditions (Elad, 2000; Harman et al., 

2004). In this study, the main objectives were to 

determine the efficiency of plant inducer i.e. 


including Bion ® , salicylic acid and T. harzianum 

against the anthracnose disease of chilli under 

greenhouse and field conditions and their effects 

to yield under field conditions. 


1. Resistant efficiency determination and 

experimental design 

1.1 Plant material 

The plant material was conducted in 

both greenhouse and under field conditions. Two 

chilli varieties were used in the experiment, 

namely Mae Ping (Mae Ping-80, from Known You 

Seed Co., Ltd - Thailand) and Man Dum (Hot 

pepper, from Chia Tai Co., Ltd - Thailand). 

1.1.1 Greenhouse experiment: Chili 

seeds were sown in polystyrene with mixed media 

containing soil, coconut husk and manure (1:1:1). 

Twenty-day-old seedlings with fully expanded 

cotyledon were tested in the greenhouse 

experiment. 

1.1.2 Field experiment The experiment 

was conducted at the field of Plant Pathology 

Department, Kasetsart University, Kamphaeng 

Saen campus from April to September 2004. Plants 

were grown from seeds in polystyrene trays in the 

greenhouse for 7 to 8 weeks, and then they were 

transplanted to the field at the end of May 2004. 

Sprinkler irrigation was applied to the crop as 

needed to maintain the optimum growth. Each 

treatment consisted of two rows, with 7 plants per 

row. Beds were mulched by black plastic after soil 

preparation. The spacing between beds was 50 cm, 

and spacing on the beds was 40 × 35 cm. Fertilizer 

N-P-K (15-15-15) was applied two times 2 nd week 

after transplanting, followed by 10 days after first 

application of each treatment. 

1.2 Plant inducer preparation 

The procedures of plant inducer 

preparation are shown in Figure 1. 

Three plant inducers, T. harzianum,

Plant inducers 

T. harzianum 

Salicylic acid 

Bion ®(50 WG) 

salicylic acid and Bion ® were tested in greenhouse 

and field conditions. They were prepared as in 

Figure 1. In greenhouse, inducer concentrations 

were prepared as follows: 1) T. harzianum of 10 8 

spore/ml, 2) 1 mM salicylic acid and 3) Bion ® 

(0.05 mg a.i./ml). Field experiment included 1) T. 

harzianum of 10 8 spore/ml, 2) 2 mM salicylic acid 

and 3) Bion ® (0.25 mg a.i./ml). However, when 

phytotoxic symptom appeared in Bion ® treatment, 

the concentration was then reduced to 0.1 mg a.i./ 

ml after the fourth sprays. 

1.3 Experimental design 

The experiment was arranged in 

Randomized Complete Block Design (RCBD) 

with 4 treatments. Ten seedlings per treatment were 

considered as each replication in the greenhouse 

experiment and field experiments. 

2. Plant inducer treatment and pathogen 

inoculation 

2.1 Greenhouse experiment 

The foliar spray on each treatment was 

applied as a fine mist one day before inoculation 

with a macerated mycelium and spore suspension 

of Colletotrichum capsici. Mycelia and spores of 

C. capsici were harvested from 10-day old culture 

on Potato Dextrose Agar (PDA). One droplet (15 

µl) of mycelium and spore suspension of C. capsici 

was placed on the wound which was punctured 

by needle at the middle of leaves. After inoculation, 


Dilute (10 9 spore/ml) 

water 

Dissolve 

water 

Dissolve 

water 

Concentration at 10 8 spore/ml (G, F) 

Adjust pH 6-6.5 

SAat 1 mM (G) 

SA1 

2N NaOH SA at 2 mM (F) 

Bion® at 0.05 mg a.i./ml (G) 

Bion ® at 0.1 and 0.25 mg a.i./ml (F) 

Figure 1 Diagram of plant inducer preparation (G: Greenhouse experiment; F: Field experiment). 

the seedlings were covered with plastic bags to 

keep moisture. 

2.2 Field experiment 

The foliar spray of each treatment was 

applied as a fine mist for nine times in total from 

the beginning of flowering stage. About 10-day 

interval of each spraying times were applied before 

harvesting and 7-day interval within harvesting 

period. The plant inducers were applied in the late 

afternoon. The disease incidence of anthracnose 

under natural infection was observed. 

3. Polyphenol oxidase activity on cotyledon 

Twenty-day-old seedlings were treated 

with plant inducers one day before inoculation by 

pathogen. Polyphenol oxidase (PPO) enzyme 

activity was checked 1, 3 and 5 days after 

inoculation. Cotyledons were homogenized in 

mortar using 5 ml of 0.1 M phosphate buffer 

(PBS), pH 6.5. Thereafter, the homogenates were 

centrifuged at 7,500 rpm, for 10 min. at 4°C. The 

supernatant (crude enzyme extract) was collected 

and kept in refrigerator at – 20°C. The reaction 

mixture consisted of 1 ml of catechol 0.006 M, 1 

ml of 0.1 M phosphate buffer (pH 6.5) and 3 ml 

sterilized water. Enzyme extract (200µl) was then 

added. PPO activity was determined by measuring 

with UV spectrophotometer at wavelength 310 nm 

after 30 min of incubation.

42 

4. Data collection and statistical analyses 

Disease incidence was observed under 

greenhouse and field conditions throughout the 

period of plantation. In greenhouse experiment, 

the disease severity of anthracnose on cotyledon 

was assessed based on the areas of necrotic 

symptom during 3 to 7 days after inoculation. 

The disease incidence of anthracnose in 

the field was expressed as the percentage of 

infected fruit per total fruits in each harvesting 

time. Total of disease incidence, fruit yield, weight 

of fruit and fruit fly infestation were inspected until 

the last harvesting time. Some other diseases were 

also observed in the field. The first harvest was 

started after the fourth spray of plant inducers. 

Ripened fruits were harvested weekly for 8 times 

on cv. Mae Ping and 7 times on cv. Man Dum. 

Fruit yield was calculated of 10 plants from each 

treatment. Data were analysed by analysis of 

variance (Proc ANOVA in SAS program). 

RESULTS 

1. Symptom of anthracnose disease observation 

and evaluation 

In the greenhouse, initial symptoms 

appeared on cotyledon as small halo around 

inoculated point from 3 days after inoculation. The 

disease symptom varied from minimum level to 

maximum level. It was clear that cotyledons 

treated with Bion ® showed less area of necrotic 

symptom than the others (Figure 2A). Under field 

conditions, it was noted that the symptom on cv. 

Mae Ping occurred mostly on ripened chilli fruits, 

whereas symptoms were found on both ripened 

and unripened chilli fruits on Man Dum variety 

(Figure 2B). 

2. Plant inducers and disease control on 

cotyledon 

The plant inducers as T. harzianum, 

salicylic acid and Bion ® were tested on induced 

resistant ability to anthracnose disease in chilli 


Figure 2 Various symptoms of anthracnose 

disease appeared on cotyledon from 3 

to 7 days after inoculation of Bion ® 

treatment (A1) and untreated control 

(A2). Symptom of anthracnose disease 

on chilli fruit (B) in the field of cvs. 

Mae Ping (B1) and Man Dum (B2). 

seedling. In Figure 3, the disease severity was 

expressed through areas of necrotic symptom on 

cotyledon 7 days after inoculation. The treatment 

of Bion ® expressed the smallest area of necrotic 

symptom in both varieties compared to the 

untreated control. The percentage of inhibition in 

each treatment was calculated based on data of 

the untreated control. It was found that Bion ® 

showed the highest activity of inhibition in the 

variety of Mae Ping and Man Dum. Figure 3B 

shows that percentages of inhibition in the 

treatment of Bion ® in Mae Ping and Man Dum 

were 71.3 and 84.3 %, respectively. 

3. PPO enzyme activity on cotyledon 

Chilli seedlings were treated with plant 

inducers one day before inoculated with a 

macerated mycelium of C. capsici. PPO enzyme 

activity was determined 1, 3 and 5 days after 

inoculation (Figure 4). The PPO enzyme activity

Areas of necrotic symptom (mm2) 

20 

15 

10 

5 

0 

Mae Ping 

10.62 

10.4 

13.25 

9.17 

2.8 

2.47 


8.62 

17.79 

Man Dum A B 

T. harzianum Salicylic a. Bion Untreated 

Percentage reduction 

compare to control (%) 

100 

80 

60 

40 

20 

0 

-20 

-40 

-60 

T. H SA Bion 

Plant inducers 

Figure 3 Areas of necrotic symptom (mm 2 ) on chilli cotyledon after 7 day after inoculation (A). 

Seedlings were treated with plant inducer as T. harzianum, salicylic acid, Bion ® and untreated 

control before inoculation with spore and macerated fungal mycelium suspension of 

C. capsici. Percent reduction of inducer treatment compared to the control (B). 

PPO activity (ASB / g sample) 

40 

30 

20 

10 

0 

30 

20 

10 

0 

30 

20 

10 

0 

1 day 

3 days 

5 days 

T1 T2 T3 T4 

Mae Ping Man Dum 

Treatment 

T1 T2 T3 T4 

Figure 4 Effects of plant inducer as T. harzianum (T1), salicylic acid (T2), Bion ® (T3), and the untreated 

control (T4) on PPO enzyme activity in cotyledon of chilli cvs. Mae Ping and Man Dum 1, 3, 

5 days after inoculation with C. capsici.

44 

was not significantly different within the 

treatments of plant inducer compared to the 

untreated control in both chilli varieties. However, 

the level of PPO enzyme activity was different 

between two chilli varieties in the same treatment 

of each time interval. PPO activity on cv. Man 

Dum was higher than that on cv. Mae Ping. In chilli 

cv. Man Dum, enzyme activity also increased over 

time from 1 to 5 days after inoculation, whereas 

on cv. Mae Ping enzyme activity did not change. 

4. The effect of plant inducers on anthracnose 

disease under field conditions 

The biotic and chemical plant inducers 

including T. harzianum, salicylic acid and Bion ® 

were tested on their inducible ability of 

anthracnose resistance on chilli under field 

conditions. 

Percentages of disease incidence among 

plant inducers in both varieties are shown in Figure 

5. The results showed that foliar sprays with any 

of plant inducers reduced anthracnose disease in 

each harvesting time, especially in treatment with 

Bion ® . The highest disease incidence was about 

80% in the untreated control at the first harvesting 

time from both tested chilli varieties. The trend of 

disease incidence along the harvesting time was 

similar in both varieties. However, disease 

incidence on cv. Mae Ping was higher than that 

Disease incidence (%) 

100 

80 

60 

40 

20 

0 

Mae Ping 

1 2 3 4 5 6 7 8 

Harvesting time (week) 

Stop spraying 


on cv. Man Dum. It was noted that after spraying 

stopped, the disease incidence increase in Mae 

Ping variety whereas the incidence seemed to 

sharply decreased in Man Dum variety. 

The effect of plant inducers on fruit 

anthracnose disease, fruit yield, weight of fruit and 

insect infestation are shown in Table 1. Among 

three plant inducers, Bion ® provided the best 

percentage of reduction of disease on both varieties 

(76.3 and 75.9%) compared to the control, 

followed by T. harzianum (24.2 and 57.4%) and 

salicylic acid (19.8 and 36.6%). However, certain 

phytotoxic effects resulted from Bion ® application 

was district at high concentration (0.25 mg a.i./ 

ml). Treated plant leaves became smaller, curved, 

and brittle. After phytotoxic symptom on leaves 

was observed, the concentration of Bion ® was 

reduced to 0.1 mg a.i./ml, thereafter no 

phytotoxicity occurred, plants produced more 

flowers and fruits resulting in fruit yield especially 

on chilli cv. Mae Ping variety (Figure 5). 

Apart from anthracnose, some other 

diseases and fruit fly infestation were also 

recorded. The result showed that all plant inducers 

were not relevant to control fruit fly infestation 

(Table 1). Some other diseases were observed in 

the field such as Choanephora blight caused by 

Choanephora cucurbitarum, collar rot caused by 

Sclerotium rolfsii and virus. 

Disease incidence (%) 

100 

80 

60 

40 

20 

0 

Man Dum 

1 2 3 4 5 6 7 


Stop spraying 

Figure 5 Disease incidence of chilli fruit anthracnose in each harvesting time on chilli cvs. Mae Ping 

and Man Dum treated by Trichoderma harzianum (—●—), salicylic acid (—■—), Bion ® 

(—▲—) and control (—*—).

5. Effects of plant inducer on fruit yield of chilli 

under field conditions 

The result of fruit yields of all treatments 

in both varieties is presented in Figure 6. Fruit 

yields of inducer treatment were not significantly 

higher than the untreated control in each harvesting 


time in both chilli varieties, except treatment with 

Bion ® . It was noted that after five weeks of 

harvesting time, the peak of Bion ® treatment in 

cv. Mae Ping increased sharply. However, the total 

fruit yield of Bion ® treatments was significant 

lower than the others. Treatment with T. harzianum 

Table 1 Effects of plant inducer on fruit anthracnose disease and the relation between disease incidence 

and fruit yield of chilli cvs. Mae Ping and Man Dum under field conditions. 

Treatment Disease% Reduction of Fruit yield Weight of Fruit weight Insect 

incidence1/ disease compared (g / plot) 1/ healthy fruit2/ (g/fruit) 1/ incidence1/ (%) to control (g) (%) 

Mae Ping variety 

T. harzianum 38.1 a 24.2 4883.6 a 3022.9 9.8 a 10.4 

Salicylic acid 40.3 a 19.8 4532.1 a 2705.6 10.1 a 10.0 

Bion ® 11.9 b 76.3 3092.9 b 2724.8 6.9 b 11.2 

Untreated control 50.3 a - 4492.1 a 2232.5 9.6 a 9.8 

CV (%) 12.6 9.6 19.8 6.6 16.8 

F-test 

Man Dum variety 

** * ns * ns 

T. harzianum 11.5 bc 57.4 3298.7 a 2919.4 5.4 8.2 

Salicylic acid 17.1 b 36.6 2684.5 b 2225.5 5.8 7.4 

Bion ® 6.5 c 75.9 780.0 c 729.3 4.8 8.3 

Untreated control 27.0 a - 2533.7 b 1849.6 5.9 6.3 

CV (%) 17.6 6.0 10.9 4.5 16.6 

F-test ** ** ** ns ns 

1/ Data (X) were collected until the last harvesting time and transformed to square root x + 05 . before analysis. 

2/ 

Means of nontransformed data were presented for clarity. Values in each column followed by the same letter were not 

significantly difference (P ≤ 0.05). 

Weight of healthy fruit was calculated based on percentages of uninfected fruit per total fruit yield. 

Fruit yield (kg/plot) 

2.0 

1.6 

1.2 

0.8 

0.4 

0.0 

1 

2 3 4 5 6 7 8 


2.0 

Mae Ping Man Dum 

1.6 


Figure 6 Fruit yield of chilli at each harvesting time on cvs. Mae Ping and Man Dum applied by 

T. harzianum (—●—), salicylic acid (—■—), Bion ® (—▲—) and control (—*—). 

Fruit yield (kg/plot) 

1.2 

0.8 

0.4 

0.0 

1 2 3 4 5 6 7 8

46 

showed the highest yield compared to the untreated 

control on cv. Man Dum (Table 1). In Mae Ping 

variety, the fruit yields and weights of healthy fruit 

of inducer treatment were not significantly higher 

than that of the untreated control. However, T. 

harzianum treatment also showed the potential for 

higher yield than the others. Fruit weight was not 

significantly increased among treatments and the 

control in each variety, but there was difference 

between two varieties. The fruit weight of chilli 

cv. Mae Ping was higher than that of cv. Man Dum. 

Therefore the total fruit yield of chilli cv. Mae Ping 

was also higher than that of cv. Man Dum. 

DISCUSSION 

In this research, from field trial all tested 

plant inducers induced the disease resistance of 

chilli against anthracnose when applied as foliar 

sprays. Among three inducers, Bion ® showed the 

best activity in controlling anthracnose compared 

to the untreated control, followed by T. harzianum 

and salicylic acid. Many reports demonstrated the 

efficiency of Bion ® to control plant disease caused 

by plant pathogenic fungi, bacteria and viruses in 

various of plant species (Görlach et al., 1996; Tally 

et al., 1999; Cole, 1999). Bion ® also activates a 

very wide spectrum of resistance under practical 

field condition e.g. on some vegetable and fruit 

crops (Tally et al., 1999). However, in this study 

the fruit yield of Bion ® treatment was reduced 

compared to the others. Phytotoxicity was 

observed in Bion ® treatment at high concentration 

(0.25 mg a.i./ml) that caused direct effect on the 

growth of chilli. After concentration was reduced 

to 0.1 mg a.i./ml, there was no phytotoxic symptom 

and plants produced more flower and fruit setting, 

especially on cv. Mae Ping. Hence, further studies 

will be needed to determine the optimal growth 

stage for application with the best rate and a 

method of suitable application. The optimization 

of these factors will lead to the development of a 

reliable and effective control against anthracnose 


infection under field conditions. 

Apart from Bion ® , T. harzianum and 

salicylic acid also have potential in management 

of anthracnose in chilli under field conditions. An 

activation of the natural plant defense system has 

been shown to occur upon exogenous application 

of salicylic acid (Manandhar et al., 1998; Malamy 

et al., 1990) and T. harzianum (De Mayer et al., 

1998; Elad, 2000). The first clear demonstration 

of induced resistance by Trichoderma sp. showed 

that treating soil with T. harzianum made 

leaves of bean plants resistance to disease that 

caused by Botrytis cinerea and Colletotrichum 

lindemuthianum, even Trichoderma sp. was 

present only on the root and not on the foliar 

(Harman et al., 2004). 

Disease incidence on chilli cv. Man Dum 

was lower than those on Mae Ping. It might suggest 

that Man Dum variety contained constitutive genes 

relevant to resistant response. Therefore, the less 

effect of inducers on disease resistance was found 

in this variety. The inducible ability of inducers 

on chilli cv. Mae Ping could be more consistent 

and reliable than that on cv. Man Dum under field 

conditions. 

The results in greenhouse showed that 

induced resistance abilities by plant inducers were 

different among plant inducers in each variety and 

between two chilli varieties. All plant inducers 

could induce disease resistance to anthracnose in 

cv. Man Dum, whereas only treatment with Bion ® 

could induce resistance to anthracnose in cv. Mae 

Ping. Among plant inducers, Bion ® showed the 

best controlling to anthracnose in both tested 

varieties. This was similar to the result in the field 

experiment. Tally et al. (1999) pointed out that 

Bion ® turned on various genes in plants which in 

turn produced specific proteins and enzymes. 

Many plant enzymes are involved in defense 

reactions against plant pathogens including 

oxidative enzyme such as polyphenol oxidase 

(PPO), peroxidase (PO) and other enzymes such 

as chitinase, β-1,3-glucanase and phenylalanine

ammonia lyase (PAL) (Avdiushko et al., 1993). 

These enzymes were reported to increase in plant 

which were treated with abiotic and biotic inducers 

in cucumber and tobacco plant. However, no clear 

relationships were found between increases in 

specific enzyme activities and the induction of 

resistance (Schneider and Ullrich, 1994). This 

suggested that resistance was not due to a particular 

enzyme but that the whole group could be 

responsible for defence. In current study, the 

results indicated that treatment with plant inducers 

did not significantly enhance the PPO activity on 

chilli cotyledon compared to the untreated control. 

This might suggest that PPO activity in tested chilli 

was not related to induce resistance treated by plant 

inducers. Further study for the relevant enzymes 

should be continued for experimentation. 

CONCLUSION 

Among tested plant inducers, Bion ® 

expressed the best activity for controlling 

anthracnose in both greenhouse and field 

conditions on two chilli cvs. Mae Ping and Man 

Dum. However, the effect of plant inducer on 

controlling anthracnose showed difference 

between greenhouse and field condition within the 

two tested varieties. Polyphenol oxidase enzyme 

activity in chilli cotyledon was not clearly affected 

by the tested plant inducers. 


The authors would like to thank the 

Evangelischer Entwicklungsdienst Church 

Development Service (EED) Stipendienreferat, 

Germany and Kasetsart University for supporting 

fund of this study and Dr. Chiradej Chamswarng 

of Kasetsart University who provided 

Trichoderma sp. for this study. 



Avdiushko, S.A., X.S. Ye and J. Kuc. 1993. 

Detection of several enzymatic activities in 

leaf prints cucumber plant. Physiological and 

Molecular Plant Pathology 42: 441-454. 

AVRDC, 1998. AVRDC Report 1997, pp. 54-57. 

Cole, D.L. 1999. The efficacy of acibenzolar-Smethyl, 

an inducer of systemic acquired 

resistance, against bacterial and fungal 

diseases of tobacco. Crop Protection 18: 267- 

273. 

De Meyer, G., J. Bigirimana, Y. Elad and M. Höfte. 

1998. Induced systemic resistance in 

Trichoderma harzianum T39 biocontrol of 

Botrytis cinerea. European Journal of Plant 

Pathology 104: 279-286. 

Elad, Y. 2000. Biological control of foliar 

pathogens by means of Trichoderma 

harzianum and potential modes of action. 

Crop Protection 19: 709-714. 

Gaffney, T., L. Friedrich, B. Vernooij, D. Negrotto, 

G. Nye, S. Uknes, E. Ward, H. Kessmann and 

J. Ryala. 1993. Requirement of salicylic acid 

for the induction of systemic acquired 

resistance. Science 261: 754-756. 

Godard, J.F., S. Ziadi, C. Monot, D. Le Corre and 

D. Silue. 1999. Benzothiadiazole (BTH) 

induces resistance in cauliflower (Brassica 

oleracea var. botrytis) to downy mildew of 

crucifers caused by Peronospora parasitica. 

Crop Protection 18: 397-405. 

G?rlach, J, S. Volrath, G. Knauf-Beite, G. Hengy, 

U. Beckhove, K.H. Kogel, M. Oostendorp, T. 

Staub, E. Ward, H. Kessmann and J. Ryals. 

1996. Benzothiadiazole, A novel class of 

inducers of systemic acquired resistance, 

activates gene expression and disease 

resistance in wheat. Plant Cell 8: 629-643. 

Harman, G.E., C.R. Howell, A. Viterbo, I. Chet 

and M. Lorito. 2004. Trichoderma species – 

opportunistic, avirulent plant symbionts. 

Nature reviews – Microbiology 2: 43-56.

48 

Kessmann, H., T. Staub, C. Hofmann, T. Maetzke, 

J. Herzog, T. Ward, S. Uknes and J. Ryals. 

1994. Induction of systemic acquired disease 

resistance in plants by chemicals. Annual 

Review of Phytopathology 32: 439 – 459. 

Kim, B.S., H.K. Park and W.S. Lee. 1989. 

Resistance to anthracnose (Colletotrichum 

spp.) in pepper, pp. 184-188. In Proceedings 

of the International symposium on 

Integrated Management Practices: Tomato 

and Pepper Production in the Tropics, 

AVRDC, Tainan, Taiwan. 

Malamy, J., J.P Carr, D.F Klessig and H. Raskin. 

1990. Salicylic acid: a likely endogenous 

signal in the resistance response of tobacco to 

viral infection. Science 259: 1002-1004. 

Manandhar, H.K, H.J Lyngs Jorgensen, S.B 

Mathur and V. Smedegaard-Petersen. 1998. 

Resistance in rice blast induced by ferric 

chloride, di-potasium hydrogen phosphate and 

salicylic acid. Crop Protection 17: 323-329. 

Murphy, A.M, L. J. Holcomble and J.P Carr. 2000. 

Characteristics of salicylic acid-induced delay 

in disease caused by a necrotrophic fungal 

pathogen in tobacco. Physiological and 

Molecular Plant Pathology 57: 47-54. 

Oostendorp, M., W. Kunz, B. Dietrich and T. 


Staub. 2001. Induced disease resistance in 

plants by chemicals. European Journal of 

Plant Pathology 107: 19-28. 

Romero, A.M., C.S Kousik and D.F Ritchie. 2001. 

Resistance to bacterial spot in bell pepper 

induced by acibenzolar-S-methyl. Plant 

Disease 85: 189-194. 

Sangchote, S. 1999. Anthracnose resistant in 

chilli. Progress report at the 20 th Anniversary 

of Kamphaeng saen Campus, Kasetsart 

University, 29 th November – 5 th December 

1999. 

Schneider, S. and W.R Ullrich. 1994. Differential 

induction of resistance and enhanced enzyme 

activities in cucumber and tobacco caused by 

treatment with various abiotic and biotic 

inducers. Physiological and Molecular Plant 

Pathology 45: 291-304. 

Sticher, L., B. Mauch- Mani and J.P Matraux. 

1997. Systemic acquired resistance. Annual 

Review of Phytopathology 35: 453 – 461. 

Tally, A., M. Oostendor, K. Lawton, T. Staub and 

B. Bassi. 1999. Commercial development of 

elicitors of induced resistance to pathogens, 

pp. 357-369. In A.A. Anurag, S. Tuzun and E. 

Bent (eds.). Induced plant defenses against 

pathogens and herbivores. APS press.


Incidence of Cymbidium Mosaic Virus and Odontoglossum Ringspot 

Virus on In Vitro Thai Native Orchid Seedlings and Cultivated 

Orchid Mericlones 

Yuphin Khentry 1 , Ampaiwan Paradornuwat 1 , Sureeya Tantiwiwat 2 , 

Salak Phansiri 3 and Niphone Thaveechai 1 * 

ABSTRACT 

Incidence of Cymbidium mosaic virus (CymMV) and Odontoglossum ringspot virus (ORSV) 

in 50 species, 16 genera of in vitro Thai native orchid seedlings and 44 cultivars, 12 genera of tissuecultured 

orchid plantlets was investigated using indirect enzyme-linked immunosorbent assay (ELISA). 

CymMV and ORSV were not detected in any of 1,000 axenic Thai native orchid seedlings. CymMV 

was detected in 6 genera namely Brassolaeliocattleya, Cattleya, Dendrobium, Epicattleya, Oncidium 

and Mokara at 27.6 % of 880 micropropagated cultivars of orchid samples. ORSV was not detected in 

any orchid sample. 

Key words: Cymbidium mosaic virus (CymMV), Odontoglossum ringspot virus (ORSV), orchid, ELISA 

INTRODUCTION 

Orchid plants are the members of 

Orchidaceae consisting more than 25,000 species, 

which are distributed almost all over the world 

but more abundantly in the tropics. There are 177 

genera, 1,125 species of orchids that originated in 

Thailand (Nanakorn and Indhamusika, 2000). 

Although there are large numbers of Thai wild 

orchids, deforestation and over-collection for 

commercial purposes have made many orchid 

species at risk of extinction. The genus 

Paphiopedilum spp. and Dendrobium cruentum 

are now included in Appendix I of Conservation 

on International Trade in Endangered Species of 

Wild Flora and Fauna (CITES). 

In vitro seed germination is being used 

for germplasm conservation and large-scale 

commercial propagation of orchid species. 

Furthermore, the tissue culture techniques, 

especially meristem culture technique have been 

employed for commercial multiplication of 

cultivated orchids. Tissue culture was introduced 

into the Thai orchid industry in late 1960’s and 

has become more and more important in Thailand. 

Moreover, it helps Thailand to continue to be the 

leader in tissue culture in Southeast Asia and the 

world leader in orchid export. The total value of 

fresh cut orchids, which were exported to 87 

countries in 2004, was 2,136 millions Baht (Office 

of Agricultural Economic, 2005). Orchids are also 

shipped out as pot plants and seedlings in flasks. 

1 Department of Plant Pathology, Faculty of Agriculture, Kasetsart University, Bangkok 10900, Thailand 

2 Department of Botany, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand 

3 Scientific Equipment Center, KURDI, Kasetsart University, Bangkok 10900, Thailand 

* Corresponding author, e-mail: agrnpt@ku.ac.th 


50 

There are approximately 60 commercial orchid 

micropropagation laboratories in Thailand. Their 

capacity may range from a few hundred thousand 

plants to over 50 million plants per year 

(Wannakrairoj, 2004). Currently, the international 

trade competition is increasing. Many countries 

restrict imports based on quality and particularly 

demand pathogen-free orchids. 

In Thailand and in many other orchidexporting 

countries, CymMV and ORSV are the 

most economically important viruses. These 

viruses reduce the growth of infected orchid plants 

as well as the quality of flowers, which effect to 

orchid industry (Person and Cole, 1986). The most 

efficient way for the spread of CymMV and ORSV 

in the orchid industry is by mass propagation of 

orchid plantlets from an infected mother stock 

through mericloning tissue culture process (Chang, 

2004). Many investigations have shown that orchid 

seedlings exported from Thailand are infected with 

these viruses (Chang et al., 2003). In order to 

produce high quality virus-free orchid plants for 

the domestic and international market, indexing 

for the absence viruses in the propagation plant 

material is an important approach. 

In the purpose of this study, the status of 

CymMV and ORSV in axenic Thai native orchid 

seedlings and cultivated orchid mericlones was 

determined using indirect enzyme-linked 

immunosorbent assay (ELISA). ELISA is still the 

most widely used method for practical plant virus 

detection throughout the world because of its 

accuracy, simplicity and low cost. ELISA has been 

used to detect CymMV and ORSV. It could detect 

purified CymMV at a low concentration of 50- 

100 ng/ml (Vejaratpimol et al., 1998) and 2.5 ng/ 

ml of purified ORSV (Wong and Lim, 1994). 


1. Detection of CymMV and ORSV in in vitro 

Thai native orchids plantlets 

Forty-five species of Thai native 


seedlings, grown in vitro, were kindly supplied by 

the ‘Production of Good Varieties and Diseasefree 

Plants for Export with Emphasis on Orchids 

Project’, Department of Plant Pathology, Faculty 

of Agriculture, Kasetsart University, Bangkok, 

Thailand. Five species were obtained from tissue 

culture laboratories in Bangkok and Ratcha Buri 

province. One leaf from each of 20 plants, of each 

species, was randomly selected, cut with sterilized 

scissor and individually assayed for CymMV and 

ORSV by indirect-ELISA technique as described 

by Clark and Adams (1977) with slight 

modifications. A 100 µg of leaf tissue was ground 

in 1 ml of 0.5M carbonate-coating buffer, pH 9.6. 

A 100 ml of each sample (without filtration) were 

loaded into ELISA wells (Costar, USA). The 

negative control was done with in vitro seedlings 

of Dendrobium hybrids. The positive control of 

CymMV was obtained from CymMV infected 

Dendrobium Sakura which showing chlorotic 

mosaic on leaves (Figure. 1A) and partially 

purified CymMV (Figure 1C). The positive 

controls of ORSV were obtained from ORSVinfected 

Cattleya spp. showing color breaking on 

flower (Figure 1B) and partially purified ORSV 

(Figure 1D). The coating plates were incubated in 

a moist chamber at 37°C for 1 hour and then, these 

plates were decanted and washed with phosphatebuffered 

saline, containing 0.05% (v/v) Tween 20 

(PBST). The healthy plants were ground in PBST 

containing 2% ovalbumin (Sigma # 5253, Sigma 

Chemical, St. Louis, USA), in a ratio of 1:30 

(w/v) and filtrated. Polyconal antibodies against 

CymMV and ORSV (provided by P. Hamelink, 

Department of Plant Pathology, Kasetsart 

University, Bangkok, Thailand) were diluted to 

1:1000 in the filtrated solution. A 100 µl of the 

diluted antiserum was added to each well and 

incubated at 37°C for 1 h after that rinsing three 

times with PBST at 3-minutes intervals. Goat-antirabbit 

gamma immunoglobulin alkaline 

phosphatase conjugate (Sigma # A 8025, Sigma 

Chemical, St. Louis, USA) was diluted to 1:2000


Figure 1 CymMV-infected Dendrobium Sakura showing mosaic on leaves (A), ORSV-infected Cattleya 

spp. showing color breaking on flower (arrow)(B), electron micrographs of CymMV particles 

(C), electron micrographs of ORSV particles (D), symptom on leaves of infected CymMV in 

vitro plantlets are not smooth, dark green areas raised somewhat above the light green tissue 

as longitudinal ridges and bumps (E) and CymMV infected plantlets showing mosaic on 

leaves (arrow)(F).

52 

in PBST containing 2% ovalbumin and added to 

each well, incubated at 37°C for 1 hour and then, 

repeatedly washed as above. A 100 µl aliquot of 

freshly prepared substrate (10 mg p-nitrophenyl 

phosphate; Sigma # N 6260, Sigma Chemical, St. 

Louis, USA) was dissolved in 10 ml of substrate 

buffer (9.7% diethanolamine, 0.02% NaN 3, pH 

9.6) and added to each well. They were incubated 

in a moist chamber at 37°C for 30 minute, after 

that 50 µl of 3 M NaOH was added to all the wells 

to stop further enzymatic reactions. Absorbance 

value of each well was measured at 405 nm with 

an ELISA microplate reader (GDV model DV 

990BV4, Italy). The color reactions produced by 

tested samples were compared with known 

negative control wells. The mean absorbance of 

infected samples that exceeded two fold of mean 

absorbance of the healthy samples was considered 

as a positive reaction (Satula et al., 1986). 

2. Detection of CymMV and ORSV in in vitrocultivated 

orchids plantlets 

Forty-four cultivars of in vitro-cultivated 

orchid plantlets were obtained from commercial 

tissue culture laboratories in Bangkok, Non 

Thaburi, Pathum Thani and Nakhon Pathom 

provinces. One leaf from each of 20 plants of each 

cultivar was randomly selected and separately 

assayed for CymMV and ORSV by indirect- 

ELISA as described previously. 

3. Bioassay 

Bioassay was used to confirm the ELISA 

results. Eighty samples that gave positive (66) and 

negative (14) reactions by indirect-ELISA were 

examined for presence of CymMV. Forty-fiveday-old 

Cassia occidentalis plants were used as 

indicator plants for CymMV. The orchid leaves 

were ground in a 0.01 M phosphate pH 7.0 buffer 

solutions at ratio 1:2 and used as inoculum. The 

inoculum was kept cool and used immediately. The 

C. occidentalis leaves were dusted with 600-mesh 

carborundum. The inoculum was rubbed gently 


onto the leaf surface with a sterile cotton bud until 

leaves appears wetted. The inoculated leaves were 

rinsed with distilled water. Inoculum from infected 

CymMV orchid plants and plain buffer solution 

were included as positive and negative controls, 

respectively. The indicator plants were maintained 

in a greenhouse and symptoms were observed after 

incubation for 3-5 days, specifically for small, 

discrete, brown lesions, which indicated infection. 

The experiment was triplicate. 

RESULTS 

A total of 1,000 tissue-cultured seedlings 

of Thai native orchid and 880 in vitro cultivated 

orchid plantlets were assayed for CymMV and 

ORSV using indirect ELISA. It was found that 

CymMV and ORSV were not detected in any 

axenic Thai native orchid seedlings (Table 1). Out 

of 880 cultivated orchid samples, 243 (27.6%) 

were positive for CymMV while ORSV was 

negative. CymMV was detected in 6 out of 10 

genera namely Brassolaeliocattleya, Cattleya, 

Dendrobium, Epicattleya, Oncidium and Mokara. 

The incidence of CymMV infection was in ranged 

between 50 % and 100 %. High infection rates 

were observed in Brassolaeliocattleya Alma Kee, 

Dendrobium Chanel, Dendrobium Chao Praya, 

Dendrobium Pravit White, Dendrobium Sakura 

and Dendrobium Shavin White (Table 2). Leaves 

of infected CymMV in vitro plantlets are not 

smooth, dark green areas raised somewhat above 

the light green tissue as longitudinal ridges and 

bumps (Figure 1E). CymMV-infected plantlets 

also showed mosaic on leaves (Figure 1F). 

All sixty-six positive samples (of eighty) 

were screened for CymMV using ELISA also 

positive by bioassay. Local, necrotic lesions 

appeared on the inoculated leaves of C. 

occidentalis after 4-6 days of incubation in the 

greenhouse (Figure 2).


Table 1 Incidence of Cymbidium mosaic virus (CymMV) and Odontoglossum ringspot virus (ORSV) 

in 50 species of in vitro seedlings of Thai native orchids using indirect enzyme-linked 

immunosorbent assay (ELISA). 

Species No. of plants infected with No. of non-infected Percentage of 

CymMV ORSV plants infection 

1. Aerides falcata Lindl. 0 0 20 0 

2. Aerides houlettiana Rchb.f. 0 0 20 0 

3. Aerides multiflora Roxb. 0 0 20 0 

4. Ascocentrum ampullaceum Schltr. 0 0 20 0 

5. Ascocentrum miniatum Schltr. 0 0 20 0 

6. Bulbophyllum morphologorum Krzl. 0 0 20 0 

7. Coelogyne cumingii Lindl. 0 0 20 0 

8. Coelogyne rochussenii 0 0 20 0 

9. Cymbidium aloifolium (Linn.) Sw. 0 0 20 0 

10. Dendrobium aggregatum Roxb. 0 0 20 0 

11. Dendrobium anosmum Lindl. 0 0 20 0 

12. Dendrobium bellatulum Rolfe 0 0 20 0 

13. Dendrobium chrysotoxum Lindl. 0 0 20 0 

14. Dendrobium crepidatum Lindl. 0 0 20 0 

15. Dendrobium cruentum Rchb. f. 0 0 20 0 

16. Dendrobium crystallinum Rchb. f. 0 0 20 0 

17. Dendrobium draconis Rchb. f. 0 0 20 0 

18. Dendrobium farmeri Paxt. 0 0 20 0 

19. Dendrobium formosum Roxb. 0 0 20 0 

20. Dendrobium friedricksianum Rchb. f. 0 0 20 0 

21. Dendrobium hercoglossum Rchb. f. 0 0 20 0 

22. Dendrobium infundibulum Lindl. 0 0 20 0 

23. Dendrobium lindleyi Steud. 0 0 20 0 

24. Dendrobium nobile Lindl. 0 0 20 0 

25. Dendrobium palpebrae Lindl. 0 0 20 0 

26. Dendrobium pendulum Roxb. 0 0 20 0 

27. Dendrobium primulinum Lindl. 0 0 20 0 

28. Dendrobium pulchellum Roxb. ex Lindl. 0 0 20 0 

29. Dendrobium scabrilingue Lindl. 0 0 20 0 

30. Dendrobium secundum (Blume) Lindl. 0 0 20 0 

31. Dendrobium sulcatum Lindl. 0 0 20 0 

32. Doritis pulcherrima Lindl. 0 0 20 0 

33. Eulophia andamanensis Rchb.f 0 0 20 0 

34. Gastrochillus calceolaris 0 0 20 0 

35. Grammatophyllum speciosum Blume. 0 0 20 0 

36. Paphiopedellum coerulea Griff 0 0 20 0 

37. Paphiopedellum concolor 0 0 20 0 

38. Phalaenopsis cornucervi Pfitz. 0 0 20 0 

39. Rhynchostylis coelestis Rchb.f. 0 0 20 0 

40. Rhynchostylis gigantea (Lindl.) Ridl. 0 0 20 0 

41. Rhynchostylis gigantea var. harrisonianum 0 0 20 0 

42. Rhynchostylis gigantea var. petotiana 0 0 20 0 

43. Rhynchostylis gigantea var. rubra Sagarik 0 0 20 0 

44. Rhynchostylis retusa (L.) Blume. 0 0 20 0 

45. Spathoglottis plicata Blume. 0 0 20 0 

46. Vanda coerulea Griff. Ex Lindle. 0 0 20 0 

47. Vanda brunnea Rchb.f 0 0 20 0 

48. Vanda denisoniana Bens. & Rchb. 0 0 20 0 

49. Vanda livouvillei Finet.Benson & Rchb.f 0 0 20 0 

50. Vandopsis gigantea (Lindl.) Pfitz. 0 0 20 0 

Total 0 0 1,000 0

54 


Table 2 Incidence of Cymbidium mosaic virus (CymMV) and Odontoglossum ringspot virus (ORSV) 

in 44 cultivars of in vitro orchid plantlets using indirect enzyme-linked immunosorbent assay 

(ELISA). 

Cultivars No. of plants infected with No. of non-infected Percentage of 

CymMV ORSV plants infection 

1. Brassolaeliocattleya Alma Kee 20 0 0 100 

2. Brassolaeliocattleya Arom Gold 0 0 0 0 

3. Brassolaeliocattleya Elizabeth Hearns 10 0 10 50 

4. Brassolaeliocattleya Free spirit 0 0 20 0 

5. Brassolaeliocattleya Golden Zell 0 0 20 0 

6. Brassolaeliocattleya Green Wich 13 0 7 65 

7. Brassolaeliocattleya Haad Yai 0 0 20 0 

8. Brassolaeliocattleya Hawaiian Passion 0 0 20 0 

9. Brassolaeliocattleya Lucky Strike ‘Mongkorn’ 0 0 20 0 

10. Brassolaeliocattleya Mem Tiang 0 0 20 0 

11. Brassolaeliocattleya White Diamond 2 0 18 10 

12. Brassolaeliocattleya White Diamond 0 0 20 0 

x Cattleya Gertrude Hausermann 

13. Cattleya Sea Breeze 10 0 10 50 

14. Cattleya hybrid 0 0 20 0 

15. Cattleya hybrid 0 0 20 0 

16. Cattleytonia Starrlyn 0 0 20 0 

17. Colmanara Wildcat ‘Bobcat’ 0 0 20 0 

18. Dendrobium Blushing White 18 0 2 90 

19. Dendrobium Burana Jade x D. Bertha Chong 0 0 20 0 

20. Dendrobium Chanel 20 0 0 100 

21. Dendrobium Chao Praya 20 0 0 100 

22. Dendrobium Earsakul 0 0 20 0 

23. Dendrobium Emma White 0 0 20 0 

24. Dendrobium Honey 0 0 20 0 

25. Dendrobium hybrid 15 0 5 75 



28. Dendrobium Juree Red 0 0 20 0 

29. Dendrobium Madame Vipa 0 0 20 0 

x Dendrobium Burana Green 

30. Dendrobium Pravit White 20 0 0 100 

31. Dendrobium Pigasus 0 0 20 0 

32. Dendrobium Sakura 20 0 0 100 

33. Dendrobium Shavin White 20 0 0 100 

34. Dendrobium Thongchai Gold 0 0 20 0 

35. Epicattleya Landwood 18 0 2 90 

36. Grammatocymbidium Lovely Melody 0 0 20 0 

37. Mokara Bota Gold 19 0 1 95 

38. Laeliocattleya Secret Love x Cattleya Queen Sirikit 0 0 20 0 

39. Oncidium Gower Ramsey 0 0 20 0 

40. Oncidium Sharry Baby 0 0 20 0 

41. Oncidium hybrid 18 0 2 90 

42. Phalaenopsis hybrid 0 0 20 0 

43. Vanda Motes Butterscotch x Vanda Doctor Anek 0 0 20 0 

44. Vanda Sanderiana x Vanda Tubtim 0 0 20 0 

Total 243 0 637 27.6

DISCUSSION 

CymMV and ORSV were not found in 

all of 1,000 in vitro Thai native orchid seedlings. 

Similarly, Porter et al. (1996) found that CymMV 

was not transmitted from CymMV-infected pod 

and pollen parents of 7,050 University of Hawaii 

(UH) Dendrobium hybrids seedlings and 

concluded that CymMV was not seed transmitted. 

Therefore, the use of seed-propagated cultivars and 

species shall be one of the most promising 

approaches to establish virus-free orchid plants and 

high quality germplasm. However genetic 

variation within offspring is unexpected. 

Cultivated orchid plantlets derived from 

the micropropagation of axilliary bud and the shoot 

tip explants were infected with CymMV but noninfected 

with ORSV. It was found that orchid tissue 

cultured plantlets were infected with CymMV 

which similarly to Wong et al. (1994) reported that 

50.5% of the thirteen orchid genera derived from 

tissue culture at the tissue culture laboratory of 

the Botanical Garden in Singapore were infected 


Figure 2 Cassia occidentalis leaf infected with CymMV showing necrotic lesions (arrow), 

5 days of inoculation (A) and the normal control leaf of Cassia occidentalis (B). 

with CymMV. In Taiwan, some in vitro plantlets 

derived from meristem of commercial cut flower 

cultivars were infected with CymMV (Chia et al., 

1991). The widespread of CymMV in the orchid 

tissue culture plantlets resulted from the use of 

infected mother plants for mass clonal 

propagation. This might cause serious damage to 

Thai orchid industry in subsequent years. 

Therefore, it is critical to screen all the plants with 

a very sensitive technique such as reverse 

transcription-polymerase chain reaction (RT-PCR) 

or ELISA before the tissue is cultured. Otherwise, 

a large population of virus-infected plant will be 

produced. In this survey, some axenic cultivated 

orchid plantlets were not infected with CymMV. 

Virus-free mother stock orchids should be well 

maintained by separate planting. There are many 

available methods to produce virus-free orchids 

from virus-infected plant including culture of 0.1 

mm apical tissue pieces of infected orchid plant 

(Morel, 1960) and, chemotherapy (Yab et al., 

1999) or chemotherapy and thermotherapy of 

infected tissue before culturing (Kim et al., 1997).

56 

Virus-free orchid plants show faster and healthier 

growth and produce much larger and longer 

inflorescences length than infected orchid plants 

(Chia and He, 1999). 

The present study reveals that CymMV 

is prevalent virus in orchid. It has spreaded widely 

in many cultivated orchid genera in Thailand. 

Plants must test for viral contamination before 

cloning to prevent the viral spreading. After tissue 

proliferation and plant differentiation, another test 

for viral infection has to be conducted before 

releasing the material from flask to further multiply 

or to transfer to community pots in greenhouses. 

It is essential to produce disease-free plantlets for 

export, especially to countries that impose strict 

plant quarantine conditions. ORSV is not prevalent 

virus in cultivated Thai orchids in this test, but 

screening regimes should be included to determine 

its existence. 

The reliability of indirect-ELISA in 

detection of CymMV of in vitro orchid plantlets 

was similar to bioassay. Hu et al. (1994) detected 

CymMV in fifty orchid samples and found that 

the results of ELISA and bioassay for detection 

were similar. However bioassay bioassay is timeconsuming 

and requires greenhouse space and it 

takes many days or weeks to get a conclusive result 

(Chang, 2004). Consequencely, ELISA is more 

rapid method for detecting CymMV than the 

mechanical inoculation bioassay (Hu et al., 1994). 

CONCLUSION 

CymMV and ORSV were not found in 

micropropagated Thai native orchids seedlings. 

CymMV was detected at 27.6% of 880 axenic 

cultivated orchids samples. CymMV was detected 

in 6 genera namely Brassolaeliocattleya, Cattleya, 

Dendrobium, Epicattleya, Oncidium and Mokara. 

ORSV was not detected in all orchid samples. This 

investigation revealed that CymMV was found in 

orchid mericlones. CymMV is the most prevalent 

virus in cultivated orchids in Thailand. This study 


suggests that the plant material must be examined 

for the existence of the virus before using them 

for mass production by tissue culture techniques. 

Use of seed-propagated cultivars provides a most 

suitable mechanism to establish virus-free 

plantings of orchids and high quality germplasm. 

It is believed that the results from this study are 

essential for tissue culture laboratories to change 

their practices for producing high-quality virusfree 

plants in the very near future. The indirect- 

ELISA will be a powerful tool for diagnosing 

CymMV in cultivated Thai orchids by large-scale 

indexing program. 

ACKNOWLEDGMENTS 

This research was granted by Graduate 

School, Kasetsart University, Bangkok, Thailand. 

We would like to express our gratitude to 

Laboratory of Dr. Pranee Hamelink, Department 

of Plant Pathology, Kasetsart University for 

providing CymMV and ORSV antisera. We also 

thank to all tissue culture laboratories for orchid 

samples. Dr. Srimek Chowpongpang, Department 

of Plant Pathology, Kasetsart University for the 

manuscript editing and helpful suggestion. A 

special thank goes to Dr. T. Hanselmann 

(University of Melbourne) for critically reading 

the manuscript. 


Chang, C.A., H.H. Lee, C.C Chen, M.J. Lin and 

C.P. Wang. 2003. Phytosanitary certification 

program of oncidium seedlings and its future 

the development of ornamental industry in 

Taiwan. Plant Pathol. Bulletin 12(3): 141- 

148 

Chang, C.A. 2004. Orchid virus disease in Taiwan 

and their control strategies, pp. 343-356. In 

Editorial Committee of Taiwan Orchid 

Growers Association (ed.). Proceedings of 

8 th Asia Pacific Orchid Conference. Liang

Hung Color printing Co., Ltd., Taiwan. 

Chia, T.F., Y.S. Chan and N.H. Chua. 1991. 

Survey of cymbidium mosaic and 

odontoglossum ringspot viruses in cultivated 

orchids by nucleic acid hybridization, pp. 56- 

59. In Proceedings of The Nagoya 

International Orchid Show 91. Naganae 

Printing Co., Ltd. Aichi-ken. 

Chia, T.F. and J. He. 1999. Photosynthetic capacity 

in Oncidium (Orchidaceae) plants after virus 

eradication. Environ. Exp. Bot. 42: 11-16. 

Clark, M.F. and A.N. Adams. 1977. Characteristics 

of the microplate method of enzymelinked 

immunosorbent assay for the detection 

of plant viruses. J.Gen.Virol. 34: 475-483. 

Hu, J.S, S. Ferreira, M.Q. Xu, M. Lu, M. Iha, E. 

Pflum and M. Wang. 1994. Transmission, 

movement, and inactivation of cymbidium 

mosaic and odontoglossum ringspot viruses. 

Plant Dis. 78(6): 633-636. 

Kim, M.S., Y.R. Lee, C.H. Song, B.H. Kim, S.B. 

Kyun and J.S. Eun. 1997. Elimination of 

odontoglossum ringspot virus (ORSV) from 

orchid (Cymbidium spp.) by chemotherapy 

and thermotherapy. RDA. J. Crop Protec. 

39(1): 19-24. 

Morel, 1960. Producing virus-free cymbidiums. 

Amer. Orchid Soc. Bull. 29: 473-478. 

Nanakorn, W. and S. Indhamusika. 2000. Queen 

Sirikit Botanic Garden (Vol.6). O.S. 

Printing House, Bangkok. 

Office of Agricultural Economics. 2005. Quantity 

and Value of Exporting Agricultural Product 

2004. Available source: http://www.oae.go.th/ 

Import-Export.php, 15 January 2005. 

Pearson, M.N. and J.S. Cole. 1986. The effect of 

cymbidium mosaic virus and odontoglossum 


ringspot virus on the growth of Cymbidium 

orchids. J. of Phytopathology 117: 193-197. 

Porter, K.G., A.R. Kuehnle and J.S. Hu. 1996. 

Lack of seed transmission of cymbidium 

mosaic virus in dendrobium. Lindleyana 

11(4): 211-213. 

Satula, C.L., J.M. Gillett, S.M. Morrissey and D.C. 

Ramsdell. 1986. Interpreting ELISA data and 

establishing the positive-negative threshold. 

Plant Dis. 70: 722-726. 

Vejaratpimol, R., C. Channuntapirat, T. Pewnim, 

K. Ito, M. Iizuka and N. Minamiura. 1998. 

Evaluation of enzyme-linked immunosorbent 

assays for the detection of cymbidium mosaic 

virus in orchids. J. Ferment. Bioeng. 86: 

65-71. 

Wannakrairoj, S. 2004. Orchid business of 

Thailand, pp.109-113. In Editorial Committee 

of Taiwan Orchid Growers Association (ed). 

Proceedings of 8 th Asia Pacific Orchid 

Conference. Liang Hung Color Printing Co., 

Ltd., Taiwan. 

Wong, S.M. and S.H. Lim. 1994. Detection of 

odontoglossum ringspot virus with and 

amplified enzyme-linked immunosorbent 

assay (ELISA) system employing a secondary 

cyclic enzyme reaction. Acta Hort. 377: 183- 

188. 

Wong, S.M., C.G. Chng, Y.H. Lee, K. Tan and F.W, 

Zettler. 1994. Incidence of cymbidium 

mosaic and odontoglossum ringspot viruses 

and their significance in orchid cultivation in 

Singapore. Crop. Prot. 13(3): 235-239. 

Yap, M.L., L.T. Lan-Chan, C.P. Yik, J. Kueh, S.M. 

Lee and H.M. Loh. 1999. Eradiation of 

viruses from commercial orchids. Singapore 

J. of Pri. Ind. 27: 11-19


Comparative Performances of Holstein-Friesian Cows under 

Smallholder and Large Scale Farmers’ Management in 

Central Rift Valley, Ethiopia 

Nega Tolla1 *, Pravee Vijchulata2 , Pornsri Chairatanayuth2 and Suwapong Swsdiphanich3 ABSTRACT 

A study was conducted in Central Rift Valley of Ethiopia to assess comparative productive and 

reproductive performances of graded Holstein Friesian cows under small and large scale farmers’ 

management. A total of 28 animals from both small (14) and large scales farms (14) in different parity 

classes were used for data collection for a period of 28 weeks. Significant differences were observed in 

daily intakes of dry matter, crude protein and P (p

eported that lactation milk yield of Holstein- 

Friesian cattle on large-scale commercial state 

dairy farms could not exceed 3000kg. Under some 

other tropical feeding and management conditions 

? or pure Holstein Friesians could produce over 

4,000 kg milk per lactation (Combellas, et al., 

1981; McDowell, 1985). Leng (1999) reported that 

Friesian cattle imported into India had produced 

an average of 6000kg milk in 300-days of lactation 

with appropriate nutritional management practices. 

In Ethiopia, market oriented small and 

large scale farms in the peri-urban and urban 

centers are the two categories of milk production 

systems (EARO, 1998). These sectors rely on 

crossbred and exotic cattle breeds under intensive 

and semi-intensive management with production 

goal of cash income. They are dependent on 

purchased roughage and concentrate feeds with 

limited grazing for milk production. Lack of 

appropriate nutritional management, health care, 

marketing system, extension and training are the 

major constrains for their low productivity (Goshu 

and Mekonen, 1997; Azage and Alemu, 1998). 

However, information on comparative production 

and reproductive performances of this genotype 

under large scale and small scale farmers’ 

management in Ethiopia is limited. Therefore, the 

main objective of this study was to compare the 

performance of graded Holstein Friesian dairy 

cows under large scale and smallholder dairy farms 

in the urban and peri-urban centers of Central Rift 

Valley of Oromia region, Ethiopia. 


This study was conducted in the Central 

Rift Valley (CRV) of Oromia, at Arsi Negelle, 

Ziway, Wonji Kuriftu and Lume districts of East 

Shoa Zone. The area is characterized with different 

altitude ranges of 1550 and 1900 meters above 

sea level and average minimum and maximum 

temperature of 20°C and 27°C respectively. It has 

an erratic and unreliable rainfall, ranging from 500 


to 900 mm per year. 

A rapid exploratory survey was 

undertaken to identify and locate the existing large 

and small scale dairy farms in the urban and periurban 

centers of the study area. Based on the 

willingness of the farm owners, the presence of 

dairy cows of graded Holstein-Friesian genotype, 

known parity and stage of pregnancy, 3 large periurban 

farms having 170 to 195 heads of dairy 

animals and 21 small scale urban farmers in the 

secondary towns having 1 to 10 heads of cattle 

were identified. The farms were categorized based 

on the existing herd size as small scale (≤10 

animals) and large scale (>10 animals). 

Accordingly, two large scale farms and 12 

smallholder farmers were randomly selected. A 

total of 28 animals from both large scale 

(14 animals with 426±85 kg average body weight) 

and small scale (14 animals with 363±8.5 kg 

average body weight) with the parity ranging from 

1 to 6 and in the last trimester of pregnancy were 

used for data collection. Animals in different 

parities were classified as early parity (1-2 

lactation) and advanced parity (3-6 lactation). 

Seven animals were used in each parity class. In 

the urban small scale production system the 

animals were entirely confined at home utilizing 

whatever space was available in the residential 

compounds. There were no sufficient exercising 

areas. Animal houses were normally small 

structures of corrugated metal sheeting or mud and 

floored with blocks of stone or cemented floor. 

Purchased cereal crop residues, such as maize 

stover, haricot bean straw, tef (Eragrostis 

abyssinica) straw were the major basal feed 

resources. Green forages of maize and native grass 

were also used seasonally. The utilization of agroindustrial 

by-products such as nougseed cake 

(Guizota abyssinica), linseed cake, cotton seed 

cake and wheat bran was minimal. In the periurban 

large scale production units animals were 

housed in sheds with well ventilated corrugated 

metal sheet and cemented floors. Their major basal

60 

feed source was based on purchase or very few 

harvested native grass hay. Green feeds of alfalfa 

and elephant grass were also used. There was 

regular supply of mixed concentrate diet. 

Data collection was started from about 

one week postpartum. The utilization of available 

feed resources and daily milk yield of each selected 

farm was monitored and recorded every five days 

for 210 days. Daily milk yield of individual animal 

was monitored and recorded for both AM and PM 

using portable spring balance. The amount and 

type of feed offered to individual animal was also 

weighed and recorded for each monitoring date. 

Both daily feed intake and milk yield for none 

collection days were estimated from average 

values of the preceding measurements. 

Accordingly the refusal of any feed type offered 

was weighed and recorded. The amount of daily 

nutrient intake over a given period was estimated 

by multiplying the nutrient contents of the feeds 

(per kg dry matter) by the daily dry matter intake 

of the respective feed. Dry feed samples from each 

farm were collected fortnightly and bulked. They 

were thoroughly mixed, sub sampled and delivered 

to laboratory for chemical analysis. Milk sample 

was collected three times, at 4, 16 and 27 weeks 

of lactation period using 100ml sampling bottle 

preserved with potassium dichromate and 

delivered the same day to the laboratory. On the 

large scale farms, any sign of estrus manifestation 

was visually observed and recorded by barn 

attendants and the veterinarian daily in the morning 

and after-noon. On the small scale farms the 

enumerators assigned for data recording visually 

observed and recorded any sign of estrus. In both 

cases mating practice was by artificial 

insemination (AI). However, on the small scale 

farms several skipped mating were observed due 

to shortage of AI facilities and/or unavailability 

of AI technicians. 

Chemical analysis 

Feeds were analyzed for dry matter 


(DM), organic matter (OM) and nitrogen (N) using 

standard procedures of AOAC (1990). Neutral 

detergent fiber (NDF) was determined as described 

by Van Soest and Robertson (1985). The in vitro 

organic dry matter digestibility (IVDOMD) was 

determined using the procedures described by 

Tilley and Terry (1963). Metabolizable energy 

(ME) content (MJ/kg DM) of feed was estimated 

from in vitro digestibility (IVDOMD × 0.016) as 

suggested by Barber et al. (1984). Metabolizable 

energy intake (MEI) was estimated by multiplying 

dry matter intake (DMI) of the feed with the values 

of their respective energy concentration, i.e.; MEI 

(MJ/d) = DMI (kg/d) × ME (MJ/kg DM) according 

to Kearl (1982) and MAFF (1985). Calcium and 

sodium contents of the feeds were analyzed using 

atomic absorption spectrophotometers according 

to Perkins (1982), and phosphorus content was 

determined according to AOAC (1990). The daily 

crude protein (CP) and ME requirement for the 

animals was estimated based on actual average 

daily milk yield and fat content according to NRC 

(1989) recommendation. The composition of milk 

fat and protein were analyzed using Gerber method 

and formaldehyde titration respectively and total 

solid was determined by oven drying the milk 

sample according to O’Mahoney (1988). 


Data on daily milk yield, milk 

compositions, postpartum reproductive 

efficiencies and nutrient intake were analyzed for 

farm scale, parity class and lactation period 

differences using the General Linear Model and 

multivariate analysis procedure of SPSS (1997). 

Mean differences between subjects under study 

were tested by pairwise comparison and least 

significant difference (LSD) method. The model 

used to analyze the effects of farm scale and parity 

classes on milk yield, reproductive traits and 

nutrient intake was: 

Y ij = µ + S i + P j + SP (ij) + e (ijk) 

Where, Y ijk = the means of daily milk yield or

µ 

reproductive traits or nutrient 

intake measured. 

= the overall mean 

Si = effect of ith farm scale (i = 1, 2) 

Pj = effect of jth Parity class 

(j = 1, 2) 

SP (ij) = interaction between farm scales 

and parity classes 

e (ijk) = error term 

The Model used to analyze effects of 

farm scale, parity class and period on milk 

composition was: 

Yijk = µ + Si + Pj + Tk + e (ijk) 

Where, Yijk = the means for milk composition 

µ = the overall mean 

Si = effect of ith farm scale (i = 1, 2) 

Pj = effect of jth Parity class (j = 1, 2) 

Tk = effect of kth period (k = 1, 2, 3) 

e (ijk) = error term (assumed 


independently and normally 

distributed) 


Chemical composition of feed resources 

Mean chemical compositions (% DM) of 

the available feed resources utilized on different 

farms are presented in Table 1. Some of the small 

scale farmers used cereal crop residues after 

soaking them with home-made brewers’ grain byproducts. 

It was observed that soaking tef and 

wheat straws with brewer’s grain by product 

substantially increased the CP, ME, Ca 

compositions and IVDMOD of the feeds. 

Feeds and nutrient intake 

Significant differences were observed 

between the large and small scale farms in daily 

Table 1 Mean chemical compositions of feed resources available for graded Holstein Friesian cows in 

central Rift Valley, Ethiopia. 

Feed types (% DM) 

DM CP ME(MJ) NDF EE ASH IVOMD Ca P Na 

Native grass hay 92 4.4 5.7 76.9 1.2 8.0 35.9 0.3 0.15 0.54 

Native grass forage 13 9.1 7.2 75.9 1.4 9.3 45.0 0.43 0.3 0.01 

Alfalfa forage 35 17.4 8.6 50.9 1.8 11.8 53.9 1.0 0.3 0.22 

Elephant grass 15 13 8.1 50.9 1.9 17.1 50.3 0.2 0.3 0.1 

Maize forage 35 10.4 10.0 62.3 1.3 9.2 62.4 0.24 0.33 0.2 

Maize stover 91 5.6 8.85 81.2 0.7 8.2 55.3 0.35 0.10 1.7 

Wheat straw* 62 13.1 10.4 59.0 8.4 6.5 65.1 0.4 0.4 0.22 

Wheat straw 92 2.7 5.81 81.6 0.9 8.2 36.3 0.09 0.05 0.001 

Tef straw* 73 12.7 9.7 59.9 5.1 6.1 60.9 0.33 0.4 0.03 

Tef straw 92 4.3 6.2 83.1 1.1 6.6 38.6 0.2 0.09 0.08 

Haricot bean straw 91 5.2 6.8 71.0 0.7 8.1 42.5 0.18 0.05 0.26 

Molasses 74 3.5 15.9 - 5.9 5.9 98.7 1.8 0.10 0.26 

Brewers grain residue 13 21.4 11.0 57.4 3.5 4.0 69.0 0.61 0.59 0.004 

Noug seed cake 91 29.7 9.3 36.4 6.2 10.9 58.1 0.69 0.99 0.03 

Linseed cake 90 29.1 10.6 24.3 9.1 8.1 66.3 0.51 0.92 0.22 

Cotton seed cake 90 22.3 7.6 48.7 6.6 5.6 47.5 0.19 0.74 0.08 

Wheat bran 88 17.1 11.6 38.4 4.6 4.4 72.4 0.11 1.0 0.05 

Mixed concentrate 89 21.3 10.6 37.4 4.7 7.1 66.3 0.34 1.2 0.25 

* Soaked in home-made brewers grain residues (Atala)

62 

intake of DM, CP, ether extract (EE), P (p

y animals on the small scale farms was roughage 

feed as compared to 38.6% of daily DMI on large 

scale farms. Similarly about 71% of NDF intake 

of animals on the small scale farms was roughage 

feeds vs. 53% of DMI on the large scale farms. 

The utilization of agro-industrial by-products such 

as nougseed cake (Guizota abyssinica), linseed 

cake, cotton seed cake and wheat bran was 

minimal (7.76% DMI). A home-made brewer’s byproducts 

(39% DMI) mixed with crop residues or 

alone was also used. Mixed concentrate in the daily 

dietary DMI of animals in small scale farms was 

only 9.6%. Since the potential intake of forage was 

inversely related to its NDF content, the DM intake 

and consequently that of other nutrients were 

limited for animals in small scale farms. Feeds and 

nutrient intake were not significantly (p>0.05) 

different among parity classes (Table 4). 

Milk yield, milk composition and reproductive 

efficiency 

The mean daily milk yield and 

composition, and postpartum reproductive 

efficiencies of Holstein Friesian cows on large and 

small scale farms are presented in Table 3. 

Difference in actual and fat corrected daily milk 

yield was highly significant (P

64 

relatively slow and short. The peak yield was 

attained at week 4 and thereafter declined 

continuously as well. Although the differences 

were not significant (P>0.05) animals in advanced 

parities (3-6) on both farm scales performed better 

than those in early (1-2) parities (Figure 2). On 

large scale farms animals in early parity (1-2) 

attained peak yield at week 4 and the trend of 

decline was slower than for those in advanced 


parities (3-6). Animals in advanced parities 


decline was relatively faster than that of their 

counterparts. On the small scale farms and animals 

in both parities, trend of milk yield increased was 

slower and shorter. Animals in advanced parities 


decline was also slower relative to those in early 

parities, which attained very short peak yield at 

Table 3 Mean daily milk yield, milk composition and reproductive efficiency of Holstein Friesian 

cows under two farm scales. 

Daily nutrients intake Farm scales 

Small Large SE P 

Number of animals 14 14 

Milk yield (kg) 11.5 15.8 0.73 *** 

FC milk yield (kg) a 11.1 14.7 0.67 *** 

Fat (%) 3.8 3.6 0.17 NS 

Protein (%) 2.9 2.8 0.08 NS 

Total solid (%) 11.3 11.3 0.24 NS 

Calving to first sign of estrus (d) 96 115 14.60 NS 

Days open 171 148 23.50 NS 

Services/conception 1.6 2.1 0.24 NS 

***=p

Milk yield (kg/d) 

20 

18 

16 

14 

12 

10 

8 

0 

2 

4 

6 

8 


10 

12 

14 

Figure 2 Mean daily milk yield of Holstein Friesian cows in two different farm scales and parity 

classes of 1-2 and 3-6. 

16 

18 

20 

Weeks of lactation 

Small scale Large scale 

Figure 1 Mean daily milk yield of cows in two different farm scales during 28 weeks of lactation. 

Milk yield (kg/d) 

20 

18 

16 

14 

12 

10 

8 

0 

2 

4 

6 

8 

10 

12 

14 

16 

18 

Weeks of lactation 

20 

22 

22 

24 

Small scale (1-2) Small scale (3-6) 

Large scale (1-2) Large scale (3-6) 

24 

26 

26 

28 

28

66 

week 4. Milk composition was not significantly 

(P>0.05) different among the two farm scales and 

parities, (Tables 3 and 4). There was significant 

effect of period on the content of milk protein 

(P0.05) different between the farm scales (Table 

3) and parity classes (Table 4). But the value of 

DO on small scale farms was longer (171) than 

on the large scale farms (148). This might be due 

to lack of regular AI services and several skipped 

mating observed. The days to first estrus (115) and 

S/C (2.1) were higher on the large scale farms than 

on the small scale farms (96 and 1.6 respectively), 

which might be due to poor heat detection 

practices. Since the numbers of animals on small 


Table 5 Effect of lactation period on milk composition of Holstein Friesian cows in two farm scales. 

Farm Scales Compositions Lactation periods 

Week 4 Week 16 Week 27 SE P 

Small No of animals 28 28 28 

Fat (%) 3.8 3.7 3.9 0.19 NS 

Protein (%) 2.6 2.4 3.5 0.13 *** 

Total solid (%) 10.7 11.5 11.8 0.28 ** 

Large Fat (%) 3.6 3.4 3.7 0.19 NS 

Protein (%) 2.5 2.4 3.5 0.13 *** 

Total solid (%) 10.6 11.5 11.8 0.27 ** 

Fat (%) 3.7 3.6 3.8 0.17 NS 

Overall Protein (%) 2.5 2.4 3.5 0.10 *** 

Total solid (%) 10.7 11.5 11.8 0.23 ** 

***=p

showed the opportunities for further improvement. 

There will be a need to investigate the nutrient 

balance of available feed supply for both farm 

scales. The ratio of rumen degradable to that of 

un-degradable protein must be assessed and 

adjusted to the recommended level. The nutritional 

values of the home-made liquor residues used by 

most small scale farmers have to be further 

assessed. The need of technical and institutional 

support will be indispensable to ensure sustainable 

supply of nutritionally balance feeds, AI and 

veterinary services which are the major areas of 

research and development intervention. This may 

need establishment of dairy farmers cooperatives 

in the study area, where other farmers elsewhere 

were successful in getting access to regular animal 

feed supply and market out-let for their products. 


The authors would like to appreciate 

Oromia Agricultural Research Institute (OARI) 

and the Ethiopian Agricultural Research 

Organization (EARO) for financially supporting 

this study. 


Abaye T., T.G. Meskel, A.G. Wold, B. Yemane 

and C. Philip. 1991. Status of dairying in 

Ethiopia and strategies for future 

development, pp. 25–36. In Proceedings of 

3 rd National Livestock Improvement 

Conference, 24–26 May 1989, Institue of 

Agricultural Research (IAR), Addis Ababa. 

AOAC (Association of Official Agricultural 

Chemists). 1990. Official Methods of 

Analysis. Washington, DC. 

Azage, T. and A.G. Wold. 1998. Prospects for periurban 

dairy development in Ethiopia, pp. 28- 

39. In Proceedings of 6 th National 

Conference of Ethiopian Society of Animal 

Production (ESAP), 15–17 May 1997, Addis 


Ababa. 

Barber,W.P., A..H. Adamson and J.F.B.Altma. 

1984. New methods of forage evaluation, 

pp.161-176. In W. Hareisgn and D.J.B. Cole 

(eds.). Recent Advances in Animal 

Nutrition. Butterworth, London. 

Chenost, M. and R. Sansoucy. 1991. Nutritional 

characteristics of tropical feed resources, pp. 

66–81. In A. Speedy and R. Sansoucy (eds.). 

Feeding Diry Cows in the Tropics. 

Proceedings of the FAO expert consultation 

held in Bangkok, Thailand, 7–11 July 1989. 

Food and Agricultural Organization of the 

United Nations, Rome. 

Chesworth J. and H. Guérin. 1992. Ruminant 

Nutrition. Macmillan Education Ltd. London. 

170 p. 

Christensen, R.A., G.L.Lynch and J.H. Clark. 

1993. Influence of amount and degradability 

of protein on production of milk and milk 

components by lactating Holstein cows. J. 

Dairy Sci. 76: 3490–3496. 

Combellas, J., N. Martinez and M. Compariles. 

1981. Holstein cattle in tropical areas of 

Venzuella. Trop. Anim. Prod. 6 (3): 214–220. 

DePeters, E.J. and J.P. Cant. 1992. Nutritional 

factors influencing the nitrogen composition 

of bovine milk: a review. J. Dairy Sci. 75: 

2043–2070. 

EARO (Ethiopian Agricultutral research 

Organization). 1998. Anim. Sci. Research 

Strategy. Short Term Plan, EARO, Addis 

Ababa. 46 p. 

Goshu, M. and H.M. Mekonen. 1997. Milk 

production of Fogera cattle and their croses 

with Friesian at Gonder, Northern Ethiopia. 

Ethiopian J. Agr. Sci. (EJAS) 16(2) 61–74. 

Gould, C.M. 1969. Unifying hypotheses – fertility 

status, palatability, ketosis, and protein intake. 

Vet. Rec. 85: 662. 

Kearl, L.C. 1982. Nutrient Requirement of 

Ruminants in Developing Countries. 

International Feedstuff Institute, Utah

68 

Agricultural Experiment Station. Utah State 

University, Logan. 381p. 

Leng, R.A. 1991. Feeding strategies for improving 

milk production of dairy animals managed by 

small farmers in the tropics, pp. 82 – 104. In 

A. Speedy and R. Sansoucy (eds.). Feeding 

Dairy Cows in the Tropics. Proceedings of 

the FAO expert consultation held in 

Bangkok, Thailand, 7–11 July 1989. Food 

and Agricultural Organization of the United 

Nations, Rome. 

Leng, R.A. 1999. Feeding strategies for 

improving milk production, pp. 403-416. In 

L. Flavey and C.Chantalakhana (eds.). 

Smallholder Dairying in the Tropics. ILRI 

(International Livestock Research Institute), 

Nairobi. 

McDowell, L.R. 1985. Meeting Constraints 

to Intensive Dairying in Tropical 

Areas. Cornell International Agriculture 

Memeograph No. 108, Ithaca, New York. 

41 p. 

McDowell, L.R., J.H. Conrad, G.L. Ellis and J.K. 

Loosli. 1983. Minerals for Grazing 

Ruminants in Tropical Regions. Center for 

Tropical Agriculture, University of Florida, 

Gainesville. 86 p. 

Ministry of Agriculture, Fisheries and Food. 1985. 

Energy Allowance and Feeding Systems for 

Ruminants. Reference Book No. 433, 

London. 73 p. 

Mukassa, M.E. 1989. A Review of Reproductive 

Performance of Bos indicus (Zebu) Cattle. 

International Livestock Center for Africa 

(ILCA), Monograph No. 6, Addis Ababa. 136 p. 


National Research Council. 1989. Nutrient 

Requirement of Dairy Cattle. 6th ed., 

National academy Press, Washington, D.C. 

157p. 

O’Mahony, F. 1988. Rural Dairy Technology– 

Experience in Ethiopia. Dairy technology 

unit, manual No.4. International Livestock 

Center for Africa, Addis Ababa. 64p. 

Perkins E. 1982. AAS Manual Model 2380. 

Norwalk, Connecticut. 

Preston, T.R. and R.A.Leng. 1987. Matching 

Ruminant Production Systems with 

Available Resources in the Tropics and Sub- 

Tropics, Penambul Books Ltd. Armldale, 

NSW. 245p. 

Sendros D. and T. Kumssa. 1998. Factors to be 

considered in formulation of livestock 

breeding policy, pp. 26-35. In Proceedings 

of the 5 th National Conference of Ethiopian 

Society of Animal Production (ESAP). 

Addis Ababa. 

Simon, S. 2005. Comparative performance of dairy 

cattle under large scale and smallholder dairy 

farms in Tanzania: a review. Available Source 

htt://www.husdyr.kvl.dk/htm/php/ Tanzania/ 

Index2.htm, August 1, 2005. 

SPSS. 1997. Statistical Procedure for Social 

Science. SPSS Inc, version 10. 

Tilley, J.A.M. and R.A. Terry. 1963. A two-stage 

technique for the in vitro digestion of forage 

crops. J. Brit. Grassland Soc. 18: 104–111. 

Van Soest, P.J. and J.B. Robertson. 1985. Analyses 

of forage and fibrous foods. A Laboratory 

Manual for Anim. Sci. 613: 98–110, Cornell 

University, Ithaca, New York.


Biochemical Properties of Nile Tilapia (Oreochromis niloticus) 

Hemoglobin 

Kriangkrai Thongsarn 1 *, Wanchai Worawattanamateekul 2 , 

Suriyan Tunkijjanukij 3 , Choosri Sribhen 4 and Apassara Choothesa 4 

ABSTRACT 

Hemoglobins of O. niloticus (averaged standard length of 18.07 ± 0.89 cm and weight of 

216.86 ± 27.13 g) were analysed of some biochemical properties. Average hematocrits were 25.80 ± 

3.71% and average hemoglobin concentrations were 7.05 ± 1.08 g/dl. It was found that there were 9 – 11 

types of hemoglobin by cellulose acetate electrophoresis. The anodic and cathodic hemoglobins were 

revealed by native polyacrylamide gel electrophoresis with approximate amounts of 45.72 and 54.28% 

of total hemoglobin, respectively. In addition, the cathodic hemoglobin was classified by cellulose acetate 

electrophoresis as cathodic I and cathodic II with approximate amounts of 47.62 and 6.66% of total 

hemoglobin, respectively. Mass spectrometry was performed to determine molecular weights of 

hemoglobin. It was found that the molecular weights of native hemoglobin was 63.0 kDa and the molecular 

weights of subunits were 15.3, 15.6 and 16.1 kDa. The pIs of anodic and cathodic hemoglobins ranged 

5.28 – 6.80 and 7.27 – 9.71 , respectively. 

Key words: Nile Tilapia, fish hemoglobin, Tilapia hemoglobin 

INTRODUCTION 

Sripanitan (1983) reported that the 

oxygen affinity of hemoglobin was dependent on 

temperature, pH, organic phosphates and pCO 2. 

The increase in pH or decrease in organic 

phosphates and temperature result in higher 

oxygen affinity and decreased P 50. In addition, 

hemoglobin greatly increases the carrying capacity 

of oxygen, CO 2 and H + as a result of reversible 

binding of the distal histidine ligand of the 

hemoglobin molecule and appropriate allosteric 

interactions between the binding sites (Perry and 

Tufts, 1998). 

Variation of water parameters in fish 

habitats such as O 2 availability, salinity, ionic 

compounds, pH and temperature and their different 

gas exchange organs such as gills, lungs, skin, 

swim–bladders, pharyngeal surface and gut 

surface cause functional heterogeneity of fish 

hemoglobin (Fago et al., 1995; Weber, 2000). As 

in mammals, fish hemoglobin is globular 

tetrameric proteins that carry oxygen for using in 

energy production process of the tissues (Horton 

1 Center of Agricultural Biotechnology, Interdisciplinary Graduate Program, Kasetsart University, Kamphaengsaen 73140, 

Thailand. 

2 Department of Fishery Products, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 

3 Department of Marine Science, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 

4 Department of Physiology, Faculty of Veterinary Medicine, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding auther, e-mail: g4409103@hotmail.com 


70 

et al., 2002). In contrast to mammals which 

commonly have several “isoHbs” that may be 

polymorphic, they often show marked functional 

differentiation (Weber, 2000). 

Weber (2002) indicated that fish 

hemoglobins were separated as anodic and 

cathodic hemoglobins by electrophoresis. The 

anodic hemoglobin’s properties of Salmonids had 

low oxygen affinity, high Bohr and Root effects 

while the cathodic hemoglobin had high oxygen 

affinity, no or reverse Bohr effect. They believed 

that, the cathodic hemoglobin might have function 

as O 2 reserve and O 2 carrier under hypoxia. 

Autoxidation is the spontaneous 

oxidation of hemoglobin (Fe 2+ ) to methemoglobin 

(Fe 3+ ) by molecular oxygen. Methemoglobin 

cannot bind oxygen and functional inert with 

respect to oxygen transport (Perry and Tufts, 

1998). It is believe that fish hemoglobin 

autoxidation is related to fish storage time. 

Therefore, this work was designed to study some 

biochemical properties of hemoglobin of nile 

tilapia (O. niloticus) as the basic information to 

study fish hemoglobin autoxidation for 

development as fish freshness indicator. 


Fish samples 

Thirty–five live O. niloticus (averaged 

standard length of 18.07 ± 0.89 cm and weight of 

216.86 ± 27.13 g) were obtained from 

Kamphaengsaen Fisheries Institute, Faculty of 

Fisheries, Kasetsart University. 

Blood and hemoglobin preparation 

Thirty–five blood was collected by 

cardiac puncture with heparinized syringes 

and kept on ice. Each sample was separated 

to two parts. The first part was used to 

investigate the hematocrit by microhematocrit 

centifuge method (Biomed Group Co., Ltd. 

Thailand) and hemoglobin concentration by 

cyanomet-hemoglobin method (Bio–Medical 


Laboratory, Thailand). The second part was 

prepared for hemolysate as described by 

Tamburrini et al. (2001) and analysed total protein 

concentration by Biuret method. These 

hemolysates were kept at –20°C for further 

investigation of hemoglobin phenotype. 

Ten hemolysates were purified by gel 

filtration chromatography using Sephadex G–100 

(fine grade) column and eluted by 0.05 M Tris– 

HCl buffer (containing 0.02% sodium azide), pH 

7.4 with flow rate of 1.8 ml/min at 8°C. Five 

hemolysates were classified of hemoglobin groups 

by anion–exchange chromatography of DEAE– 

cellulose column as described by Tamburrini et 

al. (2001). The hemoglobin fractions from gel 

filtration column chromatography and anion– 

exchange column chromatography were desalted 

by ultrafiltration with centricon–30 (Amicon, 

U.S.A.), centrifuged at 3,000 rpm at 4°C for 30 

min and kept at –20°C to determine molecular 

weight and pI of hemoglobin. 

Hemoglobin phenotype 

Hemoglobin phenotype were identified 

from thirty–five hemolysates by cellulose acetate 

electrophoresis (Helena Laboratory, U.S.A.), using 

TITAN III–H with Tris–EDTA–boric acid buffer, 

pH 8.2 – 8.6 and AFSA 2 Hemo control as 

hemoglobin standard. The result was scanned with 

GS–670 densitometer and analysed by Molecular 

Analysis Program version 1.4 (Bio–Rad 

Laboratory, U.S.A.). Native polyacrylamide gel 

electrophoresis was used for studying hemoglobin 

phenotype of fourteen hemolysates. The running 

condition was at 200 volts, 120 mA for 40 min 

with electrode buffer, pH 8.3 and the gel was 

stained with Coomassie blue R–250. 

Hemoglobin molecular weight 

Molecular weights of hemoglobin were 

determined by mass spectrometry with Bruker 

Reflex IV MALDI–TOF mass spectrometer 

(Bruker, U.S.A.) using sinapinic acid as matrix and 

hitting with nitrogen laser. The samples were

fractioned from gel filtration chromatography of 

Sephadex G–100 column. 

Hemoglobin pI 

Isoelectric focusing (IEF) (Bio–Rad 

Laboratory, U.S.A.) was used to determine the pI 

(isoelectric point) of hemoglobin. The IEF gel (pH 

ranges of 3 – 10) and broad range IEF standards 

consisted of a mixture of native proteins with pIs 

ranges of 4.45 – 9.69 were used. Electrode buffers 

contained 20 mM lysine and 20 mM arginine for 

cathode and 7% phosphoric acid for anode. The 

gels were scanned by Genius Bio Imagine System 

Gel Documentation from Lab Focus (U.S.A.) and 

were investigated with Gene Tool Analysis 

Program version 3. 03. 03. (Syngene, U.S.A.). 


Identification of hematocrit and hemoglobin of 

Oreochromis niloticus 

Thirty–five O. niloticus with average 

standard length of 18.07 ± 0.89 cm and average 


weight of 216.86 ± 27.13 g were identified for 

hematocrits and hemoglobin concentrations. It was 

found that average hematocrits were 25.80 ± 

3.71% (ranging from 15.50 – 34.00%) and average 

hemoglobin concentrations were 7.05 ± 1.08 g/dl 

(ranging from 4.46 – 9.49 g/dl). These results were 

similar to those of Sun et al. (1995) showing the 

hematocrits and hemoglobin concentrations of O. 

niloticus at 28.40 ± 0.90% and 7.02 ± 0.20 g/dl, 

respectively. 

Phenotype of Oreochromis niloticus hemoglobin 

The hemoglobin phenotypes of O. 

niloticus are shown in Figure 1. It was found that 

there were 9 – 11 types of hemoglobin which 

moved more slowly to anode (+) than human 

hemoglobin standard (A and F). These results 

might be affected by the function of fish 

hemoglobin that was different from mammal. 

Weber (2000) reported that fish 

hemoglobins had multiplicity and similarity was 

called IsoHbs such as hemoglobins of Salmonids 

group had 8 – 19 isomers that differed at genetic 

Figure 1 Phenotypes of hemoglobins by cellulose acetate electrophoresis, pH 8.2 – 8.6 (350 volts, 25 

mins). Lane 1 was AFSA 2 Hemo Control of Helena Laboratory, U.S.A. and lane 2 – 7 were 

hemoglobins of O. niloticus from hemolysates.

72 

level and showed marked polymorphism. 

The hemoglobins of Elasmobranch had several 

isomers (12 – 13 isomers) called IsoHbs and 

they were different (Weber, 2000). The seven 

polymorphisms of hemoglobins of Atlantic 

Croaker (Micropogon undulatus) from 

Chesapeake Bay in Virginia were found by native 

polyacrylamide gel electrophoresis (Shelly and 

Mangum, 1997). 

There were two hemoglobin phenotypes 

of O. niloticus as presented in Figure 2. The first 

one called anodic hemoglobin with approximate 

amount of 45.72% of total hemoglobin and the 

second one called cathodic hemoglobin with 

approximate amount of 54.28% of total 

hemoglobin. In addition, the cathodic hemoglobin 

was classified as two groups by cellulose acetate 

electrophoresis, pH 8.2 – 8.6. The first group called 

cathodic I with approximate amount of 47.62% of 

total hemoglobin and the second group called 

cathodic II with approximated amount of 6.66% 

of total hemoglobin as presented in Figure 1. 

Molecular weight of Oreochromis niloticus 

hemoglobin 

MALDI–TOF mass spectrometer was 


used to determine the molecular weights of O. 

niloticus hemoglobin. The molecular weight of 

native hemoglobin was 63.0 kDa and molecular 

weights of subunits were 15.3, 15.6 and 16.1 kDa. 

Falk et al. (1998) indicated that hemoglobins of 

five tilapia species of genera Oreochromis and 

Serotherodon were composed of 2 major alpha 

chains and 4 (O. andersonii, S. gallaeus and S. 

melanotheron) or 5 (O. aereus and O. niloticus) 

beta chains. Molecular weights of hemoglobin 

subunits were estimated ranging from 16.3 – 17.6 

kDa (kilodaltons) and the molecular weight of 

native hemoglobin with approximated value of 

65.0 – 70.0 kDa. They believed that these 

hemoglobins and globin chains might be applied 

to fish taxonomy and population studies. 

pI of Oreochromis niloticus hemoglobin 

The IEF, pH 3–10 were used to determine 

pIs of O. niloticus hemoglobins. The pIs of anodic 

and cathodic hemoglobins ranged from 5.28 – 6.80 

and 7.27–9.71, respectively. Falk et al. (1998) 

reported that pIs of hemoglobins of 5 tilapia 

species of genera Oreochromis and Serotherodon 

were found range from 5.94–8.06 and they 

differed by their globin chains. 

Figure 2 Phenotypes of hemoglobins by native polyacrylamide gel electrophoresis, pH 8.3, 12%T 

and 4%C (200 volts constant, 120 mA, 40 min). Lane 1 was AFSA 2 Hemo Control of Helena 

laboratory (U.S.A.) and lane 2 – 7 were hemoglobins of O. niloticus from hemolysates.

CONCLUSION 

Averaged hematocrits were 25.80 ± 

3.71% (ranging from 21.00 – 34.00%) and average 

hemoglobin concentrations were 7.02 ± 0.2 g/dl. 

There were 9 – 11 types of hemoglobin by 

cellulose acetate electrophoresis. The anodic and 

cathodic hemoglobins were revealed by native 

polyacrylamide gel electrophoresis. The cathodic 

hemoglobin was classified by cellulose acetate 

electrophoresis as cathodic I and cathodic II. The 

molecular weight of native hemoglobin was 63.0 

kDa and the molecular weights of subunits were 

15,3, 15,6 and 16,1 kDa. The pIs of anodic and 

cathodic hemoglobins ranged from 5.28 – 6.80 and 

7.27–9.71, respectively by isoelectric focusing. 


This work was supported by Center of 

Agricultural Biotechnology, Kasetsart University, 

Nakornpratom and National Center of Genetic 

Engineering and Biotechnology, National Science 

and Technology Development Agency, 

Pathumthani, Thailand. We would like to thank 

Dr.Sittiruk Roytrakul at BIOTEC Central Research 

Unit for his assistance in determining the 

molecular weights of O. niloticus hemoglobins by 

mass spectrometry. 


Fago, A., V. Carratore, G. D. Prisco, R. J. Feuerlein, 

L. Sottrup-Jensen and R. E. Weber. 1995. The 

cathodic hemoglobin of Anguilla anguilla: 

amino acid sequence and oxygen equilibria 

of a reverse Bohr effect hemoglobin with high 

oxygen affinity and high phosphate sensitivity. 

J. Biol. Chem. 270 (32): 18897 – 18902. 

Falk, T. M., E. K. Abban, W. Villwork and L. 

Renwrantz. 1998. Hemoglobin variations in 

some tilapiine species (Teleostei, Cichildae) 


of the genera Oreochromis and Sarotherodon. 

Genetics and Aquaculture in Africa, Actes 

de Collogue Genetique et Aquaculture en 

Afrique, Abidjan, Cute d’ lvoire, Apr. 1-4: 

147 – 152. 

Horton, H. R., L. A. Moran, R. S. Ochs, J. D. Rawn 

and K. G. Scrimgeour. 2002. Principle of 

Biochemistry. 3 rd ed. Prentice–Hall, Inc., 

Upper Saddle River. 862 p. 

Murakami, K. and S. Mawatari. 2003. Oxidation 

of hemoglobin to methemoglobin in intact 

erythrocyte by a hydroperoxide induces 

formation of glutathionyl hemoglobin and 

binding of α–hemoglobin to membrane. Arch. 

Biochem. Biophys. 417: 244 – 250. 

Shelly, D. A. and C. P. Mangum. 1997. 

Hemoglobin polymorphism in the Atlantic 

croaker, Micropogon undulatus. Comp. 

Biochem. Physiol. 118A (4): 1419 – 1428. 

Sripanitan, R. 1983. Structural Characterization 

of the Four Component of Hemoglobin 

Constant Spring. M.S. Thesis, Mahidol 

University. Bangkok. 

Sun, L. G., G. Chen and C. Chang. 1995. Acute 

responses of blood parameters and comatose 

effects in salt–acclimated tilapias exposed to 

low temperatures. J. Therm. Biol. 20 (3): 299 

– 306. 

Tamburrini, M., C. Verde, A. Olianas, B. Giardina, 

M. Corda, M. T. Sanna, A. Fais, A. M. Deiana, 

G. D. Prisco and M. Pellegrini. 2001. The 

hemoglobin system of the brown Moray 

Gymnothorax unicolor: structure/ function 

relationships. Eur. J. Biochem. 268: 4104 – 

4111. 

Weber, R. E. 2000. Adaptations for oxygen 

transport : lessons from fish hemoglobin, pp. 

22 – 37. In G. D. Prisco, B. Giardina and R.E. 

Weber (eds.). Hemoglobin Function in 

Vertebrates: Molecular Adaptation in 

Extreme and Temperate Environments. 

Springer, Verlag Italia, Milano.


Screening of Ethiopian Traditional Medicinal Herbs for their 

Nitrification Inhibiting Ability 

Wassie Haile*, Thongchai Mala, Yongyuth Osotsapar and Visoot Verasan 

ABSTRACT 

Nitrification is the main cause of reduced N-use efficiency and yield of crops. Blending 

nitrogenous fertilizer with commercial or herb-based inhibitors is one of the means of increasing N-use 

efficiency. The objective of this study was to screen Ethiopian traditional medicinal herbs along with 

Neem (none traditional Ethiopian medicinal plant) and commercial inhibitors for their nitrification 

inhibiting ability. Soil samples were collected from Awassa Agricultural Research Center, Ethiopia. 

One hundred gram of processed soil was transferred to 250 ml capacity cups to which alcohol extracts 

of 11 herbs at 1% rate of dry soil were added separately. N-seve ([2-chloro-6 (trichloromethyl) pyridine]) 

and Dicyandiamide (DCD) were added at a rate of 2 and 100µg/g of soil respectively. The cups were 

also amended with 50 mg of ammonium sulphate and the moisture was maintained at 60% water holding 

capacity (WHC) and incubated for various period at 25°C. At the end of each period, samples were 

analyzed for pH, NH 4-N and NO 3-N. The result revealed that 72% of the tested herbs and commercial 

inhibitors conserved significantly high amount of NH 4-N, an indication of inhibition of nitrification, at 

the end of 2 nd week. But from among Ethiopian medicinal herbs, only Artemis afra, Echinops spp and 

Eugenia caryophyllata inhibited nitrification at the end of 3 rd week. Averaged over two incubation 

periods, these herbs inhibited nitrification by 33, 37 and 64% respectively. Eugenia caryophyllata 

performed as effective as neem but none of the herbs out performed commercial inhibitors. Both medicinal 

herbs and commercial inhibitors prevented the soil from the acidification of the soil. There was a decline 

in inhibition percentage of all inhibitors with time but it was more drastic for herbs than commercial 

inhibitors. It is concluded that 3 Ethiopian medicinal herbs were identified as nitrification inhibitors. 

Thus, their extracts could be blended with ammonium fertilizers for increasing N-use efficiency of crops. 

Key words: medicinal herbs, nitrification, percentage inhibition ability 

INTRODUCTION 

Nitrogen is by far the most important 

nutrients required by crops or plants for their 

growth and development. However, it is deficient 

in most Ethiopian soils which resulting in limiting 

crop production. Traditional methods of soil N 

restoration methods such as fallowing, crop 

rotation with legumes and incorporation of manure 

are abandoned due to shortage of land which in 

turn is caused by ever increasing human 

population. 

In Ethiopia, to overcome N deficiency 

and increase crop production, application of Ncontaining 

chemical fertilizers has been started 

since 1971. Its consumption increased from 947 

Department of Soil Science, Kasetsart University, Kamphaengsaen Campus, Nakhon Pathom 73140, Thailand. 

* Corresponding author, email: wassiehaile@yahoo.co.uk 


metric tons in 1971 to 446 000 tons in 2002 (Taye 

et al., 2002). Appreciable increases in the yields 

of several crops have been obtained due to this 

practice. However, the costs of chemical fertilizers 

are increasing from time to time becoming 

unaffordable to subsistent farmers. More over, the 

fertilizer use efficiency of crops is very low. For 

instance, in USA, maximum of 50% of the applied 

N is utilized by corn (Nelson and Huber, 2001). 

The situation is very severe in the tropics where 

only 25 to 40 % of the applied N is utilized by 

crops in a season (Sahrawat and Mukerjee, 1977). 

Nitrification is the main cause of reduced 

N use efficiency of crops, because it converts 

ammonium to nitrate, which is freely mobile and 

subjected to losses via leaching and denitrification 

(Jarvis et al., 1996). Thus, the control of 

nitrification in the soil is indispensable to increase 

N-use efficiency (Hauck, 1984). 

There are several methods of reducing 

N loss and improving N-use efficiency of crops. 

Such methods include proper time and methods 

of application of fertilizers, use of slow release 

fertilizers and use of nitrification inhibitors 

(Hauck, 1984). 

Nitrification inhibitors are chemical 

compounds that kill ammonia oxidizing bacteria 

or interfere with their metabolism. As a result N is 

maintained in the form of NH 4 ion which is less 

subject to losses (Hauck, 1984). To date, there are 

several compounds tested and produced 

commercially worldwide. Among them N-serve 

[2-chloro-6 (trichloromethyl) pyridine] and 

dicyandiamide (DCD) are by far the most widely 

used for nitrification inhibitors (Prasad and Power, 

1995). Details on various aspects of these 

inhibitors could be found in Meisinger et al. 

(1980). 

Although, there are variations among soil 

types and crops with respects to effectiveness of 

nitrification inhibitors, appreciable increase in the 

yields and N-use efficiencies of several crops were 

obtained by blending N-fertilizers with 


nitrification inhibitors (Prasad and Power, 1995). 

There are also nitrification inhibitors of 

plant origin. Paavolainen et al. (1998) reported that 

monoterpenes that are produced by roots of 

Norway spurce (Picea abies L.) inhibits 

nitrification. Neem (Azadirachta indica) cake was 

found to inhibit nitrification effectively both in the 

laboratory and greenhouse (Sahrawat and Parmer, 

1975). Nitrification inhibition property was also 

identified in Karanaja (Pongamia glabra) 

(Sahrawat and Mukerjee, 1977) and pyrithrum 

flower. The exploitation of such plants as inhibitors 

is important, as commercial inhibitors are 

expensive for developing countries like 

In this regard, there are several Ethiopian 

traditional medicinal herbs with anti-microbial 

property against human pathogenic bacteria 

(Mintesnot and Mogessie, 1999). But they have 

never been tested for their nitrification inhibiting 

ability. The objective of this study was therefore 

to investigate the nitrification inhibiting ability of 

some Ethiopian medicinal herbs in comparison 

with commercial inhibitors. 


Soil sampling, preparation and analysis 

A composite samples of surface top soil 

(0-15 cm) from 20 plots were collected using a 

spade. These samples were taken from a research 

farm of Awassa Agricultural Research Center, 

Ethiopia. The samples were air dried ,and 

grounded to pass 2 mm sieve and were analyzed 

for the following chemical chracteristics: pH (1:2 

soil water suspension (Mclean, 1982); organic 

matter (OM) by Walkley and Black’s (1934) 

method, available phosphorus (P) by Olsen and 

Sommers (1982), Exchangeable bases (BS) by 

Thomas (1982) method and micronutrients by 

Lindsay and Norvell (1978). Total nitrogen (TN) 

and cation exchange capacity (CEC) were 

determined using procedures described in Rowell 

(1994). The soil was characterized as Awassa clay

76 

loam (Eutric Fluvisol) soil and identified as fast 

nitrifying soil (Wassie et al., 2004). Selected initial 

characterstics are presented in Table 1. 

The samples were thoroughly mixed, airdried 

and milled to pass 4 mm sieve. One hundred 

gram of processed soil was transferred into 250ml 

capacity plastic cups. 

Preparation and extraction of medicinal herbs 

As described by Woldemichael (1987), 

the Ethiopian traditional medicinal herbs and 

Neem (None Ethiopian medicinal herb included 

as positive control) and their parts used in this 

study are listed in Table 2. These plant parts were 

dried in dark places and finely grounded to powder. 

A 100 gm of each medicinal herb was mixed with 

500 ml of 95 % Ethanol (1:5 ratio) and shaked on 

a shaker for 48 hours at 200 rpm. The extracts were 

then filtered through filter paper and the alcohol 

was removed using rotary evaporator. The extracts 

were further concentrated using steam hot plate 

maintained at 40 °C. 


Preliminary screening procedure 

Plastic cups containing 100 gm of 

processed soil were amended with 50 mg of 

ammonium sulphate (Substrate for nitrifying 

bacteria) in the form of solution. Alcohol extracts 

of each medicinal herb were added to plastic cup 

separately at 1% rate of dry soil mass. Commercial 

inhibitors, N-serve ([2-chloro-6(trichloromethyl) 

pyridine]) and Dicyandiamide (DCD) were also 

added at a rate of 20 and 100 µg /g of soil 

(Sahrawat et al., 1987) respectively. Untreated 

control cups were also included in the experiment 

and three replicated cups were used for each 

treatment. The experiment was laid out in 

completely randomized design (CRD). The 

moisture content was maintained at 60 % water 

holding capacity (WHC) and incubated for 15 and 

21 days at 25 °C. At the end of each incubation 

period soil samples were taken and analyzed for 

NH 4-N, NO 3-N and pH. The NH 4-N and NO 3-N 

were analyzed using the procedure described in 

Keeney and Nelson (1982). 

Table 1 Chemical characteristics of soil used in the experiment. 

pH TN OM BS Na K Ca Mg CEC P Fe Mn Zn Cu 

(%) (%) (%) 

cmol /kg mg /kg 

7.6 0.15 3.4 68 0.4 5.53 4.19 1.03 21.2 46 18 26 6.4 0.14 

Table 2 Description of some of traditional medicinal herbs screened as nitrification inhibitors 

(Woldemichael, 1987). 

Common name Scientific name Family name Parts use 

Wormwood Artemisa afra Compositae Leaf 

Lemmon grass Cymbopogon citratus Poaceae (Graminae) Leaf 

Bitter leaf Vernonia amygdalina Del. Aseraceae Leaf 

- Croton macrostachyus Hochst. Euphorbiaceae Leaf 

Herb of grace Ruta chalepensis L. Rutaceae Fruit 

- Thymus serpyllum Lamiaceae Leaf 

- Haginia abyssinica Bruce. Rosaceae Flower 

- Echinops spp. Compositae Tap root 

Glove tree Eugenia caryphyllata Thunb. Myrtaceae Fruit 

Ginger Zingeber officinale Zingeberaceae Rhizome 

Neem Azadiracta indica Meliaceae Seed

The percentage nitrification inhibition by 

each inhibitor was calculated according to the 

following formula described in Sahrawat (1980). 

% Inhibition = 100(C-S)/C 

Where, C is the amount of NO 3 -N in the 

untreated control cups and S is the amount of NO 3 

-N in the treated cups. Data on NH 4-N and NO 3 - 

N were subjected to statistical analysis using SAS 

software version 6.2. When found significant, 

means were ranked using Duncan’s multiple range 

tests. 

Secondary screening 

From among Ethiopian traditional 

medicinal herbs tested as nitrification inhbitors in 

the preliminary screening experiment three herbs 

namely Artemis afra, Echinops spp and Eugenia 

caryophyllata were found suprior to the remaining 

others. These herbs were rescreened for 

confirmation uing the same procedure described 

in preliminary screening experiment. 



Preliminary screening 

The effect of alcohol extracts of some 

Ethiopian medicinal herbs and commercial 

inhibitors on pH, NH 4-N and NO 3-N contents of 

Awassa clay loam (Eutric fluvisol) soil after 2 and 

3 weeks of incubation period is presented in Table 

3. The tested herbs varied significantly in 

conserving N in the form of NH 4-N. At the end of 

2 nd week, 8 of herbs maintained significantly (P< 

0.01) high amount of NH 4-N and low NO 3 -N in 

the soil which is an indication of nitrification 

inhibition ability (Mesinger et al., 1980). The 

relatively low amount of NH 4-N in the soil treated 

with extracts of the remaining herbs and control 

is due to its oxidation to NO 3 -N. At the end of 3 rd 

week, only 4 of herbs maintained significantly high 

amount of NH 4-N in the soil where their respective 

extracts were added as inhibitor. The herbs were 

Table 3 Effect of alcohol extracts of some Ethiopian traditional medicinal herbs on the NH 4-N and 

NO 3 - -N (mg kg -1 )content of Awassa clay loam (Eutri fluvisol) soil amended with ammonium 

sulphate at the rate 50mg /100g of soil and incubated at 25°C. 

Incubation periods (weeks) 

Medicinal herbs/ 2 3 

Inhibitors pH NH 4-N NO 3-N pH NH 4-N NO 3-N 

Artemisa afra 6.65 89.7c* 43.5cd 6.65 42.2cd 56.3c 

Cymbopogon citratus 6.60 68.1def 48.6cd 6.66 21.1g 73.2ab 

Vernonia amygdalina 6.53 67.1def 56.0c 6.56 28.4efg 78.2ab 

Croton macrostachyus 6.65 60.8ef 75.3a 6.43 21.2g 80.7a 

Ruta chalaepensis 6.69 71.17de 49.4cd 6.90 40.0d 57.8c 

Thymus serpyllum 6.50 55.4f 74.1a 6.50 24.3fg 79.9a 

Echinops spp. 6.73 77.8cd 41.2e 6.56 51.4c 53.7c 

Haginia abyssinica 6.52 75.5cd 52.6c 6.50 36.2de 73.2ab 

Eugenia caryophyllata 7.14 127.6b 19.7f 7.13 92.9b 34.4de 

Zingeber officinale 6.47 63.3ef 63.7b 6.50 34.5def 70.2ab 

Neem (Azadiracta indica) 7.24 129.3b 22.1f 7.53 88.7b 38d 

N-serve 7.95 137.1ab 16.8f 7.90 110.9a 29.8ed 

DCD 7.03 145.5a 15.1f 7.23 114.2a 25e 

Control 6.50 55.9f 69.5ab 6.43 26.4g 79.8a 

CV (%) - 8.5 9.05 - 12.17 8.99 

* Means within colums followed by the different letters are significantly diferrent from each other at P< 0.05 according to 

DMRT

78 

Eugenia caryophyllata, Echinops spp., Artemisa 

afra and neem (Azadiracta indica). These herbs 

were previously reported to have antibacterial 

9killing bacteria) nature against human pathogenic 

bacteria (Mintesnot and Mogessie, 1999). Thus, 

their effectiveness as nitrification inhibitor in this 

study is probably due to their bactericidal property 

against ammonia oxidizing bacteria. The 

occurrence of plant extracts and their exudates as 

nitrification inhibitor are also reported by several 

authors (Pacncholy, 1973; Sahrawat and Mukerjee, 

1977; Rice and Paavolainen et al., 1998). 

Eugenia caryophyllata performed as 

effective as neem but none of the herbs performed 

as effective as N-serve and DCD. Similar to the 

present finding, the significant nitrification 

inhibitory property of neem has been reported by 

several authors (Sahrawat and Parmar, 1975; 

Prasad and Power, 1995). 

According to Prasad and Power (1995), 

epinibin, nibin, desacetylnimbin, salanin, 

desacetylacinin and azadiractin collectivelly 

known as melancins are chemical constituent of 

neem responsible for markedly retardation of 

nitrification. 

In the study of these commercial 

inhibitors, N-serve and DCD were superior in their 

nitrification ability in inhibiting nitrification. These 

inhibitors retard nitrification by inactivation of 

cytochrome oxidase of ammonia oxidizing 

bacteria (Meisinger et al., 1980). 

There was a decline in NH 4-N and a 

parallel build up of NO 3-N in treated soil with all 

inhibitors with time, though the decline varied 

among herbs. This is due to the fact that inhibitors 

are organic compounds themselves and are 

subjected to the attack by microorganisms in the 

soil eventually leading to loss of inhibitory 

property (Yesuf and Vancleemput, 2000). 

Effectiveness of Ethiopian medicinal 

herbs along with neem and commercial inhibitors 

expressed as percent inhibition of nitrification in 

Awassa clay loam soil is shown in Table 4. Of all 


herbs only Artemisa afra, Echinops spp., Eugenia 

caryophyllata and Neem effectively inhibited 

nitrification. Averaged over two incubation 

periods, these herbs inhibited nitrification by 33, 

37 and 64 and 60 % respectively. Comparatively, 

N-serve and DCD were superior and inhibited 

nitrification by 70 and 73% respectively. 

Secondary screening 

The effect of alcohol extracts of Artemisa 

afra, Echinops sp. and Eugenia caryphyllata and 

commercial inhibitors on NH 4-N and NO 3-N 

contents of soil is presented in Table 5. These 3 

herbs conserved significantly (P

commercial inhibitors on soil pH which was 

monitored in the secondary screening experiment 

is shown in Table 7. Both medicinal herbs and 

commercial inhibitors maintained the pH of the 

soil at the higher level than that of the control. 

This is probably due to the fact that the inhbitors 

prevented the oxidation of NH 4-N to NO 3-N. As a 


result the production of H ions were suppressed. 

In a similar study, Yesuf and Vancleemput (2000) 

reported that inhibitor treated soils have higher pH 

than untreated ones. Thus, the nitrifcation 

inhibitors in addition to inhibiting nitrification, are 

also prevented the soil from acidification. 

Table 4 Nitrification inhibition percentage by Ethiopian traditional medicinal herbs. 

% Inhibition 

Medicinal herbs/Inhibitors Incubation periods (weeks) 

2 3 Mean 

Artemisa afra 37c* 29c 33 

Cymbopogoncitratus 30dc 8d 19 

Vernonia maygdalina 19e 0 d 10 

Croton macrostachyus 0e 0d 0 

Ruta chalaepensis 28cd 27c 28 

Thymus serpyllum 0e 0d 0 

Echinops spp 40c 32c 37 

Haginia abyssinica 24d 8d 16 

Eugenia Carophyllata 71ab 57ab 64 

Zingeber officinale 8e 12d 10 

Neem(Azdiracta indica) 68b 52b 60 

N-serve 75ab 63ab 70 

DCD 78a 68a 73 

Control - - - 

CV(%) 13 23 - 

* Means within colums followed by the different letters are significantly diferrent from each other at P< 0.05 according to 

DMRT. 

Table 5 Effect of Ethiopian medicinal herbs on the NH 4-N and NO 3-N content in Awassasoil (Eutric 

fluvisol) amended with ammonium sulphate and incubated for various periods of time. 

Incubation periods (weeks) 

Medicinal herbs/Inhibitors 2 3 4 2 3 4 

NH 4 + -N (mg/kg) NO3-N (mg/kg) 

Artemis afra 78.81c* 39.12cd 34.7c 42.74b 56.09b 69.4b 

Echinps spp 92.17c 52.31c 39.9c 41.9b 47.31b 54.4c 

Eugenia caryophyllata 131.96b 102.44b 83.3b 22.39c 30.03c 38.5d 

N-serve 150.89a 114.25ab 93.5ab 15.73c 25.7c 35.3d 

DCD 156.19a 122.86a 96.3d 13.87d 23.87c 33d 

Control 49.23d 39.11d 20.1d 70.24a 72.39a 81a 

CV (%) 7.26 11.29 11.3 8.98 14.27 10.3 

* Means with in colums followed by the different letters are significantly differrent from each other at P< 0.05 according to 

DMRT.

80 

CONCLUSION 

In conclusion, out of Ethiopian 

traditional medicinal herbs tested in this 

experiment, three of them namely Eugenia 

caryophyllata, Echinops spp. and Artemisa afra 

were identified to be nitrification inhibitors. They 

were found to be effective up to the end of 4 weeks 

of incubation. The highest percentage nitrification 

inhibion was achieved at the end of 2 weeks of 

incubation and the least at the end of 4 weeks of 

incubation. Eugenia caryophyllata was ranked the 

best among Ethiopian traditional medicinal herbs 


Table 6 Nitrification inhibition percentage by Ethiopian traditional medicinal herbs. 

Medicinal herbs/inhibitors % Inhibition 

Incubation periods (Weeks) 

2 3 4 

Artemisa afra 39c* 22b 14c 

Echinops spp. 43c 35b 33b 

Eugenia caryophyllata 68b 59a 38b 

N-serve 78ab 64a 56a 

DCD 80a 67a 59a 

Control - - - 

CV (%) 9 17 20 

* Means within colums followed by the different letters are significantly different from each other at P< 0.05 according to 

DMRT. 

Table 7 Effect of nitrification inhibitors on the 

pH of Awassa clay loam (Eutric fluvisol) 

soil amended with 50 mg of ammonium 

sulphate and incubated at 25°C. 

Medicinal herbs/ Incubation periods (weeks) 

Inhibitors 2 3 4 

Artemisa afra 6.9 6.8 6.7 

Echinops spp 6.9 6.8 6.7 

Eugnia caryophyllata 7.0 6.9 6.8 

N-serve 7.4 7.3 6.9 

DCD 7.3 7.3 7.0 

Contro 6.5 6.7 6.5 

Mean 7.0 6.97 6.76 

SE± 0.13 0.13 0.07 

and gave comparable result as that of Neem in 

inhibiting nitrification. However, none of the herbs 

out shined the commercial inhibitors. Application 

of both heral and commercial inhibitors together 

with ammonium sulphate prevented the soil from 

acidification compared to the soil that received 

ammonium sulphate alone. Thus, alcohol extracts 

of these herbs could be blended with ammonium 

containing fertilizers to increase N-use efficiency 

of crops. They are particulary applicable to 

subsistent farmers who can not afford to buy 

commercial inhibitors. However, further extensive 

laboratory and field studies are required. On the 

other hand, to the authors’ knowledge there has 

never been experience of testing and using 

commercial inhibitors in Ethiopia. Thus, this study 

proved that commercial inhibitors that are N-serve 

and DCD were found to be effective inhibitors. 

Thus, These inhibitors could be imported are used 

by large scale private farmers which can afford 

the incurred expenes. 

ACKNOLEDGWEMENTS 

The senior author would like to thank Mr. 

Kumsa Baissa and Mr. Abebe Abay for their 

unreserved support rendered to him in the course 

of this experiment. He is also highly indepted to 

Ethiopian government for sponsoring his study.


Hauck, R.D. 1984. Technological approaches to 

improving the efficiency of nitrogen fertilizer 

use by crop plants, pp. 551 – 560. In R.D. 

Hauck (ed.). I Nitrogen in crop production. 

ASA-CSSA-SSSA, Madison, WI. 

Jarvis, C. S., E. A. Stockdale, M. A. Shepherd and 

D. S. Powlson.1996. Nitrogen mineralization 

in temperate agriculture soils. Process 

and measurements. Adv. in Agron. 57: 187- 

233. 

Keeney, D.R. and D.W. Nelson. 1982. Nitrogeninorganic 

forms, pp. 648-654. In A. L 

Page and D. R. Keeney (eds.) Methods 

of soil analysis part 2. Chemical 

and microbiological properties of soil 

analysis. Agronomy Monograph No. 9 ASA, 

Madison. 

Lindsay, W. L and W. A. Novell. 1978. 

Development of a DTPA soil test for Zn, Fe, 

Mn and Cu. Soil Sci. Soc.Am.J.42:421-428. 

Mclean, E.O. 1982. Soil pH and lime requirement. 

pp. 199-224. In A. L. Page, R. H. Miller and 

D. R. Keeney (eds.). Methods of soil analysis 

part II, 2 nd ed. ASA, Madison. 

Meisinger, J.J., G. W. Randall and M. L. 

Vitosh.1980. Nitrification Inhibitors : 

Potential and Limitations, Spec.Pub.38, 

ASA and SSSA, Madison,WI. 129p. 

Mintesnot, A. and A. Mogessie. 1999. Assessment 

of the antibacterial activity of some traditional 

medicinal plants on food- borne pathogens. 

Ethiop. J. Health Dev. 3: 211-216. 

Nelson, N. D. W and Huber. 2001. Nitrification 

Inhibitors for Corn Production. National 

Handbook. Iowa State University, extension, 

electronic version, IOWA. 

Oslen, R. S. and L. E. Sommers. 1982. Phosphorus. 

pp. 403-427. In A. L Page , R.H. Miller and 

D. R. Keeney ( eds.). Methods of Soil 

Analysis , Part II 2 nd ed. ASA, Madison. 

Paavolainen, L., V. Kutunen and A. Smolander. 


1998. Inhibition of nitrification in forest soil 

by monoterpenes. Plant and Soil 205: 147- 

154. 

Prasad, R. and F.J. Power. 1995. Nitrification 

inhibitors for agriculture, health and the 

environment. Adv. in Agron. 54: 233-278. 

Rice, E. L. and S. K. Panchloy. 1973. Inhibition 

of nitrification by climax ecosystem III. 

Inhibition other than tannin. Am. J. Bot. 61: 

1095-1130. 

Rowell, D. L. 1994. Soil Science: Methods and 

applications. Longman Scientific and 

technical, Singapore. 350 p. 

Sahrawat, K.L. and S.K. Mukerjee. 1977. 

Nitrification inhibitors. I. Studies with 

karanajin, a furano flaviniod from karanaja 

(Pongamia glabra vet.) Plant and Soil 

47: 27-36. 

Sahrawat, K.L. and B.S. Parmar.1975. Alcohol 

extract of Neem (Azadirzacta indica L) seed 

as nitrification inhibitor. Indian J. Soc. Soil 

Sci. 23: 131-134. 

Sahrawat, K.L., D.R. Keeney and S.S. Adams. 

1987. Ability of nitrapyrin, diciandiamide and 

acetylene to retard nitrification in mineral and 

organic soil. Plant and Soil 101: 179-182. 

Sahrawat, K.L. 1980. On criteria for comparing 

the ability of compounds for retardation of 

nitrification in soil. Plant and Soil 55: 487- 

490. 

Taye, B., H. Yesuf, S. Sahlemedhin, G. Amanuel, 

H. Hassena, D. G. Tanner, T. Tesfaye and G. 

Takele. 2002. Optimizing Fertilizer Use in 

Ethiopia: Correlation of Soil Analysis with 

Fertilizer Response in Hetosa Woreda, Arsi 

Zone, Addis Ababa. 

Thoman, G. W. 1982. Exchangeable cation. p.159. 

In A. L. Page, R.H. Miller and D.R. Keeney 

(eds.). Methods of Soi, Analysis, Part II 2 nd 

ed. ASA, Madison. 

Walkley, A and I.A. Black . 1934. An examination 

of the digitijareff method for determining soil 

organic matter and a proposed modification

82 

of chromic acid titration method. Soil Sci. 153: 

100-107. 

Wassie, H., T. Mala, Y. Osotsapar and V. Verasan. 

2004. Investigation on the nitrification 

potential of some soils occurring in sothern 

and central Ethiopia. Kamphaengsean Acad. 

J. 2: 111-125 

Woldemichael, K. 1987. A glossary of Ethiopian 

Plant Names. 4 th ed. Addis Ababa 245 p. 


Yesuf, A., O. Vancleemput. 2000. The effect of 

urease inhibitors and their mixture on the 

mineral N content of sandy loam soil after urea 

application, pp. 23-34. In W. Chekol and E. 

Mersha (eds.). Nutrient management for 

improved soil/crop productivity in Ethiopian 

agriculture. Proceedings of the 5 th coference 

of ESSS, Addis Ababa.


Seasonal Characteristics of Wood Formation in the Elite 

Genetic – Based Eucalyptus camaldulensis Dehnh. 

Teera Veenin 1 *, Tadashi Nobuchi 2 , Minoru Fujita 2 and Somkid Siripatanadilok 3 

ABSTRACT 

Seasonal characteristics of wood formation were investigated using the knife-cutting method 

in five superior clones of six years old Eucalyptus camaldulensis in the plantation of eastern Thailand. 

In the observation of wound tissue formed through knife insertion, cambial initials at the time of knife 

– cutting could be estimated. Based on these positions, pursuit of time course of wood formation or 

radial growth was carried out. 

In rainy season all clones showed active radial growth. The radial growth rate decreased in dry 

season. In the comparison of radial growth among five clones, clones 3 and 4 maintained better growth 

rate than other clones even in dry season. 

Key words: Eucalyptus camaldulensis, wood formation, radial growth, elite genetic-based, seasonal 

characteristics 

INTRODUCTION 

In the tropical monsoon area such as 

Thailand with long dry season, Eucalyptus 

camaldulensis, the fast growing species which can 

be adaptable to many climate condition (Kauman 

et al., 1995), particularly, in the dry habitat area 

has been selected as one of the potential plantation 

species. Seedling or young plants for plantation 

are generally obtained from seeds or cuttings. For 

producing much amount of young plants, the group 

of Faculty of Forestry, Kasetsart University started 

to produce young plants through tissue culture 

technology in 1989. They could succeed to get 

young plants from tissue culture, succeeded in 

naturalization and planted them in field in 1992. 

The growth of each plant in the field of one-year- 

old were varied distinctively. The trees from 

superior clones showed better growth and more 

uniformity than the trees from seedling 

(Siripatanadilok et al., 1992). 

The one important target in E. 

camaldulensis plantation from tissue culture is to 

evaluate the seasonal characteristics of wood 

formation or radial growth of each clone. 

Especially, the selection of clones with good 

growth rate in dry season is the key research point. 

From the view point of the methodology 

to investigate the seasonal characteristics of wood 

formation or radial increment, the pinning method 

(Wolter, 1968) is excellent method because it is 

simpler and easier, marks accurately and can be 

adapted to a range of small to large diameter trees 

(Nobuchi et al., 1993). In this method cambial 

1 Department of Forest Products, Faculty of Forestry, Kasetsart University, Bangkok 10900, Thailand. 

2 Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan. 

3 Department of Forest Biology, Faculty of Forestry, Kasetsart University, Bangkok 10900, Thailand. 

* Correspording auther, e-mail E-mail:teera50@yahoo.com 


84 

initials at the time of pinning are recorded in the 

wound tissue (Nobuchi et al., 1995). This method 

mostly has been applied to the study of temperate 

zone trees (Shimaji and Nagatsuka, 1971; Kuroda 

and Shimaji, 1983, 1984) and rarely applied to 

tropical trees. 

For the pinning method to tropical trees, 

Shiokura (1989) and Nobuchi et al. (1995, 1996) 

used a nail instead of a pin because tropical trees 

have generally thick and hard bark. The wound 

tissues induced by either a pin or a nail insertion 

were short in longitudinal direction, therefore, it 

was difficult to obtain appropriate sections from 

the center of tiny wound tissues. To make this 

method easier “knife-cutting method” proposed by 

Fujiwara (1992) was the alternative method to 

improve the pinning method. Kuroda and Kiyono 

(1997) investigated the seasonal rhythms of xylem 

growth measured by knife-cutting method that 

they called the wounding method. 

The researches of the seasonal 

characteristics of wood formation of genus 

Eucalyptus are very much limited. Ogata et al. 

(1996) analysed the seasonality of wood formation 

in E. tereticornis and E. saligna by knife-cutting 

method and Kondo et al. (1996) analysed the radial 

growth in E. dalrympleana by knife-cutting 

method. Particularly, the radial growth compared 

among clones of E. camaldulensis are not reported. 

In this report, the knife-cutting method 

was applied to compare the radial growth among 


5 superior clones of E. camaldulensis and the data 

were discussed in correlation with the climatic data 

near the plantation site. The objective of this study 

aims to screen for the eucalyptus clone which is 

fast growing even in dry condition. 


Five superior clones of six-year-old E. 

camaldulensis in the plantation site at Sakaew 

province in eastern part of Thailand were selected 

as sample trees. The name of clones were T5, Kitti, 

S9, Y2 and K2, which were coded in this report as 

clone 1, 2, 3, 4 and 5, respectively. Diameter at 

breast height ranged from 18.5 to 24.1 cm. 

Knife-cutting method 

Knives of 0.5 mm in thickness and 10 

mm in width were used for marking. To investigate 

radial growth around the trunk, marking positions 

were done at four sides of trunk (southwest side 

and opposite, southeast side and opposite). The 

knife was inserted deep enough to reach the border 

of cambium and xylem and removed immediately. 

As shown in Figure 1, the distance between two 

points was about 4 cm in tangential direction and 

about 10 cm in longitudinal direction. The 

marking was done every month from August 1998 

to July 1999. Table 1 indicates the dates of 

marking. 

Blocks of 3 cm in radial depth containing 

� � � � � � 

Aug. 98 Oct. 98 Dec. 98 Sep. 98 Nov. 98 Jan. 99 

� � � � � � 

Feb. 99 Apr. 99 Jun. 99 Mar. 99 May. 99 Jul. 99 

Southwest side Southeast side 

Figure 1 Schematic illustration of the marking point on a trunk.

the wounding tissues were collected on August 28, 

1999 by a saw and a chisel without felling the trees. 

Light microscopy 

Wood blocks containing the marked parts 

were incised through the center of long slit-like 

wound tissue by handsaw. Thereby, the wood 

blocks were separated into two parts. After first 

part of wood blocks were softened in boiling water, 

transverse sections 20-30 µm in thickness were 

cut using a sliding microtome. They were stained 

with safranin fast green. Wound tissues were 

observed under a light microscope. The other part 

of the wood block was also used for the 

measurement of radial growth. 

Measurements of radial growth 

Second part of wood blocks containing 

the wound tissues were sanded with a series of 

sand papers of No.40, 60, 80 and 120. Each of the 

sanded transverse sections were scanned using 

Epson ES-8000 image scanner and Adobe 

Photoshop program. For the measurement of 

radial growth in the computer, a growth ring 

boundary was used as a reference line. That is, 

radial growths were measured by computer as the 

distance between growth ring boundary and 

estimated cambial initials at the time of each 

marking. 


Characteristics of growth rings 

The characteristics of growth ring 


boundary as a reference line were investigated in 

transverse sections and scanned pictures. The 

concentric line was necessary for the measurement 

of radial growth of different marking positions 

around the circumference of the trunk. 

As shown in Figure 2, the growth ring 

boundaries of all clones were not so distinct as 

those of temperate zone trees. It was because 

vessel diameters showed little variation throughout 

one year. However, late wood-like zone had 

slightly smaller diameter of vessels together with 

thicker fiber cell walls. Based on those 

characteristics, the growth ring boundary was 

adopted as a reference line. 

Anatomical characteristics of wound tissue 

A transverse section of wound tissue 

induced by knife-cutting is shown in Figure 3. 

Nobuchi et al. (1995) and Ogata et al. (1996) 

reported that the wound tissue was divided into 

two zones. In zone 1 in Figure 4, cells in cambial 

initials and enlarging zone were directly affected 

by knife-cutting. They were crushed and cell wall 

formation were interrupted in the cells living at 

the time of marking. Therefore, they retained the 

cell wall organization, even if they were deformed. 

In zone 2, cells in cambial initials and enlarging 

zone were not crushed but indirectly affected by 

knife-cutting resulting in the formation of 

abnormal cells. This zone showed a region of 

normal wood (NW), callus-like cells (C), radially 

flat cells (FC) and a layer with small diameter of 

vessels (S). 

In the region with small diameter of 

Table 1 Dates of marking. 

Southwest side and opposite side Southeast side and opposite side 

August 12, 1998 September 12, 1998 

October 23, 1998 November 21, 1998 

December 26, 1998 January 23, 1999 

February 21, 1999 March 17, 1999 

April 27, 1999 May 22, 1999 

June 20, 1999 July 17, 1999

86 

vessels a point where the number of cell rows 

increased tangentially or the group of radially flat 

cells was noted. This increase could have been 

caused by the anticlinal division of cambial initials. 

The line which connected the innermost point of 

anticlinal divisions (arrowheads) was considered 

theoretically to be the location of cambial initials 


at the time of marking. Practically, this line 

coincided with the line of the innermost smallsize 

vessels. This line was, therefore, adopted as 

the marker of cambial initials at the time of 

marking and used for measurement (Nobuchi et 

al., 1995) 

Figure 2 A transverse section of clone 2 showing growth ring boundary or reference line. Arrowheads 

represent the reference line. Arrows in Fig. 2-5 indicate the direction of bark side. Scale bars 

= 250 µm. 

Figure 3 A transverse section of clone 2 showing wound tissue induced by knife-cutting. The wound 

tissue was divided into two zones. Scale bars = 250 µm.

Seasonality of radial growth and its relation to 

monthly rainfall 

The distance between the reference line 

and the estimated cambial initials at the time of 

each marking was measured (line 1 in Figure 5). 

This indicates the radial growth in the terms from 

the latest growth ring boundary to the time of a 

given marking. However, It is quite natural that 

some errors might arise because the radial growth 

is not homogenous along the circumference of the 

trunk. To avoid these errors, the radial growth 

was expressed as the percentage of line 1 divided 

by line 2 in Figure 5 (Ogata et al., 1996). 

The percentage of radial growth rate 

compared among 5 clones are shown in Table 2. 

The radial growth rate in southwest side and 

southeast side of the trunks were the great 

dispersions. Shimaji and Nagatsuka (1971) 

reported that the greater dispersions of data might 

be attributed to the greater tangential width of 

pinning area. To eliminate such dispersions of 

data, the pinning area should be as narrow as 

possible. The dispersion of data did not overrule 

the fact that each growth curve was accurate 

growth pattern of the limited locus of each tree 

stem. Therefore, the growth curve was separated 

into two as shown in Figure 5 (a, b). 


Figure 5 A scanned transverse section of clone 3 

showing the measurement procedure of 

growth rate. Line 1 shows the distance 

from the growth ring boundary (GRB) 

to the estimated cambial initials at the 

time of marking (CIM) and line 2 shows 

the distance from GRB to the cambial 

initials at the time of collecting wood 

block (CIC). Radial growth rate was 

calculated as the percentage of line 1 

divided by line 2. 

Figure 4 A transverse section of clone 2 showing the enlargement of zone 2. Arrowheads show the 

estimated cambial initials at the time of marking. Normal wood (NW), callus-like cells (C), 

radially flat cells (FC) and small diameter of vessels (S). Scale bars = 100 mm.

88 


Table 2 Growth rate (%) of each marking in southwest side (SW) and southeast side (SE) of the trunks 

in 5 clones and monthly rainfall. 

Year and Clone 1 Clone 2 Clone 3 Clone 4 Clone 5 Monthly 

Month Radial growth Radial growth Radial growth Radial growth Radial growth rainfall 

rate (%) rate (%) rate (%) rate (%) rate (%) (mm) 

SW SE SW SE SW SE SW SE SW SE 

1998 Aug. 18.7 33.0 27.5 39.4 32.1 153.0 

Sep. 39.2 32.5 36.6 53.9 46.8 233.6 

Oct. 26.7 44.9 42.2 41.8 37.4 202.1 

Nov. 46.1 51.1 44.4 64.9 62.7 8.1 

Dec. 36.8 41.6 48.1 41.1 49.1 0.0 

1999 Jan. 63.0 59.3 59.4 67.0 70.5 15.0 

Feb. 37.8 52.8 56.7 53.2 51.3 0.0 

Mar. 49.3 52.2 63.1 67.4 66.5 122.0 

Apr. 47.0 60.7 59.8 50.5 50.0 78.9 

May. 72.2 77.4 80.0 77.1 80.1 129.0 

Jun. 66.9 81.6 82.6 80.7 79.4 306.0 

Jul. 87.2 86.6 92.5 86.9 89.9 83.4 

Figure 5 shows a transverse section with 

a wound tissue formed through knife insertion. 

Line 1 indicates the distance between the growth 

ring boundary (GRB) and the estimated cambial 

initials at the time of marking (CIM). Line 2 shows 

the distance between GRB and the cambial initials 

at the time of collecting wood block (CIC). Radial 

growth rate was expressed as the percentage of 

line 1 divided by line 2. 

Data of radial growth rate (%) together 

with monthly rainfall are indicated in Table 2. The 

data of radial growth rate are also shown in Figure 

6 (a, b) and corresponding monthly rainfalls are 

shown in Figure 7. 

In Figure 6, 7 and Table 2, radial growth 

was judged to be active in rainy season and radial 

growth rate decreased in dry season for all 5 clones. 

This result was in coincidence with Nobuchi et 

al. (1995, 1996) who investigated seasonal 

characteristics of wood formation in Hopea 

odorata, Shorea henryana and Tectona grandis in 

Thailand. 

In the former part of dry season, wood 

formation was considered not to stop even though 

the growth rate decreased. Considering that 

E. camaldulensis is the evergreen tree species, it 

is well understood. In the latter half of dry season, 

radial growth rate showed conspicuous decrease. 

We could not identify whether the wood formation 

completely stopped or not. For this investigation 

the observation of cambial pattern is necessary. 

Seasonal changes of soil moisture content are also 

important research point to discuss the necessary 

water supply to the cambial activities. 

For the comparison among 5 clones, the 

radial growth patterns especially in the latter half 

of dry season were investigated. Some clones 

including clones 1, 2 and 5 showed more 

conspicuous decrease of growth rate. Clones 3 and 

4 showed better growth rate. Especially clone 3 

showed continuous radial growth even at the end 

of dry season which the soil moisture content 

would be very low. This result supported the idea 

that clone 3 would be more tolerant in dry habitat. 

When we consider the selection of E. 

camaldulensis clones, this is important information.

Radial growth rate (%) 

Radial growth rate (%) 

100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 


Aug.98 Oct.98 Dec.98 Feb.99 Apr.99 Jun.99 

Sep.98 Nov.98 Jan.99 Mar.99 May.99 Jul.99 

Clone1 

Clone2 

Clone3 

Clone4 

Clone5 

a 

Clone1 

Clone2 

Clone3 

Clone4 

Clone5 

Figure 6 Comparison of the growth curves among five clones in the southwest side (a) and the southeast 

side (b). 

Rainfall (mm) 

350 

300 

250 

200 

150 

100 

50 

0 

Aug.98 Sep.98 Oct.98 Nov.98 Dec.98 Jan.99 Feb.99 Mar.99 Apr.99 May.99 Jun.99 Jul.99 

Figure 7 Monthly rainfall near the plantation site at Sakaew province from August 1998 to July 1999. 

b

90 

CONCLUSIONS 

In the investigation of seasonal 

characteristics of wood formation using the knifecutting 

method in five superior clones of E. 

camaldulensis, the results were concluded as 

follows : 

1. The radial growth of all 5 clones 

increased in rainy season and decreased in dry 

season. In the dry season, the decrease of growth 

rate was more conspicuous in the latter half. 

2. In the comparison among 5 clones, 

some clones including 1, 2 and 5 showed more 

conspicuous decrease than 3 and 4 in dry season. 

Especially, clone 3 showed continuous growth 

even in the dry season, which might support this 

clone to be the candidate for the plantation in drier 

habitat. 


The authors would like to thank Japan 

Society for the Promotion of Science (JSPS) for 

financial support during the research in Japan and 

to Kasetsart University Research and Development 

Institute (KURDI), which supported the research 

funds in Thailand. 

Special thanks are extended to Forest 

Industry Organization (FIO) for permission of field 

experiment and to the members of Laboratory for 

Structure of Plant Cells, Graduate School of 

Agriculture, Kyoto University, Japan for valuable 

advice and research facilities. 


Fujiwara, K. 1992. A marking method using small 

knife for soft X-ray microdensitometry. 

Jumoku Nenrin 5: 27-30. 

Kauman, W.G., I. Gerard, H. Jiging and W. 

Huaijun. 1995. Processing of Eucalypts. 

Commonwealth Forestry Review 74: 147- 

154. 

Kondo, Y., T. Fujii, K. Fujiwara and R. Nakada. 

1996. An analysis of the radial growth in 

eucalyptus and tulip tree by knife-cutting 

method (in Japanese). Abstracts of the 46 th 

Annual Meeting of the Japan Research 


Society. Kumamoto, Japan. 

Kuroda, K. and K. Shimaji. 1983. Traumatic resin 

canal formation as a marker of xylem growth. 

For. Sci. 29(3): 653-659. 

Kuroda, K. and K. Shimaji. 1984. The pinning 

method for marking xylem growth in 

hardwood species. For. Sci. 30(2): 548-554. 

Kuroda, K. and Y. Kiyono. 1997. Seasonal 

rhythms of xylem growth measured by the 

wounding method and with a banddendrometer 

: an instance of Chamaecyparis 

obtusa. IAWA J. 18(3): 291-299. 

Nobuchi, T., T. Fujisawa and H. Saiki. 1993. An 

application of the pinning method to the 

marking of the differentiating zone and to the 

estimation of the time course of annual ring 

formation in Sugi (Cryptomeria japonica). 

Mokuzai Gakkaishi. 39: 716-723. 

Nobuchi, T., Y. Ogata and S. Siripatanadilok. 

1995. Seasonal characteristics of wood 

formation in Hopea odorata and Shorea 

henryana. IAWA J. 16(4): 361-369. 

Nobuchi, T., S. Janmahasatien and M. Sakai. 1996. 

Seasonal change of wood formation and some 

characteristics of heartwood formation in Teak 

(Tectona grandis L.) Plantation. Kasetsart 

J. (Nat. Sci.) 30(2): 254-263. 

Ogata. Y., T. Nobuchi and M. Fujita. 1996. An 

analysis of the seasonality of wood formation 

in cinnamon trees and eucalyptuses by knifecutting 

method. Kyoto Uni. Bull. For. 68: 

116-126. 

Shimaji, K. and Y. Nagatsuka. 1971. Pursuit of 

the time sequence of annual ring formation. 

J. Jap. Wood Res. 17(3): 122-128. 

Shiokura, T. 1989. A method to measure radial 

increment in tropical trees. IAWA Bull.n.s. 

10(2): 147-154. 

Siripatanadilok, S. and B. Thaiutsa. 1992. 

Application of vegetative propagation to 

improve timber yield of red gum (Eucalyptus 

camaldulensis Dehnh). RD & E Project Semi 

Annual Report No. 6. Department of Forest 

Biology, Faculty of Forestry, Kasetsart 

University. 28 p. 

Wolter. K.E. 1968. A new method for marking 

xylem growth. For. Sci. 14: 102-104.


Development of Catalase Gene Nuclear DNA-Based Marker for 

Population Genetic Analysis in Thai Teak (Tectona grandis L.f.) 

ABSTRACT 

Jongkon Cheua-ngam 1 and Hugo Volkaert 1,2 * 

The teak Catalase gene was amplified and cloned using consensus PCR primers designed 

based on sequences from other plant species. The obtained teak Cat DNA sequences were used to 

develop specific primers. The specific primer set could successfully amplify a specific DNA fragment 

from teak in all populations studied. Using PCR-SSCP, 5 different alleles were detected. The tested 

nuclear gene primer set had considerable potential as a DNA marker for population analysis in teak. 

This approach could be used to specifically amplify fragments in other plant species and applied to 

study evolution, population genetics, outcrossing rate and mating patterns. 

Key words: Catalase, teak, Tectona grandis, population genetics, SSCP 

INTRODUCTION 

Thailand used to be an important supplier 

of teak timber to the world market. However, due 

to overharvesting, yields declined after about 1960 

(De’Ath, 1992; Graudal et al., 1999). The 

unsustainable extraction of timber from natural 

forest has become a major cause of concern about 

deforestation, forest degradation and soil erosion. 

Therefore, a complete ban of commercial 

exploitation of natural forests was decreed in 1989. 

Plantations of quality hardwood species under 

sustainable management are seen as an alternative 

to timber extraction from natural forests. A genetic 

improvement program is an essential component 

of a successful and sustainable plantation project. 

Knowledge of genetic variation within and 

between populations of teak is important for both 

conservation of remaining genetic resources and 

breeding for plantation varieties. 

Molecular genetic analysis techniques 

are now widely used to study genetic diversity and 

gene flow in plants. DNA markers are being used 

for fast surveys of genetic variation in several crop 

improvement programs, as a tool in classification, 

selection, and breeding. The development of 

appropriate DNA-based markers for teak would 

make it possible to study gene-flow and genetic 

diversity aspects of this important species. A recent 

development in population genetics is the 

application of genealogical approaches to quantify 

diversity and gene flow processes (coalescent), but 

these methods require information on both allele 

frequencies and genealogical relationship of the 

alleles. Recent progress in marker development 

has been based on the detection of nucleotide 

sequence variation within introns of a few specific 

protein-coding genes (e. g. Palumbi and Baker, 

1 Center for Agricultural Biotechnology, Kasetsart University, Kamphaengsaen Campus, Nakhon Pathom 73140, Thailand. 

2 National Science and Technology Development Center (NSTDA-BIOTEC), Science Park, KlongLuang, Pathumthani 12120, 

Thailand. 

* Corresponding author, e-mail: hugo.v@ku.ac.th 


92 

1994; Strand et al., 1997). Now nuclear gene 

sequences are available in genome databases and 

promise to greatly assist the search for new nuclear 

markers especially expressed sequence tag 

libraries (ESTs). The extensive EST databases 

reveal homologs from various plant species at 

shallow or deep levels of evolutionary history from 

which consensus primers for PCR could be 

developed. Therefore, the aim of this study was to 

develop specific Catalase (Cat) nuclear marker 

tools for teak that could be used in management 

of genetic resources and teak breeding programs. 


Sample collection 

Individual young leaves were collected 

from trees in a forest remnant near Potharam, 

Ratchaburi province in April 2003. Additional leaf 

or bud samples were collected from trees in 6 

natural populations in Thailand during August and 

September 2004 (Table 1). With a few exceptions, 

trees sampled from the natural populations were 

at a distance of at least 100 m between each 

individual as measured by GPS in the field. 

Approximately 100 mg tissue was transferred to a 

1.5 ml tube containing extraction buffer within 12 

hours from collection for further processing in the 

laboratory. 

DNA extraction 

The fresh young leaf samples from the 


Potharam population were extracted using DNeasy 

Plant Mini Kit (Qiagen) according to the 

manufacturer’s instructions. The samples of 

individual teak from natural populations were 

extracted using a modified CTAB method (Doyle 

et al., 1989). The DNA concentration of all 

samples was estimated on agarose gel by 

comparing to a standard of known concentration 

(Fermentas). 

Development of catalase gene primers 

Development of consensus primers 

PCR primer pair for Cat genes was 

designed based on information of DNA sequences 

obtained from other plant species. Sequences from 

different plant species were retrieved from publicly 

accessible DNA databases (GenBank/EMBL/ 

DDBJ) by keyword searching. The sequences were 

aligned using the ClustalW program (http:// 

www.ebi.ac.uk/clustalw/index.html). The resulting 

alignments were improved by visual inspection 

using the GeneDoc program. Several criteria for 

selecting consensus PCR primers were used. (1) 

All conserved regions, at least 7 codons long, were 

identified. The primer region should optimally 

be conserved in all plant sequences or at least in 

all dicot plant species. (2) The conserved regions 

were compared for minimal redundancy at the 3’ 

end of a potential primer. (3) Two primers should 

be located in opposite direction, preferably one or 

more introns apart. (4) The length of the target 

DNA sequences should not be longer than 2000 

Table 1 List of plant material and collecting sites. 

Population Geographic region Sample size (tree) 

Tg8-01 Photharam, Ratchaburi 23 1 + 46 2 

Tg8-02 SaiYoke National Park, Kanchanaburi 20 

Tg8-03 MaeMoei National Park, Tak 34 

Tg8-04 ChiangDao National Park, Chiang Mai 29 

Tg8-05 PratooPhaa, Lampang 48 

Tg8-06 SakYai Forest Park, Uttaradit 40 

Tg8-07 Srisatchanalai National Park, Sukhothai 44 

1 individual trees sampled all over the Potharam District 

2 all teak trees in a small forest patch behind WatBoht, Potharam District

p as determined from available genomic 

sequences. (5) The length of potential primer sites, 

their G-C content (especially at the 3’ end), and 

the position of the 3’ end relative to the reading 

frame were additional criteria considered. By 

placing the critical 3’ end at a second codon 

position nucleotide for forward and at the first 

codon position for reverse primer, the possibility 

of a mismatch preventing annealing may be 

reduced. In some cases, primer pairs with suboptimal 

melting temperatures may be necessary 

due to the paucity of conserved regions in the 

sequences. Additionaly, (6) the sequence of 

candidate primers was compared to databases 

using BLAST. If the result of BLAST indicated 

that the primer was similar to sequences from fungi 

or bacteria, this primer sequence was discarded. 

DNA fragments were PCR amplified 

using the consensus primer pairs in a total reaction 

mixture of 25 µl, containing 200 µM dNTPs 

(Promega), 5 pmole of each primers, 1× PCR 

buffer with 2 mM MgCl 2 (Qiagen), 0.3 units of 

Taq DNA polymerase (Qiagen), and 20 ng of 

genomic DNA template. PCR products were run 

on agarose gel. The reproducible PCRs were 

selected for cloning using TOPO-TA cloning kit 

(Invitrogen). Individual clones were picked, and 

plasmid DNA was purified and sent for 

sequencing. 

Development of teak gene specific primers 

The DNA sequences obtained from the 

cloned fragments were integrated in the alignments 

containing the sequences of other plant species. 

The intron - exon boundaries (GT-AG) were 

determined in the teak sequences. The conserved 

region of teak sequences were chosen to design 

specific primers. 

Application of markers to population analysis 

PCR-SSCP 

DNA samples were amplified from 


genomic DNA by using the specific primers in 96 

well plates in a total reaction mixture of 15 µl, 

containing 200 µM dNTPs (Promega), 2.5 mM 

MgCl 2 (Fermentas), 5 pmole of each primer, 

1× PCR buf fer with (NH 4) 2SO 4 (Fermentas), 0.3 

unit of Taq DNA polymerase (Fermentas) and 20 

ng of genomic DNA template. Amplification was 

carried out at 94°C for 3 min, followed by 35 

cycles of 45 sec at 94°C, 45 sec at 50 °C annealing 

temperature, 90 sec at 72°C, and a final extension 

at 72°C for 5 min (15 min when the fragment was 

to be used for cloning). PCR products were 

electrophoresed on 1% agarose gel using 

1xTAE buffer at 50 V for 40 min, stained with 

ethidium bromide and photographed. Four 

volumes of loading dye (98% formamide, 0.025% 

bromophenol blue, 0.025% xylene cyanol, and 

10mM NaOH) were added to the PCR products, 

then denatured at 95°C for 10 min, and 

immediately placed on ice-cold water to stabilize 

single strands. Electrophoresis was performed on 

polyacrylamide gel using non-denaturing 

conditions (Single-Strand DNA Polymorphism, 

Orita et al., 1989). 3.5 µl of the aliquots were 

loaded on a 30 cm× 40 cm × 0.4 mm 

polyacrylamide (Sequagel MD, National 

Diagnostics, U.S.A.) gels attached to glass plate 

in 0.6x TBE buffer using Hoefer SQ3 Sequencer 

(Amersham Pharmacia Biotech), run in a 4°C 

refrigerator at constant 8 watt for 16 h and 

visualized by silver staining. 

Cloning and sequencing 

From the banding patterns that appeared 

on the SSCP gel, the nuclear allelotypes were 

scored. Representative individuals were selected 

for PCR cloning. The PCR products were purified 

using QIAquick ® PCR Purification Kit (Qiagen) 

according to the manufacturer’s instructions then 

ligated into pGEM-T Vector Systems I (Promega) 

according to the manufacturer’s instructions. The 

ligations were transformed into competent cell 

(Escherichia coli, ‘DH10B’) by electroporation.

94 

The clones were grown overnight at 37 °C in LB 

(Luria-Bertani) medium agar with 100 µg/ml of 

antibiotic (ampicilin), and then screened by bluewhite 

colony selection. Each cloned copy of PCR 

amplification represented a single allele, and 

unambiguous sequence of both alleles in an 

individual could be obtained by sequencing 

multiple clones. Single colonies were selected for 

PCR amplification (reaction in each primer used 

same as corresponding PCR-SSCP method) to 

check the presence of insert and type of allele. The 

remainder of the same single colony was grown 

overnight in an incubator shaker at 37°C, 150 rpm 

in 700 µl of LB medium broth with 100 µg/ml of 

ampicilin in 1.5 microcentifuge tube, then kept in 

25 % glycerol and stored at -80°C. The PCR 

products of the clone were electrophoresed on 

polyacrylamide gel (same method was used PCR- 

RF-SSCP). Two µl of bacterial culture containing 

plasmids representing different alleles were used 

to inoculate 5 µl of LB medium broth with 100 

µg/ml of ampicilin in 15 µl tube and grown 

overnight in incubator shaker at 37°C, 150 rpm. 

The plasmids were extracted using Wizard ® Plus 

SV Minipreps DNA Purification System 

(Promega) according to manufacturer’s 

instruction, and then sent for sequencing at 

Macrogen, Inc. (Seoul, Korea). 

RESULTS 

Development of consensus primer 

73 DNA sequences coding for Cat genes 

representing 38 plant species were obtained from 

databases. The retrieved sequences consisted of 

partial and complete cDNAs, ESTs, and genomic 

DNA fragments. The sequence alignment allowed 

the design of a primer set of general usability 

among plants. Comparison of expressed and 

genomic DNA sequences indicated that the intronexon 

boundaries were not 100 % conserved among 

loci and plant species. 

The positions of the consensus primers 


were chosen in exon 3 and exon 4. The primer 

sequences were 5-GGT TTC TTT GAR GTY ACN 

CAY GA-3 for forward and 5-TG ATG AGC ACA 

YTT NGG NGC RTT-3 for reverse primer (with 

R = A or G, Y = C or T, N = A, C, G or T). In 

Arabidopsis thaliana (At1g20630) these primers 

would amplify a fragment of approximately 

1100 bp. 

The PCR reactions using the CAT primer 

set amplified a fragment in all populations, 

producing a single band on agarose gel. Two 

positive clones were obtained from this primer set 

and sequenced. The length of the inserts was 

856 bp. 

Development and analysis of specific nuclear 

gene primer sets 

The initial sequence from two teak clones 

were compared to GenBank databases using 

BLAST for verification of gene identity. The 

position of the specific primer for teak were chosen 

in exon 3 and 4, spanning the intervening intron 

(based on comparison with the complete sequence 

of A. thaliana, At1g20630). The sequences of the 

teak specific primers were 5’-CGATTCTCCAC 

TGTCATCCA-3’ for forward and 5’-GGAA 

GTTGTTTCCCACCAAA-3’ for reverse primer. 

The specific primers of Cat gene 

successfully amplified a fragment from all teak 

populations. On agarose gel, a single band of 

approximate 400 bp was visible. Different alleles 

could not be distinguished (Figure 1). However, 

using SSCP, several alleles could be identified in 

all populations (Table 2). Different alleles from 

the Potharam population are shown in Figure 2. 

Using sub-optimal annealing temperature during 

the PCR, a second locus was revealed on SSCP. 

Therefore, PCR amplification at the highest 

possible annealing temperature (52°C) was 

necessary. After cloned fragments were sent for 

sequencing, the size of the DNA amplified 

fragments varied between 388 to 413 bp.

DISCUSSION 

Development of primer set 

The specific teak catalase primer set 

could successfully amplify a DNA fragment from 

the nuclear DNA, and produced a single band on 

agarose gel, but no differentiation of allele was 

observed. However different alleles could be 

identified in each population when electrophoresis 

was done on polyacrylamide gel using SSCP 

conditions. The sequences obtained from cloned 

fragments indicated that the target gene was 

amplified. A similar approach has been used to 

amplify specific DNA fragments from other plants 

such as, Musa spp. Xylia xylocarpa and 

Andrographis paniculata (data not shown). These 

results demonstrated that sequences from other 

species could provide enough information to 

design primers that amplified low copy number 

nuclear genes from various plant taxa. Therefore 

these primers could be called “universal primers” 

and useful for population studies in other plant 

species as well. Because each amplification 

product should include an intron, the differences 


in size among the bands could, in general, be 

attributable to variation in intron length among 

duplicated loci. Once a particular band has been 

sequenced, it is possible to design new primers 

that are specific to that band in the teak genome. 

Single amplification products can then be easily 

screened for variation by SSCP direct sequencing. 

Because intron sequences evolved at a 

much higher rate than exon sequences, it was 

expected that nuclear gene based markers would 

show higher polymorphism than protein based 

Table 2 Number of alleles from 7 populations 

amplified from the Cat locus. 

Population Number of alleles 

Potharam 4 

SaiYoke 3 

MaeMoei 4 

ChiangDao 3 

PratooPha 3 

SakYai 3 

Srisatchanalai 4 

All populations 5 

Figure 1 PCR products amplified from Potharam teak samples using the teak specific CAT primer set. 

Lane M is a size standard (λ/Hind III+EcoRI), the approximate size of the fragments is 400 

bp.

96 

isozymes. The copy number for nuclear 

biosynthetic genes is generally low and the rate of 

mutation is higher than for chloroplast (Fineschi 

et al., 2004) and mitochondrial genes, therefore 

they can be used to quantify diversity and study 

evolutionary processes, population genetics, 

outcrossing and mating patterns in teak tree or 

other plants. 

Application of specific primer to population 

genetic analysis 

The obtained sequences confirmed 

variation within the intron, while the exons were 

conserved (Mason-Gamer et al., 1998; Ishii and 

McCouch 2000; Natari et al., 2003). 

The Cat locus showed polymorphism in 

all populations examined. The number of alleles 

that could be detected using this approach was 

markedly higher than the number of alleles that 

were detected in a previous report using isozyme 

assays. Kjaer et al. (1996) reported only 2 to 4 

alleles at 10 isozyme loci for trees from a 

provenance trial containing populations from 

India, Indonesia and Thailand. 

Initial analysis of the SSCP banding 

patterns obtained could be problematic since it was 


not known in advance how many bands could be 

expected and how the different alleles would 

behave during the electrophoresis. Simpler 

banding patterns were assumed to represent 

homozygous individuals while banding patterns 

that consisted of combinations of two different 

simple banding patterns were assumed to represent 

heterozygous individuals (Figure 2). In some 

samples obtained from an international 

teak provenance trial the number of different 

banding patterns observed was too large for a 

straightforward analysis without testing for 

segregation in progenies (data not shown). Further 

studies, including analysis of segregation in 

progenies, will be needed to clarify the allelic 

relationships among the observed SSCP banding 

patterns. 

For sequencing of the different alleles, 

several problematic sequences were obtained from 

cloned fragments, resulting from Taq polymerase 

errors and recombination (Cronn et al., 2002) 

between different alleles from a heterozygous 

individual during the PCR amplification step. 

Therefore, multiple clones were sent for 

sequencing with at least 2 clones from 

homozygous trees and four clones from 

Figure 2 SSCP polyacrylamide gel of the Cat marker amplified from individual teak trees from the 

Potharam population. Different alleles are assigned. Note that without progeny testing the 

different types of banding pattern observed make allelic inferences tentative.

heterozygous trees to confirm the sequence 

identity of each allele. Palumbi and Baker (1994) 

estimated the frequency of Taq polymerase error 

rate, at ~ 1 transition substitution per 1000 bases, 

by sequencing multiple clones in heterozygotes 

and comparing several individual sequences from 

the same allele. To avoid the problem of DNA 

polymerase errors retained in individual clones, 

PCR products could be sent for directly 

sequencing. 

CONCLUSION 

Although no sequence information was 

available for the particular target species, enough 

information on related species was stored in public 

databases allowed to design PCR amplification 

primers. The primer set of Cat was designed from 

consensus sequence regions in other plant species 

and could successfully amplified a fragment of the 

teak genome. The obtained sequences could be 

used to design specific primers that amplify a 

single locus from the teak genome. A high level 

of polymorphisms was observed when analyzed 

using the SSCP separation method. The different 

alleles in each population could be distinguished 

and counted. Therefore, DNA based markers 

developed from low copy nuclear genes have 

potential for use in population genetic studies in 

teak and other plant species. 


We would like to thank European 

Commission DG XII Research (project number: 

ICA4-CT-2001-10094) for financial support and 

the Center for Agricultural Biotechnology, 

Kasetsart University for providing a research 

assistantship. We would like to thank the National 

Parks, Wildlife and Plant Conservation 

Department for granting permission to collect 

samples in some National Parks. We acknowledge 

the assistance of Mr. Jertpong Makaramani, Dr. 


Somran Sutdee, and the staffs at the National 

Parks, Mr. Prayart Chaichon from Pratoo Phaa, 

Lampang province, and the villagers from 

WatBoht, Potharam, Ratchaburi province for 

assistance during the collection of plant materials. 


Cronn, R., M. Cedroni, T. Haselkorn, C. Grover 

and J. F. Wendel. 2002. PCR-mediated 

recombination in amplification products 

derived from polyploid cotton. Theor. Appl. 

Genet. 104: 482-489. 

De’ Ath, C. 1992. A history of timber exports from 

Thailand with emphasis on the 1870-1937 

period. Nat. Hist. Bull. Siam Soc. 40: 49-65. 

Doyle, J.J. and J.L. Doyle. 1989. Isolation of plant 

DNA from fresh tissue. Focus 12: 13-15. 

Fineschi, S., S. Cozzolino, M. Migliaccio and G. 

G. Vendramin. 2004. Genetic variation of relic 

tree Species: the Case of Mediterranean 

Zelkova abelicea (Lam.) Boisser and Z. sicula 

Di Pasquale, Garfì and Quézel (Ulmaceae). 

Forest Ecology and Management 197: 273- 

278. 

Graudal, L., E.D. Kjaer, V. Suangtho, P. Saardavut 

and A. Kaosa-ard. 1999. Technical Note 52, 

DANIDA Forest Seed Center, Humlebaek. 

36. 

Ishii, T. and S.R. McCouch. 2000. Microsatellite 

and microsynteny in the chloroplast genome 

of Oryza and eight other gramineae species. 

Theor. Appl. Genet. 100: 1257-1266. Kjaer, 

E.D., H.R. Siegismund and V. Suangtho. 1996. 

A Multivariate study on genetic variation in 

teak (Tectona (L)). Silv. Gen. 45(5-6): 361- 

368 

Marson-Gamer, R.J., C. F. Weil and E. A. Kellogg. 

1998. Granule-Bound starch synthase: 

structure, function, and phylogenetic utility. 

Mol. Biol. Evol. 15(12): 1658-1673. 

Natari, L., T. Giordani and A. Cavallini. 2003. 

Sequence variability of a dehydrin gene within

98 

Helianthus annuus. Theor. Appl. Genet. 106: 

811-818. 

Orita, M., H. Y. Suzuki, H. Kanazawa, T. Sekiya 

and K. Hayashi. 1989. Rapid and sensitive 

Detection of point mutations and DNA 

polymorphisms using the polymerase chain 

reaction. Genomics 5: 874-879. 


Palumbi, S. R. and C.S. Baker. 1994. Contrasting 

population structure from nuclear intron 

sequences and mtDNA of Humpback whales. 

Mol. Biol. Evol. 11(3): 426-435. 

Strand, A.E., J. Leebens-Mack and B.G. Milligan. 

1997. Nuclear DNA-based markers for plant 

evolutionary biology. Mol. Ecol. 6: 113-118.


Effects of Na + , K + and Ca 2+ Accumulation on the Expression 

of Ca 2+ -ATPase Gene in Rice KDML 105 

Wunrada Surach 1 , Mingkwan Mingmuang 1 and Amara Thongpan 2 * 

ABSTRACT 

Having been exposed to 150 mM NaCl supplemented in the growth medium, the roots of 

KDML105 started to accumulate Na + in three hours but the leaves showed the amount of Na + change 

after one day. Addition of either 0.4 mM or 10 mM of CaCl 2 to the 150 mM NaCl supplemented 

medium, on the other hand, caused a small drop in Na + accumulation in the leaves and roots on day 8. 

K + , however, significantly decreased in both leaves and roots as the result of NaCl in the medium. As 

for Ca 2+ deposition in KDML105, NaCl supplemented alone in the medium had little effect but the 

combined NaCl and CaCl 2 caused the rise of Ca 2+ in both leaves and roots. 

The expression of Ca 2+ -ATPase gene in the leaves of KDML105 seemed to be directly activated 

by the exposure to Na + . The combined effect of Na + and Ca 2+ also showed the drastic change in Ca 2+ - 

ATPase expression, and rice, therefore, try to maintain Ca 2+ homeostasis by producing more Ca 2+ - 

ATPase mRNA especially in the leaves. Since the roots are not the main accumulation site of Na + , the 

effect of NaCl alone on the Ca 2+ -ATPase gene in the roots was not as distinctively seen. However, the 

addition of 10 mM CaCl 2 caused a large increase of Ca 2+ -ATPase gene expression in the roots, while 

both Na + and Ca 2+ accumulation seemed to affect more to the gene expression in the leaves. 

Key words: Na + , K + , Ca 2+ , rice KDML105, Ca 2+ -ATPase gene, Ca 2+ -ATPase mRNA 

INTRODUCTION 

Aromatic rice “Khao Dok Mali 105” 

(KDML105) is one of the most popular Thai rice, 

having high demand both for local consumption 

and international market. To have this rice cultivar 

widely propagated in all types of soil, i.e., high 

salinity, or under stress condition of drought, 

improvement has been continuously made on 

KDML105 through selection (BSU, 2001). In 

saline soil, NaCl can disturb ion homeostasis 

causing the imbalance of ions, i.e., Na + , K + and 

Ca 2+ which results in hyperosmotic stress and ion 

toxicity. So, plants have to acclimatize to survive 

in these conditions. One mechanism of 

acclimatization is to adjust osmotic pressure by 

ion compartmentation (Niu et al., 1995) where 

Ca 2+ -ATPase is vitally needed. 

In determining the levels of Na + , K + and 

Ca 2+ accumulated in different parts of KDML105 

rice and how each type of ion could cause the 

change in Ca 2+ -ATPase expression we would be 

able to understand the mechanism of salt tolerant 

rice and enable us to make a better plan to improve 

this rice cultivar. 

1 Department of General Science, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

2 Department of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Cornesponding auther : e-mail.fsciart@ku.ac.th 


100 



KDML 105 rice seeds (Oryza sativa L.) 

from Pathumthani Rice Research Center were 

germinated in pots containing sandy soil 

impregnated with Hoaglands’ solution (Hoagland 

and Arnon, 1950) at room temperature for two 

weeks. Plantlets were transferred and grown 

hydroponically for a week in Hoagland’s solution. 

The leaves were collected for RNA extraction. 

Salt induction was carried out by adding 

150 mM NaCl alone or having NaCl combined 

with either 0.4 mM or 10 mM CaCl 2 to Hoaglands’ 

solution. Seedlings were maintained in this 

hydroponic culture for 8 days. 

Measurement of Na + , K + and Ca2+ Plants were rinsed with distilled water 

to remove surface ions, then the shoots and the 

roots were separated, and oven dried at 80°C for 

48 hours. The samples were chopped and digested 

in glass tubes containing 5 ml of concentrated 

HNO3, H2SO4 and HClO4 (5:2:1) and placed in a 

heat block at 180°C. After complete digestion, the 

sample volumes were adjusted to 50 ml with 

distilled water. The amounts of Na + , K + and Ca2+ were determined using an atomic absorption 

spectrophotometer. Duncan’s multiple range test 

was employed for test of significance. 

Isolation of total RNA 

Total RNA was extracted from the 

KDML105 leaves using guanidinium 

isothiocyanate according to the method described 

by Chomczynski and Sacchi (1987) with some 

modifications. Approximately one gram of the 

leaves was ground in liquid nitrogen and 

suspended in a 10 ml buffer solution containing 4 

M guanidinium isothiocyanate, 25 mM sodium 

citrate, 0.5% sarkosyl, and 0.1 M βmercaptoethanol. 

One millilitre of 2 M sodium 

acetate (pH 4.0), 10 ml of phenol and 2 ml of 


chloroform: isoamyl alcohol (24:1 v/v) were added 

and vortexed, followed by centrifugation at 3,500 

g for 10 min. The supernatant was collected and 

mixed with an equal volume of cold isopropanol. 

DNA and RNA were allowed to precipitate at –20° 

C for 1 h, followed by centrifugation at 11,000 g 

for 10 min. The pellet was washed once using 1 

ml of cold 70% ethanol and centrifuged at 5,000 g 

for 5 min. It was then dissolved in 500 µl of 0.1% 

DEPC-treated water and added with 500 µl of 6 

M LiCl. The solution was left at 4°C overnight, 

followed by centrifugation at 11,000 g for 10 min. 

The precipitated RNA pellet was washed again in 

cold 70% ethanol as described above and dissolved 

in 0.1% DEPC-treated water. The purity and 

concentration of total RNA was determined using 

spectrophotometer at 260 and 280 nm. RNA was 

stored at –80°C until use. 

cDNA synthesis 

First strand cDNA was synthesized from 

the total RNA by the process of reverse 

transcription. Approximately 5 µg of RNA was 

mixed with 0.5 µg of oligo (dT) 12-18 primer and 

adjusted to the volume of 12 µl with distilled water. 

The mixture was heated at 70°C for 10 min and 

quickly cooled on ice. Superscript II RNaseH - kit 

solution (Invitrogen) containing 4 µl of 5X first 

strand buffer, 2 µl of 0.1 M dithiothreitol and 1 ml 

of reverse transcriptase (200 U) was added to the 

sample together with 1 µl of 10 mM dNTP. Then 

the mixture was incubated at 42°C for 50 min. The 

reaction was terminated by incubation at 70°C for 

15 min. It was kept at –20°C until use. 

Determination of the levels of gene expression 

using real-time PCR 

Real-time PCR was performed using an 

ABI 7700 Sequence Detection System. The 

optimization of the real-time PCR reaction was 

performed according to the manufacturer’s 

instructions (PE Applied Biosystems). The PCR 

conditions were standardized using core reagent

protocol. All reagents were provided in the core 

reagent kit, including AmpliTaq-Gold polymerase 

(PE Applied Biosystems). After optimization, 

nucleotide primers were used for the detection and 

quantification of Ca 2+ -ATPase. 

Gene expression of Ca 2+ -ATPase was 

analyzed by Fluorescent PCR employing FAM 

labeled oligonucleotide probe. The primers and the 

probe were designed using ABI PRISM TM Primer 

Express TM (Applied Biosystems). A forward 

primer (5'-GCTGCTGGAATACGTG TTATGG- 

3') and a reverse primer (5'-TCACGGCATATCG 

CCTCTGCTGTTT-3') were applied to produce a 

single 110 bp PCR product. A probe (5'- 

TTAGTTCAAAGAGCTT CACAGGGA-3'), 

labeled with FAM (6-carboxylfluorocien) reporter 

at 5' end and TAMRA (6-carboxytetramethylrodamine) 

quencher dye at 3' end, was used 

to produce fluorescent signal. Real-time PCR 

amplification was carried out in 5 µl of 10X PCR 

buffer, 6 mM MgCl 2, 0.5 mM dNTPs, 20 ?M of 

each primer for Ca 2+ -ATPase forward primer and 

Ca 2+ -ATPase reverse primer, 1 µl of cDNA sample, 

1U AmpliTaq-Gold polymerase. Real-time PCR 

conditions consisted of an initial denaturation step 

of 50°C for 2 min, 95°C for 10 min, followed by 

40 cycles of 95°C for 15 s, and 60°C for 1 min. 

Genomic DNA was used as a reference. 

Experiments were performed in triplicate. Ca 2+ - 

ATPase transcripts were calculated by comparing 

to the reference generated by the controlled DNA. 

Ca 2+ -ATPase expression was determined from the 

levels of mRNA using real-time PCR in the form 

of normalized genome equivalent (NGE). 


Na + accumulation 

After transferring the three-week-old 

KDML105 to grow in the Hoaglands’ medium 

containing 150 mM NaCl and supplemented with 

either 0, 0.4, or 10 mM CaCl 2, the amounts of Na + 

in the leaves and the roots were determined at 3 


hours, 1 day, 2 days, 5 days, and 8 days periods. 

The results showed that Na + content in the leaves 

at the first three hours remained the same in all 

treatments (Figure 1A). The change in Na + 

accumulation was observed after 1-day growth and 

gradually increased from 0.8% to as high as 4.5% 

on day 8. CaCl 2 seemed to have positive effect on 

Na + accumulation in the leaves comparing to 

having only NaCl in the medium. In the roots, 

however, Na + accumulation profile was quite 

different from that of the leaves (Figure 1B). There 

was a jump in Na + level right at the first three hours 

of salt exposure. All the treatments with either 150 

mM NaCl alone or having it combined with 

different amount of CaCl 2 made the Na + in the 

roots changed from 0.5% to 0.8%, but the 

increment of Na + at the later days of exposure was 

not drastically altered. The highest percentage of 

Na + accumulation was found on day 8 at 1.2%. 

The slight change of Na + in the roots could be the 

results of continuous transport of Na + to the leaves 

where the real accumulation was taking place. This 

type of high Na + accumulation in the leaves is also 

seen in cotton seedling treated with 150 mM NaCl 

(Binzel et al., 1987), in rice treated with 171 mM 

NaCl (Garcia et al., 1997), and in grass treated 

with 0-600 mM NaCl (Marcum, 1999). Although 

it was clearly seen that both NaCl and CaCl 2 

caused the high Na + accumulation in the leaves of 

KDML105 in the beginning (up to day 5), but the 

longer exposure time of CaCl 2 with the 150 mM 

NaCl in the medium (day 8) seemed to lower the 

Na + accumulation than having 150 mM NaCl 

alone (Figure 1). This might be due to the 

inhibitory action of Ca 2+ to the influx of Na + 

through cation channel and at the same time 

inducing the reaction of SOS (salt-overlysensitive) 

protein system to eliminate extra Na + 

from the cells as well as blocking the influx of 

Na + through K + channel (Yokoi et al., 2002), which 

is also seen in citrus fruit, cotton seedling, and 

wheat (Cramer et al., 1985; Ben-Hayyim et al., 

1987; Davenport et al., 1997).

102 

The amount of Na + (%) 

6 

5 

4 

3 

2 

1 

0 

3h 1d 2d 5d 8d 

K + accumulation 

In the control medium having neither 

NaCl nor CaCl 2, the K + accumulation in the leaves 

and the roots of KDML105 rice were at steady 

levels of 2.6% and 0.6%, respectively. However, 

having been exposed to either 150 mM NaCl alone 

or the combination of 150 mM NaCl with either 

0.4 mM or 10 mM CaCl 2, the drop in K + 

accumulation were both observed at the low levels 

of 1.2% in the leaves and 0.4% in the roots (Figure 

2A and B). There was only a slight change from 

the beginning of 3-hour exposure to the salt until 

at the end of 8-day period. The decrease of K + 

might result from the competitive effect of Na + in 

the medium with the flow of K + through inwardrectifying 

K + channel (Niu et al., 1995). At the 

same time, the maintenance of K + in rice at a 

certain level is necessary because of its several 

important roles in plants, i.e., enzymes activation, 

stomata opening, respiration and photosynthesis, 

homeostasis, as well as the growth of leaves and 

roots (Osothsapa, 2000). The additive effect of 

CaCl 2 and NaCl on K + accumulation was observed 

in both leaves and roots of KDML105 as also 

reported in cotton seedling (Cramer et al., 1985) 

but not at distinctive level 


A 1.4 

B 

The amount of Na + (%) 

1.2 

0.8 

0.6 

0.4 

0.2 

control 0 mM CaCl2 0.4 mM CaCl2 10 mM CaCl2 1 

0 

3h 1d 2d 5d 8d 

Figure 1 The levels of Na + in (A) the leaves and (B) the roots of KDML105 grown in Hoaglands’ 

solution supplemented with 150 mM NaCl and different concentrations of CaCl 2 (0, 0.4 and 

10 mM CaCl 2). 

Ca 2+ accumulation 

The Ca 2+ accumulation in the leaves of 

KDML105 was kept at rather steady levels from 

the beginning to the 8-day exposure of rice in the 

medium having either 150 mM NaCl alone or 

combining with different concentrations of CaCl 2, 

which was in the ranges of 0.8-1.2% (Figure 3A). 

As for the roots, having only 150 mM 

NaCl or the combination with 0.4 mM CaCl 2 made 

the roots less susceptible to Ca 2+ comparing to the 

control after the 3-hour exposure to the salt but 

not at later stages (day 5 and 8). However, the 

higher concentration of CaCl 2 (10 mM) gave an 

increase amount of Ca 2+ to the ranges of 0.4% to 

0.5% (Figure 3B). This result also agreed with the 

pattern of Ca 2+ accumulation found in the leaves 

and the roots of cotton (Cramer et al., 1985). The 

importance of having Ca 2+ at the same level in 

plant is as expected since Ca 2+ plays a vital role in 

cellular homeostasis and signal transduction. The 

changes in Ca 2+ content not only cause the change 

in membrane permeability, it also triggers the 

expression of Ca 2+ -ATPase and hence protein 

synthesis (Osothsapa, 2000).

The amount of Ca 2+ (%) 

The amount of K + (%) 

3 

2.5 

2 

1.5 

1 

0.5 

1.4 

1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

0 

3h 1d 2d 5d 8d 

3h 1d 2d 5d 8d 

Changes in Ca 2+ -ATPase expression 

Ca 2+ -ATPase expression was determined 

from the levels of mRNA using real-time 

PCR in the form of normalized genome equivalent 

(NGE). Ca 2+ -ATPase expression in the leaves of 

KDML105 was found to change at drastic 


A 

The amount of K + (%) 

0.8 

0.6 

0.4 

0.2 

0 

3h 1d 2d 5d 8d 

control 0 mM CaCl2 0.4 mM CaCl2 10 mM CaCl2 Figure 2 The levels of K + in (A) the leaves and (B) the roots of KDML105 grown in Hoaglands’ 



A 

The amount of Ca 2+ (%) 

0.6 

0.5 

0.4 

0.3 

0.2 

0.1 


3h 1d 2d 5d 8d 

Figure 3 The levels of Ca 2+ in (A) the leaves and (B) the roots of KDML105 grown in Hoaglands’ 



levels after 3-hour exposure to 150 mM NaCl from 

0.3 NGE to 2 NGE (7 folds), and even higher 

when 0.4 mM CaCl 2 was added (6 NGE, 19 folds). 

After one-day exposure to the salt, the levels of 

expression were all dropped to only 1.2 NGE 

(4 folds) but went up again on day 2 (2.2 - 5.6 

B 

B

104 

NGE, 7-16 folds). Finally, all seemed to be at 

steady levels from day 2 to day 8, except those 

having 0.4 mM CaCl 2 treatment which leveled 

off on day 5 and day 8 (Figure 4A). The high 

fluctuation of Ca 2+ -ATPase expression might 

result from the activating effect of NaCl to increase 

cytoplasmic level of Ca 2+ (Niu et al., 1995) which 

in turn induces the signal transduction system in 

response to salt stress (Geisler et al., 2000). 

However, the activating effect was temporary and 

Ca 2+ level needed to be controlled at a certain level 

to maintain the cellular functions. The high 

synthesis of Ca 2+ -ATPase, therefore, is required 

to eliminate extra amount of Ca 2+ in cytoplasm. 

CaCl 2 was also found to have accumulating 

effect on Ca 2+ -ATPase expression as well as 

helping decrease the Na + accumulation in the cell 

at later period. These results agreed with the 

report of Wimmers et al. (1992) on the 310% 

increase in Ca 2+ -ATPase mRNA of tomato leaves 

after one-day exposure to 50 mM NaCl. The 

other types of this change were also reported 

in different organisms, i.e., 190% increase of 

Ca 2+ -ATPase mRNA found in Arabidopsis after 

one-day exposure to 100 mM NaCl (Geisler 

et al., 2000) and 10 folds increase of Ca 2+ -ATPase 

in amoeba after having exposed to 80 mM 

Normalized genome equivalent 

(NGE) 

7 

6 

5 

4 

3 

2 

1 

0 

3 h 1d 2d 5d 8d 


A 

NaCl for one month (Moniakis et al., 1995). 

As for Ca 2+ -ATPase expression in the 

roots, it was found that NaCl alone or combining 

with lower concentration of CaCl 2 (0.4 mM) 

had little effect on the change of Ca 2+ -ATPase 

mRNA (Figure 4B). This is contrary to the 180% 

increase of Ca 2+ -ATPase mRNA in the roots of 

tomato after having been exposed to 50 mM 

NaCl for one day (Wimmers et al., 1992). The 

less effect of NaCl in the roots of KDML105 could 

be the result of fast transportation of ions to 

different parts of the plant especially to the 

leaves where accumulation was more prominent 

and hence the distinctive change in Ca 2+ -ATPase 

mRNA. However, the addition of high 

concentration of CaCl 2 (10 mM) could cause 

the change in Ca 2+ -ATPase mRNA after one-day 

exposure to attain a very high level of 6.1 

NGE (8 folds) and dropped to 4.3 NGE (5 folds) 

on day 2 (Figure 4B). Then on day 5, Ca 2+ - 

ATPase mRNA level went to as high as 9.4 NGE 

(13 folds) before dropping to the normal 

level again on day 8. This observation could 

result from the saturated salt condition in the roots 

where its effect on Ca 2+ -ATPase expression was 

unavoidable, but it could return to the original level 

when the salt was finally eliminated. 

Normalized genome equivalent 

(NEG) 


8 

6 

4 

2 

0 

3 h 1d 2d 5d 8d 

Figure 4 Ca 2+ -ATPase in (A) the leaves and (B) the roots of KDML105 grown in Hoaglands’ solution 

supplemented with 150 mM NaCl and different concentrations of CaCl 2 (0, 0.4 and 10 mM 

CaCl 2). 

B

It was evident that the change in Ca 2+ - 

ATPase expression as seen from the mRNA level 

was directly activated by the effect of NaCl and 

CaCl 2 and Ca 2+ accumulation in cytoplasm (Niu 

et al., 1995). The change in Ca 2+ -ATPase level was 

then inevitably needed to restore the Ca 2+ 

homeostasis and made the rice survive in the high 

salt condition. 

CONCLUSION 

1. NaCl in the medium caused the 

increase of Na + in the KDML105 roots at the early 

stage of the exposure (3 hours) and kept at the 

same level throughout most of the experimental 

period except at day 8, while the Na + change in 

the leaves was observed after one-day exposure 

but continuously increased at distinctive level. 

Addition of CaCl 2 to the NaCl containing medium 

helped decreasing Na + accumulation in the leaves 

at day 8 but this effect was not distinctively seen 

in the roots. 

2. NaCl showed a negative effect on the 

K + level in both leaves and roots of KDML105 

from the early three hour after exposure to the salt. 

The change in K + over time, however, was not 

clearly seen. The early drop of K + level was mainly 

due to the competition of Na + with K + for the same 

K + channel. 

3. NaCl alone had little effect on the 

Ca 2+ levels in the leaves and the roots of 

KDML105 but having NaCl combined with CaCl 2 

made a distinctive change in Ca 2+ accumulation 

at both parts to a considerable level. However, this 

additive effect of CaCl 2 also caused the decrease 

in Na + level as a result of the inhibitory action of 

Ca 2+ to the influx of Na + to the cell. 

4. NaCl caused the drastic change in 

Ca 2+ -ATPase mRNA of the KDML105 leaves. 

This effect was even greater when superimposed 

by CaCl 2 because both types of salt contributed to 

the Ca 2+ accumulation in cytoplasm. The plant, 

therefore, tried to eliminate the surplus of Ca 2+ by 


synthesizing more Ca 2+ -ATPase even at the lower 

exposed amount of 0.4 mM CaCl 2. On the other 

hand, Ca 2+ -ATPase mRNA in the roots of 

KDML105 was affected by CaCl 2 (10 mM) but 

not by NaCl alone. Since the roots were not the 

sites for Na + accumulation per se, fast 

transportation of Na + made the roots more stable 

and less likely to have high Ca 2+ -ATPase 

expression at lower concentration of 0.4 mM of 

CaCl 2. However, the high concentration of 10 mM 

CaCl 2 was too much to be eliminated by a mere 

transportation. The roots, therefore, needed to 

synthesize more Ca 2+ -ATPase to assist in coping 

with this salt stress condition. 


This reserch was supported by the 

Rockefeller Foundation (grant # 98001), and 

Graduate School, Kasetsart University. We would 

also like to thank Pathumthani Rice Research 

Center for supplying KDML105 cultivar, and to 

Soil and Fertilization Analysis Unit, Research and 

Experimental Greenhouse for the use of ion 

analysis equipment. 


Ben-Hayyim, G., U. Kafkafi and R. Ganmore- 

Neumann. 1987. Role of internal potassium 

in maintaining growth of cultured Citrus cells 

on increasing NaCl and CaCl 2 concentrations. 

Plant Physiol. 85: 434-439. 

Binzel, M.L., P.M. Hasegawa, D. Rhodes, S. 

Handa, A.K. Handa and R.A. Bressan. 1987. 

Solute accumulation in tobacco cells adapted 

to NaCl. Plant Physiol. 84: 1408-1415. 

Bioservice Unit (BSU). 2001. Science and 

Technology with Thai Rice. NSTDA, 

Bangkok, 158 p. 

Chomczynski, P. and N. Sacchi. 1987. Single-step 

method of RNA isolation by acid guanidinium 

thiocyanate-phenol-chloroform extraction.

106 

Anal. Biochem. 162: 156-159. 

Cramer, G.R., A. L?uchli and V.S. Polito. 1985. 

Displacement of Ca 2+ by Na + from the 

plasmalemma of root cells. Plant Physiol. 79: 

207-211. 

Davenport, R.J., R.J. Reid and F.A. Smith. 1997. 

Sodium-calcium interactions in two wheat 

species differing in salinity tolerance. 

Physiologia Plantarum 99: 323-327. 

Garcia, A.B., E.J. de Almeida, S. Iyer, T. Gerats, 

M.V. Montagu and A.B. Caplan. 1997. Effects 

of osmoprotectants upon NaCl stress in rice. 

Plant Physiol. 115: 159-169. 

Geisler, M., N. Frangne, E. Gomes, E. Martinoia 

and M.G. Palmgren. 2000. The ACA4 gene of 

Arabidopsis encodes a vacuolar membrane 

calcium pump that improves salt tolerance in 

yeast. Plant Physiol. 124: 1814-1827. 

Hoagland, D.R. and I.D. Arnon. 1950. The Water- 

Culture Methods for Growing Plants 

without Soil. California Agricultural 

Experimental Station. Circular, 286 p. 


Marcum, K.B. 1999. Salinity tolerance 

mechanisms of grasses in subfamily 

Chloridoideae. Crop Sci. 39: 1153-1160. 

Moniakis, J., M.B. Coukell and A. Forer. 1995. 

Molecular cloning of an intracellular P-type 

ATPase from Dicteostelium that is upregulated 

in calcium-adapted cells. J. 

Biol.Chem. 270: 28276-28281. 

Niu, X., R.A. Bressan, P.M. Hasegawa and J.M. 

Pardo. 1995. Ion homeostasis in NaCl stress 

environments. Plant Physiol. 109: 735-742. 

Osothsapa, Y. 2000. Nutrition Element of Plant. 

Kasetsart University, Bangkok. 424 p. 

Wimmers, L.E., N.N. Ewing and B.A. Bennett. 

1992. Higher plant Ca 2+ - ATPase: Primary 

structure and regulation of mRNA abundance 

by salt. Proc. Natl. Acad. Sci. USA. 89: 9205- 

9209. 

Yokoi, S., R.A. Bressan and P.M. Hasegawa. 2002. 

Salt stress tolerance of plants. JIRCAS 

Working Rep. 25-33.


Molecular Identification of Cycas by Restriction Fragment 

Length Polymorphism (RFLP) and Random Amplified Polymorphic 

DNA (RAPD) 

Pattamon Sangin1 , Amara Thongpan2 , Anders J. Lindstrom3 and Mingkwan Mingmuang1 * 

ABSTRACT 

RAPD and RFLP were used to identify nineteen species of Cycas. Ten species of these Cycas 

namely C. chamaoensis, C. macrocarpa, C. pectinata, C. clivicola, C. pranburiensis, C. litoralis, 

C. tansachana, C. siamensis, C. nongnoochiae and C. simplicipinna are locally found in Thailand while 

the nine remaining species of C. seemannii, C. wadei, C. bougainvilleana, C. chevalieri, C. diannanensis, 

C. nathorstii, C. edentata, C. parvulus and C. micholitzii are from several countries around the world 

but collectively planted at Nong Nooch Tropical Botanical Garden. 

In the RAPD study, twenty random primers were screened to amplify the genomic DNA of 

nineteen species of Cycas. Only five primers, i.e., OPB-1, OPB-8, OPB-14, OPB-15 and OPB-17 of ten 

nucleotides long were found to give polymorphic DNA patterns. These eighty-seven bands of Cycas 

DNA at the size of 0.35 -2.5 kb could be used to indicate the differences of these Cycas. As for RFLP, 

three probes were synthesized from 5S rRNA gene, 5S rRNA repeat unit gene of C. clivicola and 18S 

rRNA gene of C. pranburiensis. The probes were hybridized with the genomic DNA of Cycas which 

had been digested with restriction enzymes BamHI, EcoRI and DraI. The phylogenetic trees were 

constructed based on their similarity index derived from DNA polymorphism of RAPD and RFLP 

separately. The RAPD data classified nineteen species of Cycas into two major groups which mostly 

corresponded to their geographic origins, i.e., one group of Thailand origin and another of other countries. 

However, the RFLP data gave a different set of grouping showing more to their morphological 

characteristics but less on their geographic origins. 

Key words: Cycas, RFLP, RAPD, phylogenetic tree, geographic origin 

INTRODUCTION 

Cycads are ancient plants with a long 

continuity line of heredity. They are classified into 

3 families, 11 genera and 250 known species 

distributed all over the world (Stevenson, 1992) 


2 Deparment of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

3 Nong Nooch Tropical Botanical Garden, Chonburi 20250, Thailand. 

* Corresponding author : email : fscimkm@ku.ac.th 

but are mainly found along the intertropical belt, 

i.e., Africa, India, Indonesia, and North Australia. 

Cycas is the only genus naturally grown in 

Southeast Asia. However, representatives of 

different species and genera of Cycad from several 

distinctive parts of the world are collectively 


108 

grown for academic purpose at Nong Nooch 

Tropical Botanical Garden, Chonburi province, 

Thailand. 

Although the species of Cycad can be 

generally identified from corralloid roots, 

girdlingleaf traces, secondary compounds 

production or even from the isozymes (Caputo et 

al., 1993), these physico-morphological properties 

are also affected by the growth conditions and 

environment which make the closely related 

species difficult to identify. At present, there are 

10 species of morphologically similar Cycas found 

in Thailand but their true genetic identity and 

evolutionary line are not yet known. To effectively 

group them and determine their evolutionary 

relationships, biomolecular technique is needed. 

Random Amplified Polymorphic DNA 

(RAPD) technique is a powerful tool to detect 

DNA polymorphism and classify closely related 

species in most plants (Williams et al., 1990), and 

it is considered a fast and easy approach to solve 

the obscured identification problem. Many kinds 

of plant have been fingerprinted using RAPD 

markers (Dettori and Palombi, 2000) to evaluate 

their phylogenetic relationship, genetic variability 

and genetic relationships (Lerceteau et al., 1997; 

Ahmad, 1999). On the other hand, Restriction 

Fragment Length Polymorphism (RFLP) 

technique makes the use of specific probes from a 

cDNA or genomic DNA library of the investigated 

species for the same purpose. This technique has 

been widely used in studying genetic variation and 

phylogenetic relationships among populations, 

species and varieties (Lerceteau et al., 1997; 

Garcia-Mas et al., 2000; Sun et al., 2001). Since 

ribosomal DNA (rDNA) are organized in tandem 

repeat units, they are often used as probes for 

RFLP. The copy number of these units varies from 

a few hundred to thousands and are different from 

species to species (Rogers and Bendich, 1987). 

This paper reported on the use of RAPD 

and RFLP markers to determine the relationship 

among Cycas species and to assess the 


organization level of genetic diversity through 

phylogenetic tree. 



Young cycad leaves of nineteen Cycas 

species were collected from the plants grown at 

Nong Nooch Tropical Botanical Garden, 

Chonburi, Thailand. These are C. chamaoensis, 

C. macrocarpa, C. pectinata, C. clivicola, 

C. pranburiensis, C. litoralis, C. tansachana, 

C. siamensis, C. nongnoochiae, C. simplicipinna, 

C. seemannii, C. wadei, C. bougainvilleana, 

C. chevalieri, C. diannanensis, C. nathorstii, 

C. edentata, C. parvulus and C. micholitzii. 

DNA extraction 

Genomic DNA was extracted using 

hexadecetyltrimethyl ammonium bromide 

(CTAB) method described by Doyle and Doyle 

(1990) with some modification. About 0.3 g of 

young leaves was ground to fine powder in liquid 

nitrogen. One milliliter of preheated (65°C) 2X 

CTAB isolation buffer (2% CTAB, 1.4 M NaCl, 

20 mM EDTA, 100 mM Tris-HCl, pH 8.0 and 10 

mM 2-mercaptoethanol) was added to the sample. 

The homogenate was incubated at 65°C for 1 h 

and then extraction was made using one equal 

volume of chloroform:isoamyl alcohol (24:1). The 

mixture was centrifuged at 10,000 g for 10 min at 

room temperature. The aqueous phase was 

collected and mixed with 1/5 volume of 5X CTAB 

(5% CTAB and 0.7 M NaCl) and 2/3 volume of 

isopropanol. The nucleic acid pellet was air-dried 

and resuspended in 100 mM TE buffer (10 mM 

Tris, 1mM EDTA, pH 7.0). RNase A was added 

to the sample at the final concentration of 10 ng/ 

ml. After incubating at 37°C for 30 min, the 

sample was extracted with phenol:chloroform 

(1:1). 70% ethanol was added at 2X volume of 

DNA to make it precipitated. Finally, the DNA 

pellet was air-dried and resuspensed in 30 ml TE

uffer. DNA pattern and concentration was 

detected on a 1% agarose gel and by UV 

spectrophotometer. 

RAPD 

The protocol for RAPD analysis was 

adapted from that of Williams et al. (1990). The 

volume of the final reaction (25 µl) consisted of 1 

X buffer (10 mM Tris-HCl, pH 8.0 and 50 mM 

KCl), 3.0 mM MgCl2, 1.25 U Taq DNA 

polymerase, 200 µM dNTP, 10 mM random primer 

(Operon), 25-100 ng of genomic DNA. 

Amplifications were made in a Perkin Elmer 9600 

thermocycler with an initial denaturing step of 1 

min at 94 °C, followed by 45 cycles of 1 min at 

94°C, 1 min at 36°C, 2 min at 72°C and a final 

extension of 5 min at 72°C. PCR products were 

subjected to 1% agarose gel electrophoresis run 

at 90 V and DNA bands were visualized by 

ethidium bromide staining. 

RFLP 

One gram of genomic DNA was cut with 

the restriction enzymes DraI, BamHI, EcoRI, 

HindIII or stuI and then subjected to 1% agarose 

gel electrophoresis. The gels were blotted onto a 

positively charged nylon membrane (Boehringer 

Mannheim) by vacuum blotter. The probes were 

amplified by PCR using 18S rRNA, 5S rRNA and 

5S rRNA repeat unit primer and PCR products 

were labeled with digoxigenin according to the 

protocol of Dig High prime DNA Labeling and 

Detection Starter Kit II (Roche). Hybridization 

was detected by enhanced chemiluminescence on 

Kodax X-ray film with 0.5-2 h exposure time. 

Data analysis 

DNA fragments were scored as presence 

(1) or absence (0) for each primer or restriction 

enzyme used. These scores were used to calculate 

their genetic similarity according to Nei and Li 

(1979), using NTSYS-pc1.80 from which the 

phenograms were constructed using a UPGMA. 



Primer selection and levels of polymorphism 

Twenty primers were screened for the 

genomic DNA amplification of nineteen Cycas 

species. Only 5 primers (Table 1) were found to 

give polymorphic DNA patterns. The total 

numbers of 87 bands with the fragments ranging 

from 0.35- 2.5 kb are shown in Figures 1 and 2. 

Amplification with primers OPB-1, OPB-8 and 

OPB-17 revealed unique bands of 0.8 kb, 0.6 kb 

and 0.7 kb, respectively. These bands could be 

used as genetic markers to identify the 

relationships among Cycas since they were 

specific to certain groups of Cycas (Nicolosi et 

al., 2000). 

RAPD analysis 

RAPD data were subjected to UPGMA 

and NTSYS-pc (Version 1.8). The similarity index 

showed that the relationships among all nineteen 

species fell in the range of 0.816-0.516 (Table 2). 

Maximum similarity was between C. edentata and 

C. litoralis (83.9%), while the least similarity were 

found between C. wadei and C. pranburiensis, 

C. tansachana and C. bougainvilleana, C. wadei 

and C. bougainvilleana (50.6 %). The distribution 

of species within the clusters showed apparent 

relation with geographical origin. The dendrogram 

(Figure 3) classified the 19 species into two major 

clusters, A and B. Cluster A contained all Cycas 

of Thailand geographical origin while cluster B 

contained those from other countries. 

Table 1 Nucleotide sequences of random 

oligonucleotide primers which showed 

polymorphism. 

Primer no. Sequence 

OPB-1 5′ GTTTCGCTCC 3′ 

OPB-8 5′ GTCCACACGG 3′ 

OPB-14 5′ TCCGCTCTGG 3′ 

OPB-15 5′ GGAGGCTGTT 3′ 

OPB-17 5′ AGGAACGAAG 3′

110 

Cluster A was further separated into two 

groups, A I and A II. A I was subdivided into two 

groups, I and II. Group I consisted of 

C. chamaoensis, C. macrocarpa, C. pectinata, 

C. clivicola and C. siamensis. It is interesting to 

find that all Cycas of group I are of Thailand origin 

especially C. chamaoensis, C. macrocarpa, 

C. clivicola and C. siamensis were found only in 

Thailand but C. pectinata found also in 

China, Thailand and Vietnam was somewhat 


separated as indicated by isozyme (Yang and 

Meerow, 1996). Group II contained C. micholitzii, 

C. simplicipinna, C. edentata, C. litoralis and 

C. pranburiensis. The similarity index suggested 

that C. micholitzii wass closely related to 

C. simplicipinna which also belonged to the same 

Stangerioides section based on their morphology 

(Pu and Chiu, 1990; Stevenson, 1992) (Table 5). 

The similarity index showed the close relationship 

of C. edentata and C. litoralis and both were also 

Figure 1 RAPD amplification of 19 Cycas species using the primers (A) OPB-1 showing a common 

band of 0.8 kb, (B) OPB-8 showing a common band of 0.6 kb, (C) OPB-14 having no common 

band , M: 100 bp DNA size marker, lane 1 C. chamaoensis, lane 2 C. macrocarpa, lane 3 C. 

clivicola, lane 4 C. micholitzii, lane 5 C. simplicipinna, lane 6 C. pectinata, lane 7 C. 

pranburiensis, lane 8 C. edentata, lane 9 C. litoralis, lane 10 C. tansachana, lane 11 

C. siamensis, lane 12 C. nongnoochiae, lane 13 C. wadei, lane 14 C. seemannii, lane 15 

C. bougainvilleana, lane 16 C. parvulus, lane 17 C. chevalieri, lane 18 C. nathorstii, lane 19 

C. diannanensis. (B) and (C) indicated the similar DNA pattern found between 

C. edentata (lane 8) and C. litoralis (lane 9) as well as those of C. parvulus (lane 16) and 

C. diannanensis. (lane 19).

in Rumphiae subsection. Group A II comprised 

of C. tansachana, C. nongnoochiae C. seemannii 

and C. wadei. The similarity index showed 

the closer phylogenetic relationships of 

C. nongnoochiae and C. seemannii than that 

of C. tansachana. However, morphological 

identification of C. nongnoochiae and 


C. tansachana indicated that these 2 species 

belonged to the same Indosinenses section and also 

of Thailand origin (Table 5). It should be noted 

here that all ten Cycas species of Thailand origin 

were classified into cluster A using RAPD 

technique and also supported by their 

morphological characteristics in the close 

Figure 2 RAPD amplification of 19 Cycas species using the primers (A) OPB-15 , (B) OPB-17 

showing a common band of 0.7 kb, M: 100 bp DNA size marker, lane 1 C. chamaoensis, lane 

2 C. macrocarpa, lane 3 C. clivicola, lane 4 C. micholitzii, lane 5 C. simplicipinna, lane 6 C. 

pectinata, lane 7 C. pranburiensis, lane 8 C. edentata, lane 9 C. litoralis, lane 10 

C. tansachana, lane 11 C. siamensis, lane 12 C. nongnoochiae, lane 13 C. wadei, lane 14 

C. seemannii, lane 15 C. bougainvilleana, lane 16 C. parvulus, lane 17 C. chevalieri, lane 18 

C. nathorstii, lane 19 C. diannanensis. (A) and (B) indicated the similar DNA pattern found 

between C. parvulus (lane 16) and C. diannanensis (lane 19).

112 

Table 2 Similarity index of 19 Cycas species as identified by RAPD. 

Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 

1 1.000 

2 0.816 1.000 

3 0.747 0.793 1.000 

4 0.713 0.759 0.782 1.000 

5 0.667 0.667 0.667 0.770 1.000 

6 0.770 0.747 0.724 0.690 0.690 1.000 

7 0.724 0.678 0.701 0.713 0.644 0.701 1.000 


8 0.644 0.667 0.690 0.701 0.678 0.621 0.598 1.000 

9 0.713 0.667 0.713 0.770 0.724 0.644 0.667 0.839 1.000 

10 0.644 0.667 0.690 0.678 0.563 0.621 0.552 0.701 0.724 1.000 

11 0.747 0.724 0.793 0.736 0.690 0.678 0.632 0.736 0.736 0.713 1.000 

12 0.667 0.644 0.690 0.724 0.655 0.644 0.575 0.724 0.701 0.678 0.690 1.000 

13 0.621 0.598 0.598 0.586 0.609 0.552 0.506 0.678 0.655 0.678 0.667 0.701 1.000 

14 0.655 0.655 0.632 0.644 0.575 0.563 0.609 0.690 0.690 0.690 0.632 0.713 0.644 1.000 

15 0.609 0.586 0.586 0.667 0.598 0.540 0.609 0.644 0.644 0.506 0.609 0.644 0.506 0.586 1.000 

16 0.621 0.575 0.598 0.770 0.724 0.621 0.575 0.701 0.747 0.632 0.598 0.724 0.609 0.644 0.736 1.000 

17 0.667 0.621 0.575 0.632 0.563 0.575 0.621 0.563 0.632 0.540 0.598 0.655 0.655 0.621 0.736 0.701 1.000 

18 0.701 0.655 0.609 0.690 0.552 0.609 0.609 0.598 0.644 0.575 0.655 0.667 0.598 0.540 0.793 0.644 0.736 1.000 

19 0.621 0.575 0.575 0.701 0.632 0.575 0.552 0.609 0.655 0.655 0.621 0.701 0.586 0.667 0.667 0.816 0.701 0.667 1.000 

Note: The names of the nineteen Cycas species are shown in Table 5.

elationship of sub-groupings. 

Cluster B consisted of C. bougainvilleana, 

C. nathorstii, C. chevalieri, C. parvulus 

and C. diannanensis. The similarity index 

indicated that C. bougainvilleana was closely 

related to C. nathorstii and both are in Cycas 

section (Table 5). The next related one in this group 

was C. chevalieri. As for C. parvulus and 

C. diannanensis, they were identified by this 

method as closely related to one another as also 

supported by Wang (1996) who grouped them in 

the same species based on their morphological 

characteristics of having broad leaflets with flat 

or undulate margins. All of Cycas in cluster B are 

from other countries. 

RFLP analysis 

For RFLP, three probes were synthesized 


0.00 0.25 0.50 0.75 1.00 

A 

B 

AI 

AI 

from 5S rRNA gene, 5S rRNA repeat unit gene 

and 18S rRNA gene (Figures 3 and 4). The probes 

were hybridized with genomic DNA of all nineteen 

Cycas species, which had been digested with 

restriction enzyme BamHI, EcoRI, DraI, HindIII 

and StuI. It was found that three restriction 

enzymes: BamHI, EcoRI and DraI could be used 

to identify Cycas species as also seen in the genetic 

relationships of Leymus species (Anamthawat- 

Jonsson and Bodvarsdottir, 2001). However, 

HindIII and StuI could not be used to identify the 

differences among the nineteen Cycas species 

because restriction sites of HindIII and StuI might 

be in the conserved regions on Cycas (data not 

shown). 

The results showed 33 polymorphic 

bands which were further used to construct 

phylogenetic tree by NTSYS-pc program as 

C. chamaoensis 

C. macrocarpa 

C. pectinata 

C. clivicola 

C. siamensis 

C. michoitzii 

C. simplicipinna 

C. edentata 

C. litoralis 

C. pranburiensis 

C. tansachana 

C. nongnoochiae 

C. seemannii 

C. wadei 

C. bougainvilleana 

C. nathorstii 

C. chevalieri 

C. parvulus 

C. diannanensis 

Figure 3 Phylogenetic tree of nineteen Cycas species using RAPD technique. The drogram was 

generated from similarity index based on UPGMA. The species in bold letters were of Thailand 

origin. 

I 

II

114 

subjected to UPGMA and resulting in the 

similarity index in the range of 1.0 -0.595 

(Table 3). Maximum similarity at 100% was found 

between C. wadei and C. chevalieri as well as 

C. edentata and C. litoralis. Although both pairs 

showed 100% similarity, but only C. edentata and 

C. litoralis was in accordance with RAPD analysis. 

The least similarity of 50% were found between 

C. bougainvilleana and C. pranburiensis, 

C. siamensis and C. pranburiensis, 

C. nongnoochiae and C. pranburiensis. The 

constructed dendrogram also separated nineteen 

Cycas species into two clusters, A and B (Figure 

6). Cluster A was further divided into group I, II 

and III. Group I comprised of C. chamaoensis, 

C. pectinata and C. bougainvilleana. It was found 


that C. pectinata and C. chamaoensis were closely 

related to each other and both were also in the 

Indosinenses section based on their morphological 

classification (Table 5). Group II consisted of 

C. macrocapa, C. clivicola, C. simplicipinna, 

C. edentata, C. litoralis and C. nathorstii. The 

similarity index indicated that C. macrocapa and 

C. clivicola were closely related to one another 

and having the same Thailand origin, while 

C. edentata and C. litolaris had very high similarity 

index which could not be separated from each 

other and they were also in the same Rumphiae 

subsection (Table 5). Group III comprised 

of C. wadei, C. chevalieri, C. parvulus, 

C. diannanensis and C. seemannii. However, 

C. wadei and C. chevalieri showed very high 

Figure 4 Southern analysis of genomic DNA of 19 Cycas species digested with BamHI and hybridized 

with (A) 18S rRNA and (B) 5S rRNA lane 1 C. clivicola, lane 2 C. macrocarpa, lane 3 

C. nathorstii, lane 4 C. wadei, lane 5 C. litoralis, lane 6 C. edentata, lane 7 C. chamaoensis, 

lane 8 C. seemannii, lane 9 C. pranburiensis, lane 10 C. simplicipinna, lane 11 C. micholitzii, 

lane 12 C. siamensis, lane 13 C. nongnoochiae, lane 14 C. tansachana, lane 15 

C. diannanensis, lane16 C. parvulus, lane 17 C. pectinata, lane 18 C. bougainvilleana, lane 

19 C. chevalieri.


Figure 5 Southern analysis of genomic DNA of 19 Cycas species digested with DraI (A) and hybridized with 5S rRNA (B) digested with Dra I and 

hybridizyed with 18S rRNA lane 1 C. litoralis, lane 2 C. edentata, lane 3 C. seemannii, lane 4 C. pranburiensis, lane 5 C. chevalieri, lane 6 C. 

parvulus, lane 7 C. pectinata, lane 8 C. diannanensis , lane 9 C. bougainvilleana, lane 10 C. wadei, lane 11 C. siamensis, lane 12 

C. micholitzii, lane 13 C. nongnoochiae, lane 14 C. simplicipinna, lane 15 C. nathorstii , lane 16 C. chamaoensis, lane 17 C. tansachana, lane 

18 C. macrocarpa, lane 19 C. clivicola (C) digested with EcoRI and hybridized with 18S rRNA lane 1 C. chevalieri, lane 2 C. parvulus, lane 

3 C. seemannii, lane 4 C. pranburiensis, lane 5 C. siamensis, lane 6 C. chamaoensis, lane 7 C. pectinata, lane 8 C. diannanensis, lane 9 

C. bougainvilleana, lane 10 C. wadei, lane 11 C. micholitzii, lane 12 C. pectinata, lane 13 C. litoralis, lane 14 C. edentata, lane 15 

C. simplicipinna, lane 16 C. nathorstii, lane 17 C. macrocarpa, lane 18 C. nongnoochiae, lane 19 C. tansachana. C. pranburiensis (lane 4) 

gave a unique 5.5 kb band as shown in (C) and the similar band pattern between C. litoralis (lane 13) and C. edentata (lane 14).

116 

Tabel 3 The similarity index of 19 Cycas species as identified by RFLP data. 

Species 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 

1 1.000 

2 0.765 1.000 

3 0.824 0.882 1.000 

4 0.706 0.706 0.824 1.000 

5 0.765 0.824 0.765 0.647 1.000 

6 0.882 0.765 0.824 0.647 0.765 1.000 

7 0.647 0.647 0.647 0.529 0.588 0.588 1.000 


8 0.706 0.824 0.706 0.588 0.882 0.765 0.588 1.000 

9 0.706 0.824 0.706 0.588 0.882 0.765 0.588 1.000 1.000 

10 0.824 0.706 0.765 0.706 0.588 0.706 0.529 0.529 0.529 1.000 

11 0.794 0.676 0.735 0.794 0.676 0.676 0.500 0.676 0.676 0.735 1.000 

12 0.735 0.559 0.676 0.676 0.559 0.618 0.500 0.559 0.559 0.853 0.824 1.000 

13 0.794 0.735 0.794 0.676 0.735 0.735 0.735 0.735 0.735 0.618 0.706 0.647 1.000 

14 0.735 0.618 0.618 0.559 0.676 0.735 0.618 0.735 0.735 0.559 0.647 0.588 0.647 1.000 

15 0.853 0.676 0.735 0.618 0.735 0.853 0.500 0.794 0.794 0.676 0.824 0.765 0.706 0.765 1.000 

16 0.735 0.794 0.794 0.676 0.735 0.676 0.735 0.735 0.735 0.618 0.706 0.647 0.824 0.706 0.706 1.000 

17 0.794 0.735 0.794 0.676 0.735 0.735 0.735 0.735 0.735 0.618 0.706 0.647 1.000 0.647 0.706 0.824 1.000 

18 0.765 0.882 0.765 0.647 0.882 0.765 0.647 0.941 0.941 0.588 0.735 0.618 0.794 0.735 0.794 0.794 0.794 1.000 

19 0.706 0.824 0.765 0.647 0.765 0.706 0.647 0.765 0.765 0.588 0.676 0.618 0.794 0.618 0.676 0.912 0.794 0.824 1.000 

Note: The names of the nineteen Cycas species are shown in Table 5.

similarity index to one another but they were in 

different sections, i.e., C. wadei belonged to 

Wadeanae section while C. chevalieri was in 

Stangerioides section (Table 5). 5S rRNA gene 

and 18S rRNA gene of these 2 species might share 

the same restriction sites, which resulted in the 

overlapping of the pattern. As for C. parvulus 

and C. diannanensis they were closely related 

as indicated by both their China origin and 

their belonging to the same Stangerioides section 

(Table 5). 

Group B comprised of C. micholitzii, 

C. siamensis, C. tansachana, C. nongnoochiae and 

C. pranburiensis. It was found that C. micholitzii 

was closely related to C. siamensis but they were 

not in the same morphological section. However, 

C. tansachana and C. nongnoochiae were closely 

related and both were in the same Indosinenses 

section while C. pranburiensis was not found to 

relate to other eighteen Cycas species and was 

considered an out-group. This was also confirmed 

by the extra band of 5.5 kb which was unique to 

only C. pranburiensis. The phylogenetic tree 


0.00 0.25 0.50 0.75 1.00 

A 

B 

C. chamaoensis 

C. pectinata 

C. bougainvilleana 

C. macrocarpa 

C. clivicola 

C. simplicipinna 

C. edentata 

C. litoralis 

C. nathorstii 

C. wadei 

C. chevalieri 

C. parvulus 

C. diannanensis 

C. seemannii 

C. micholitzii 

C. siamensis 

C. tansachana 

C. nongnoochiae 

C. pranburiensis 

Figure 6 Phylogenetic tree of nineteen Cycas species using RFLP technique. The dendrogram was 

generated from similarity index based on UPGMA. The species in bold letters were of Thailand 

origin. 

derived from RFLP technique, however, illustrated 

that Cycas of Thailand origin were randomly 

distributed in both cluster A and B. 

Levels of polymorphism between RAPD and 

RFLP 

Nineteen Cycas species were identified 

using two different marker systems. The level of 

polymorphism detected by each marker and their 

similarity index which indicated their cluster 

formation are shown in Table 4. For RAPD 

analysis, only five out of twenty primers were 

found to give polymorphic pattern among nineteen 

Cycas species. The number of polymorphic bands 

was 87, giving an average of 17.4 polymorphic 

bands per total number of primers used. RFLP 

analyses, on the other hand, three probes were used 

giving 33 total bands and the average number of 

polymorphic bands per total number of probes 

was16.5. Although the average number of 

polymorphic bands per total number of 

polymorphic probes/primers obtained from RFLP 

and RAPD were quite close but the cluster 

I 

II 

III

118 

formation as seen from phylogenetic tree by RAPD 

data agreed more to the geographic origins while 

those of RFLP data inclined to support the 

grouping by morphological characters. The 

differences found among the dendrogram 

generated by RAPD and RFLP could be partially 

explained by the different number of polymorphic 

bands (87 for RAPD and 33 for RFLP). The high 


number of RAPD bands may cover more genome 

than RFLP bands, which cover only the conserved 

regions. Since the results from RAPD method 

corresponded well with both their geographic 

origins and morphology, it may be a more efficient 

method for identifying different species of Cycas 

than using RFLP. 

Table 4 Characteristic of RFLP and RAPD markers used in analyzing the genetic variability of 19 

Cycas species. 

RAPD RFLP 

Total number of probes/primers 20 3 

Number of polymorphic probes/primers 5 2 

Total number of bands amplified by polymorphic probes/primers 87 33 

Average number of polymorphic bands per total number of polymorphic 

probes/primers 17.4 16.5 

Table 5 The origins and morphological sections of nineteen Cycas species. 

Number name origin morphological section/subsection* 

1 C. chamaoensis Thailand Indosinenses/Indosinenses 

2 C. macrocarpa Thailand Cycas/Cycas 

3 C. clivicola Thailand Indosinenses/Indosinenses 

4 C. micholitzii Vietnam Stangerioides/Stangerioides 

5 C. simplicipinna Thailand Stangerioides/Stangerioides 

6 C. pectinata China, Vietnam Indosinenses/Indosinenses 

and Thailand 

7 C. pranburiensis Thailand Cycas/Cycas 

8 C. edentata Philippines Cycas/Rumphiae 

9 C. litoralis Thailand, Vietnam, Cycas/Rumphiae 

Burma and Malaysia 

10 C. tansachana Thailand Indosinenses/Indosinenses 

11 C. siamensis Thailand Indosinenses/Indosinenses 

12 C. nongnoochiae Thailand Indosinenses/Indosinenses 

13 C. wadei Philippines Wadeanae/- 

14 C. seemannii Fiji Cycas/Rumphiae 

15 C. bougainvilleana Papua New Guinea Cycas/Rumphiae 

16 C. parvulus China Stangerioides/Stangerioides 

17 C. chevalieri Vietnam Stangerioides/Stangerioides 

18 C. nathorstii Sri Lanka Cycas/Cycas 

19 C. diannanensis China Stangerioides/Stangerioides 

* classified by Stevenson (1992)

CONCLUSION 

The RAPD and RFLP were used to 

identify nineteen species of Cycas. In the RAPD 

study, only five primers, i.e., OPB-1, OPB-8, OPB- 

14, OPB-15 and OPB-17 were found to give 

polymorphic band patterns. RAPD data were 

subjected to UPGMA and analysis was done by 

using NTSYS-pc (version 1.8). The similarity 

index revealed that the species relationships fell 

in the range of 0.816-0.516. The dendrogram 

separated nineteen Cycas species into two clusters 

A and B. Cluster A comprised all ten Thailand 

species of C. chamaoensis, C. macrocapa, 

C. pectinata, C. clivicola, C. siamensis, 

C. simplicipinna, C. litoralis, C. pranburiensis, 

C. tansachana, C. nongnoochiae as well as 

C. micholitzii, C. seemannii, C. edentata and 

C. wadei from other countries. Cluster B 

comprised of C. bougainvilleana, C. nathorstii, 

C. chevalieri, C. parvulus and C. diannanensis all 

of which are from other countries. The 

phylogenetic tree of RAPD seems to show good 

correlation with geographical distribution. 

For RFLP, three probes were synthesized 

from 5S rRNA gene, 5S rRNA repeat unit gene 

and 18S rRNA gene. The probes were hybridized 

with genomic DNA of all nineteen species of 

Cycas which had been digested with restriction 

enzyme BamHI, EcoRI and Dra I. RFLP data were 

subjected to UPGMA and analysis was done by 

using NTSYS-pc (version 1.8). The similarity 

index showed that the species relationships fell in 

the range of 1.0-0.595. The constructed 

dendrogram also separated them into two clusters 

A and B. Cluster A comprised of C. chamaoensis, 

C. pectinata, C. bougainvilleana, C. macrocapa, 

C. clivicola, C. simplicipinna, C. edentata, 

C. litoralis, C. nathorstii, C. wadei, C. chevalieri, 

C. parvulus, C. diannanensis and C. seemannii. 

Cluster B comprised of C. micholitzii, 

C. siamensis, C. tansachana and C. nongnoochiae 

while C. pranburiensis was not found to relate to 


eighteen other species and was considered an outgroup. 

The distribution of nineteen Cycas species 

within the dendrogram of RFLP has no apparent 

relation with the geographical origin but more 

related to their morphological characteristics. 


This work was financially supported by 

Kasetsart University Research and Development 

Institution, Grant No 22.5.4.1. We would like to 

thank Nong Nooch Tropical Botanical Garden for 

providing specimen to use in this work. 


Ahmad, F. 1999. Random amplified polymorphic 

DNA (RAPD) analysis reveals genetic 

relationships among the annual Cicer species. 

Theor. Appl. Genet. 98: 657-663. 

Anamthawat-Jonsson, K. and K.S. Bodvarsdottir. 

2001. Genomic and genetic relationships 

among species of Leymus (Poaceae:Triticeae) 

inferred from 18S-26S ribosomal gene. Amer. 

J. Bot. 88: 553-559. 

Caputo, P., C. Marquis, T. Wurtzel, D. Stevenson 

and T.E. Wurtzel. 1993. Molecular biology 

in Cycad systematic, pp. 213-219. In D.W. 

Stevenson and K.J. Norstog (eds.). 

Proceedings of CYCAD 90, the Second 

International Conference on Cycad 

Biology. Palm & Cycad Societies of Australia, 

Milton, Queensland. 

Dettori, T.M. and A.M. Palombi. 2000. 

Identification of Feijoa sellowiana Berg 

accession by RAPD marker. Scienta 

Horticulturae 86: 276-290. 

Doyle, J.J. and J.L. Doyle. 1990. Isolation of plant 

DNA from fresh tissue. Focus 12: 12-15. 

Garcia-Mas, J., M. Oliver, H. Gomez-Paniagua 

and M.C. de Vicente. 2000. Comparing AFLP, 

RAPD and RFLP markers for measuring 

genetic diversity in melon. Theor. Appl.

120 

Genet. 101: 860-864. 

Lerceteau, E., T. Robert, V. Petiar and D. 

Crouzillat. 1997. Evolution of the extent of 

genetic variability among Theobroma cacao 

accessions using RAPD and RFLP marker. 


Nicolosi, E., N.Z. Deng, A. Genetile, La S. Malfa, 

G. Continella and E. Tribuloto. 2000. Citrus 

phylogeny and genetic origin of important 

species as investigated by molecular marker. 


Nei, M. and W.H. Li. 1979. Mathematical model 

for studying genetic variation in terms of 

restriction endonuclease. PNAS. 79: 5269- 

5273. 

Pu, H. and W. Chiu. 1990. Genetic variation in 

Cycas section Stangerioides using anchored 

microsatellite primers, pp. 334-344. In C.J. 

Chen (ed.). Proceeding of the Fourth 

International Conference on Cycad Biology 

(Panzhihua, China). International Academic 

Publishers, Beijing. 

Roger. S.O. and A.J. Bendich. 1987. Ribosomal 

RNA genes in plants variability in copy 


number and in the intergenic spacer. Plant 

Mol. Biol. 9: 509-520. 

Stevenson, D.W. 1992. A formal classification of 

the extant Cycad. Brittonia 44: 220-223. 

Sun, C.Q., X.K. Wang, Z.C. Li, A. Yashimura and 

N. Iwana. 2001. Comparison of the genetic 

diversity of common wild rice (Oryza 

rufipogon Grif f.) and cutivated rice (O.sativa 

L.) using RFLP markers. Theor. Appl. Genet. 

102: 157-162. 

Wang, D.Y. 1996. Taxonomy, pp 1-142 In F.X. 

Wang and H.L. Liang (eds.). Cycad in China. 

Guangdong Science and Technology Press, 

Guangzhou. 

Williams, G.K.J., R.A. Kubelik, J.K. Livak, A.J. 

Rafalski and V.S. Tingey. 1990. DNA 

polymorphisms amplified by arbitrary primers 

are useful as genetic markers. Nucleic Acid 

Research 18: 6531-6535. 

Yang, S. and A.W. Meerow. 1996. The Cycas 

pectinata (Cycadaceae) complex:Genetic 

structure and gene flow. Int. J. Plant Sci. 157: 

468-483.


Physiological Study and Alcohol Oxidase Gene(s) of Thermotolerant 

Methylotrophic Yeasts Isolated in Thailand 

Nantana Srisuk 1 *, Savitree Limtong 1 , Hiroya Yurimoto 2 , 

Yasuyoshi Sakai 2 and Nobuo Kato 2 

ABSTRACT 

Two thermotolerant methylotrophic yeasts, Candida sithepensis sp. nov. and Pichia siamensis 

sp. nov. isolated in Thailand were studied for their physiology and alcohol oxidase gene(s). Optimum 

and maximum temperatures for growth on methanol of C. sithepensis sp. nov. were 34°C and 37°C 

whereas those of P. siamensis sp. nov. were 37°C and 40°C, respectively. Effect of methanol on the 

methylotrophic growth of the two yeast strains was investigated within the ranges of 0.25% to 2% (v/v) 

methanol. At 34°C, C. sithepensis sp. nov. exhibited a decline in µ max from 0.16 h -1 to 0.13 h -1 with 

respect to increase in methanol concentration. When cultivated at 37°C, the µ max of C. sithepensis sp. 

nov. was found to be drastically decreased from 0.16 h -1 to 0.10 h -1 within the same ranges of methanol 

concentration studied. P. siamensis sp. nov. showed an insignificant decrease of µ max with respect to an 

increase in methanol supplement when cultivated at 37°C. 

Key enzymes involved in methanol dissimilatory pathway, i.e., alcohol oxidase (AOD), 

glutathione-dependent formaldehyde dehydrogenase (FLD), formate dehydrogenase (FDH) and catalase 

(CAT) of the two yeasts were assessed. As for the favorable growth of C. sithepensis sp. nov. at its 

optimum temperature, fairly high level of specific AOD activity was observed but lower level of specific 

AOD activity was exhibited at maximum temperature. In contrast, specific FLD and FDH activity 

profiles were found to be higher when the yeast C. sithepensis sp. nov. was grown at its maximum 

temperature compared to the methylotrophic growth at its optimum temperature. Catalase production 

appeared to be slightly disturbed under all conditions studied. In case of P. siamensis sp. nov., stationary 

phase growth under low methanol supplement at its optimum temperature appeared soon after 24h 

cultivation. 

Nucleotide sequences of AOD gene(s) of the two yeasts were investigated. C. sithepensis sp. 

nov. showed two AOD genes designated as AOD1 and AOD2 whereas P. siamensis sp. nov. revealed 

only one AOD gene within the genome. 

Key words: methylotrophic yeasts, alcohol oxidase gene, thermotolerant, physiological study 

1 Department of Microbiology, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

2 Laboratory of Microbial Biotechnology, Division of Applied Life Sciences, Graduate School of Agriculture, Kyoto University, 

Kyoto 606-8502, Japan. 

* Corresponding author, e-mail: fscints@ku.ac.th 


122 

INTRODUCTION 

Methylotrophic yeasts are widely used 

for heterologous protein production (Cereghino 

and Cregg, 2000; Gellissen, 2000; Faber et al., 

1995). Two yeast genera, i.e., Pichia spp. and 

Candida spp. have now been recognized as 

methylotrophic (Houard et al., 2002). Several 

species: P. pastoris, P. methanolica, P. angusta, 

C. boidinii and C. methanolovescens are known 

as representatives of these two genera (Demain et 

al., 1998). Methanol metabolism starts with the 

oxidation of methanol to formaldehyde and 

hydrogen peroxide. The reaction is catalyzed by 

alcohol oxidase and the hydrogen peroxide 

produced is subsequently removed by the reaction 

catalyzed by catalase. This beginning part of 

methanol metabolism is localized in peroxisomes 

whereas the subsequent reactions in methanol 

assimilation and dissimilation are localized in the 

cytosol ( Fukui et al., 1975; Douma et al., 1985; 

Goodman, 1985; Veenhuis and Harder, 1987). 

Dissimilation of formaldehyde proceeds to formate 

and further to CO 2, and is catalyzed by 

formaldehyde dehydrogenase and formate 

dehydrogenase, respectively, resulting in the 

generation of two molecules of NADH (Veenhuis 

and Harder, 1987). 

The merits of methylotrophic yeasts 

applied to various processes are due to the 

capability to utilize methanol as the sole carbon 

and energy source. However, industrial 

applications have mostly been limited to the 

inability of these yeast strains to grow at high 

temperature. This affects not only the yeast growth 

but also the production of yeast proteins or 

metabolites. It is, therefore of advantage to search 

for thermotolerant methylotrophic yeast strains to 

overcome this detrimental effect of high 

temperature. Thailand is one of the tropical 

countries having reasonably high ambient 

temperature throughout the year. As a result of such 

climate, the thermotolerant microbial resources 


have therefore been the target of interest. Limtong 

et al. (2004) described three new thermotolerant 

methylotrophic yeasts isolated in Thailand, i.e., 

Candida krabiensis sp. nov., C. sithepensis sp. nov. 

and Pichia siamensis sp. nov. 

In this study it was aimed to determine 

the optimum and maximum temperatures for 

growth of the two new yeast isolates described by 

Limtong et al. (2004). In addition, the effect of 

methanol on the methylotrophic growth at 

optimum and maximum temperatures was 

evaluated. Methanol utilizing enzymes of 

these yeast isolates were investigated under 

different growth conditions. Furthermore, alcohol 

oxidase gene(s) of the isolates of interest were also 

studied. 


Yeast strains 

All yeast isolates, except Pichia angusta, 

were isolated in Thailand from natural samples 

(flowers, fruits, leaves, soils and water) by 

the selective and enrichment isolation 

technique. Pichia angusta, on the other hand, was 

obtained from Laboratory of Microbial 

Biotechnology, Division of Applied Life Sciences, 

Graduate School of Agriculture, Kyoto University, 

Japan. 

Cultivation and media 

Yeast isolates were maintained in YPD 

slant culture. Methanol synthetic medium (Sakai 

et al., 1995) was used for the shake flask 

cultivation. Methanol concentration varied as 1% 

(v/v) for determination of optimum and maximum 

temperatures for growth, 0.25-2% (v/v) to study 

the effect of methanol on methylotrophic yeast 

growth and 0.75-2% (v/v) for the production of 

methanol utilizing enzymes. Yeast cultivation (100 

ml) to determine temperatures and methanol 

concentrations critical to yeast growth was 

conducted aerobically at various temperatures, i.e.,

20-40°C to identify optimum and maximum 

temperatures for growth; 34°C, 37°C and 40°C, 

depending on yeast isolate, for the study of yeast 

growth as well as methanol utilizing enzymes 

production which might be affected by methanol 

supplement. All incubations were performed with 

rotary shaking at 180 rpm in 250 ml Erlenmeyer 

flasks. However, 400 ml culture in 1 L Erlenmeyer 

flasks were used to investigate the methanol 

utilizing enzymes. 

Growth determination 

Yeast growth was determined by 

measuring the optical density at 610 nm (OD 610). 

Preparation of cell-free extracts 

The methanol-grown cells were washed 

in 50mM potassium phosphate buffer (pH 7.5), 

and then transferred to a 2 ml Eppendorf tube 

containing an equal volume of 0.5 mm glass beads. 

The tube was shaken vigorously for 30 s on a minibeadbeater 

(Biospec Products, UK) and chilled on 

ice for 30 s. This procedure was repeated six times. 

Cell debris was discarded after centrifugation at 

16,000g for 5 min at 4°C. The supernatant was 

then subjected to enzyme activity assay and protein 

content determination. 

Assay of methanol utilizing enzymes and 

protein content 

Alcohol oxidase activity was determined 

by the ABTS/POD method (Tani et al., 1985) 

whereas catalase activity was determined as a 

decrease in H 2O 2 in the reaction mixture 

(Bergmeyer, 1955). Activities of formaldehyde 

dehydrogenase and formate dehydrogenase were 

measured by determining the rate of NADH 

formation at 340 nm as described by Schutte et al. 

(1976). Protein concentrations of cell-free extracts 

were determined by the method of Bradford (1976) 

using a protein assay kit (Bio-Rad Laboratories) 

having bovine serum albumin as the standard. 

Methanol in the culture medium was determined 


by gas chromatography (Shimadzu GC-9A, Japan) 

and chromatopac recorder (Shimadzu C-R3A, 

Japan). 

DNA methods 

Genomic DNA of the yeast strains was 

prepared using the protocols described by Higgins 

and Cregg (1998). The primer employed for PCR 

amplification of AOD gene was designed. Two 

mixed primers, AODFp (5′-AAYCCS 

TGGGTSTAYYTSCCNGG-3′) and AODRp (5′- 

GGRTAYTCNTRRAARTGRAACAT-3'), were 

obtained as a result of the primer design based on 

the amino acid sequences of highly conserved 

regions in several AOD encoding genes 

(NPWVYLPG and NFHFLEY). Using these two 

primers and genomic DNA from strain C. 

sithepensis sp. nov. or P. siamensis sp. nov. as a 

template, a 1.1 kb fragment was specifically 

amplified in a BIO-RAD I Cycler thermal cycler. 

Purified PCR product was subjected to the TAcloning 

into the bacterial host, Escherichia coli 

DH5a using pGEM-T Easy Vector System 

(Promega). Transformed colonies were picked up 

to prepare the plasmid DNA. The plasmids 

obtained were checked for the correct insertion of 

AOD gene by digestions with the restriction 

enzymes EcoRI and NotI. The PCR amplified 

AOD fragment was then labeled with dioxetane 

chemiluminescence using the Gene ImagesTM 

AlkPhos DirectTM labelling and detection system 

from Amersham Pharmacia Biotech. DNA 

hybridization was performed at 55°C as 

recommended by the manufacturer. Southern 

analysis was then carried out to investigate the 

number(s) of AOD gene of both yeast isolates by 

individual digestion of the genomic DNA with the 

restriction enzymes BamHI, EcoRI, HindIII, SacI, 

XbaI and XhoI. DNA sequences of the cloned 

fragments were determined using dye deoxy 

method in DSQ2000L DNA sequencer (Shimadzu 

Co., Ltd) and nucleotide sequences were analyzed.

124 


Screening for thermotolerant methylotrophic 

yeasts 

253 isolates of methylotrophic yeast 

were obtained from natural samples such as fruits, 

flowers, leaves, soil and natural water using the 

selective isolation technique at 30°C. Further 

screening on 1% methanol agar medium revealed 

that 124 isolates were able to grow at 37°C and 

54 isolates showed good to very good growth. 

Among these, 10 isolates exhibited higher 

maximum specific growth rate than that observed 

in P. angusta when cultivated under shake 

flask cultivation in synthetic broth containing 

1% methanol as sole carbon and energy source. 

In comparison to P. angusta, the isolates identified 

as P. siamensis sp. nov. and C. sithepensis sp. 

nov. (Limtong et al., 2004) were selected for 

further studies according to the highest maximum 

specific growth rate (0.152 h -1 ) for P. siamensis 

sp. nov. and to the highest maximum cell 

density determined as OD 610 (14.100) for C. 

sithepensis sp. nov. Results are summarized 

in Table 1. In addition, when cultivated at 

25°C and 30°C, C. sithepensis sp. nov. also 

exhibited higher maximum cell density than 

P. angusta. 


Further screening at 40°C indicated that 

17 isolates showed methylotrophic growth on 1% 

methanol agar medium. However, P. siamensis sp. 

nov. was the only isolate that showed the 

methylotrophic growth at 40°C under shake flask 

cultivation in synthetic broth containing 1% 

methanol. C. sithepensis sp. nov. and P. siamensis 

sp. nov. were, therefore, subject to further 

characterization. 

Determination of optimum and maximum 

temperatures for growth 

Optimum and maximum temperatures 

for growth of the selected yeast isolates were 

determined from the relationship between m max 

and growth temperatures as shown in Figure 1. 

Results indicated that C. sithepensis sp. nov. 

possessed the ranges of optimum temperature of 

30-34°C and a maximum temperature of 37°C. 

Although negligible difference in maximum 

specific growth rate could be observed, it should 

be noted that 34°C were preferably used as an 

optimum temperature for further investigation due 

to a slightly greater growth rate obtained. Similar 

pattern was shown between P. siamensis sp. nov. 

and P. angusta regarding their optimum and 

maximum temperatures which appeared to be 

37°C and 40°C, respectively. 

Table 1 Maximum specific growth rate and cell density of ten yeast isolates cultivated in 1% methanol 

synthetic broth at 37°C. 

isolates µ max (h -1 ) OD 610 max / (time) 

Pichia siamensis sp. nov. 0.152 10.100 / (60h) 

S051 0.145 9.950 / (84h) 

FS30 0.143 7.650 / (48h) 

Candida sithepensis sp. nov. 0.143 14.100 / (72h) 

M02 0.142 11.350 / (60h) 

FS95 0.141 11.400 / (72h) 

FS101 0.140 11.950 / (60h) 

FS26 0.139 12.450 / (60h) 

FS56 0.138 9.950 / (72 h) 

PT27 0.128 9.950 / (60 h) 

Pichia angusta 0.123 12.100 / (60h)

Effect of methanol on methylotrophic yeast 

growth 

Effect of methanol on yeast growth was 

individually investigated at optimum and 

maximum temperatures for growth. The two 

isolates of interest, C. sithepensis sp. nov. and P. 

siamensis sp. nov., were cultivated along with the 

reference strain P. angusta under shake-flask 

cultivation in synthetic medium supplemented 

with various methanol concentrations (0.25% - 

2%). Results shown in Figure 2 indicated that, at 

its optimum temperature (34°C), C. sithepensis sp. 

nov. exhibited a decline in µ max from 0.16 h -1 to 

0.13 h -1 (decreased to 81.25% of its original value) 

when methanol concentration increased up to 2%. 

When cultivated at its maximum temperature (37° 

C), the maximum specific growth rates of C. 

sithepensis sp. nov. were found to decrease from 

0.16 h -1 to 0.10 h -1 (decreased to 62.5% of its 

original value) within the same ranges of methanol 

concentration studied. P. siamensis sp. nov. 

showed a decline in µ max from 0.14 h -1 to 0.12 h -1 

(decreased to 85.71% of its original value) with 

respect to an increase in methanol supplement 

µ max (h -1 ) 

0.20 

0.15 

0.10 

0.05 

0.00 


from 0.25% to 2% when cultivated at 37°C. 

Methylotrophic growth of P. siamensis sp. nov. at 

40°C appeared to be inferior to that observed at 

37°C cultivation as shown by generally lower level 

of µ max (0.08-0.10 h -1 ). However, similar profile, 

i.e., negligible change in µ max was noticed (Figure 

3). Effect of methanol concentrations on 

methylotrophic growth of the reference strain, i.e., 

P. angusta at 34°C, 37°C and 40°C was also 

investigated (Figure 4). Results obtained in 

P. angusta were found to be comparable to those 

observed in P. siamensis sp. nov. in terms of 

negligible changes in µ max (0.11-0.14 h -1 ) at all 

cultivation temperatures. 

Study of growth and methanol utilizing 

enzymes production 

It is interesting to find that, at the 

maximum temperature for growth of the two 

isolates of interest, C. sithepensis sp. nov. exhibited 

a drastic response to methanol concentration in 

the culture medium whereas a decline in m max of 

P. siamensis sp. nov. was observed in a lesser 

extent. Contrary to the methylotrophic growth at 

15 20 25 30 34 37 40 45 

Temperature (°C) 

Figure 1 Maximum specific growth rate of the methylotrophic yeast C. sithepensis sp. nov. ( ), P. 

siamensis sp. nov. ( ) and P. angusta ( ), on 1% methanol synthetic broth under shake flask 

cultivation at different temperatures for growth.

126 

optimum temperature, insignificant differences 

between the two isolates could be pointed out. 

This may indicate some physiological role of 

AOD gene(s) of the strain grown at 

high temperature as described in P. methanolica 

µ max (h -1 ) 

µ max (h -1 ) 

0.20 

0.15 

0.10 

0.05 

0.00 

0.20 

0.15 

0.10 

0.05 


(Nakagawa et al., 2002). This part of experiment 

was therefore aimed to investigate enzymatic 

profiles of the methylotrophic yeast C. sithepensis 

sp. nov. and P. siamensis sp. nov. at their 

optimum and maximum temperatures for growth 

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 

Methanol (%v/v) 

Figure 2 Maximum specific growth rates of C. sithepensis sp. nov. cultivated at 34°C ( ) and 37°C 

( ) in the synthetic methanol broth supplemented with various concentrations of methanol. 

0.00 

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 


Figure 3 Maximum specific growth rates of P. siamensis sp. nov. cultivated at 37°C ( ) and 40°C ( ) 

in the synthetic methanol broth supplemented with various concentrations of methanol.

under low and high methanol supplement to the 

culture. 

Depending on yeast isolate and 

experimental results obtained at optimum and 

maximum temperatures for growth, the profile of 

methanol utilizing enzymes produced under low 

and high methanol concentrations would be 

individually assessed. Despite the fact that, in some 

cases, low methanol concentrations, e.g., 0.25% 

appeared to yield an amenable growth rate, 

however, the culture possessed low cell density 

and shortly reached a stationary phase growth due 

to the depletion of carbon and energy source, i.e., 

methanol. This obviously resulted in an inadequate 

amount of harvested cells needed for enzyme 

activity assay. The culture density was, therefore, 

taken into account when “low” methanol 

concentration was assigned. Activities of catalase 

(CAT) and some key enzymes responsible for 

dissimilatory pathway of methanol utilization, i.e., 

alcohol oxidase (AOD), glutathione-dependent 

formaldehyde dehydrogenase (FLD) and formate 

dehydrogenase (FDH) were assessed with growth 

µ max (h -1 ) 

0.20 

0.15 

0.10 

0.05 


of the methylotrophic yeast C. sithepensis sp. nov. 

and P. siamensis sp. nov. along with those observed 

in P. angusta. 

A time-course determination of growth, 

methanol and specific enzyme activities of the 

isolate C. sithepensis sp. nov. was conducted at 

various growth conditions (Figure 5). Methanol 

concentration appeared to introduce insignificant 

effect to yeast growth at its optimum temperature 

cultivation whereas a retarded growth pattern was 

observed as a result of high methanol supplement 

at its maximum temperature. As for the favorable 

growth at optimum temperature, fairly high level 

of specific AOD activity was observed (maximum 

activity of 7.65 U/mg protein) as indicated in 

Figure 5C. However, lower level of specific 

activity was exhibited at maximum temperature 

for growth. During exponential growth of C. 

sithepensis sp. nov. (12-36 h of cultivation), 

specific AOD activity appeared to be lower when 

methanol concentration increased. These results 

agreed with high AOD activity observed at low 

methanol concentration in the inflow growth 

0.00 

0.00 0.25 0.50 0.75 1.00 1.25 1.50 1.75 2.00 2.25 


Figure 4 Maximum specific growth rates of the isolate P. angusta cultivated at 34°C ( ), 37°C ( ) 

and 40°C ( ) in the synthetic methanol broth supplemented with various concentrations of 

methanol.

128 

medium of C. boidinii (Volfova et al., 1992). 

However, AOD production profiles appeared to 

be contrary to those observed for FLD production 

shown in Figure 5E. Specific FLD activity profiles 

was found to be higher when the yeast 

OD 610 


100 

10 

1 

0 

2.00 

1.50 

1.00 

0.50 

0.00 

0 12 24 36 48 60 

Times (h) 


A 

B 

C. sithepensis sp. nov. grown at its maximum 

temperature compared to the methylotrophic 

growth at its optimum temperature. Similar 

pattern, although in the lesser extent, was 

shown for FDH production. Figure 5D shows 

Specific activity (u/mg protein) 

8 

6 

4 

2 

0 

4000 

3000 

2000 

1000 

0 

0.0 

0 12 24 36 48 60 

Times (h) 

Figure 5 Growth (A), methanol residue (B) and specific activities of alcohol oxidase (C), catalase (D), 

glutathione-dependent formaldehyde dehydrogenase (E) and formate dehydrogenase (F) of 

C. sithepensis sp. nov. cultivated at various growth conditions. 34 °C and 0.75% (v/v) 

methanol, (●); 34 °C and 2% (v/v) methanol, ( ); 37 °C and 0.75% (v/v) methanol, (◆); 37 

°C and 1.5% (v/v) methanol, ( ) 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

0.5 

0.4 

0.3 

0.2 

0.1 

F 

C 

D 

E

insignificant differences on specific CAT activity 

observed under all conditions studied. A relatively 

stable CAT production of C. sithepensis sp. nov. 

was concurrent to those appeared in C. boidinii 

cultures (Volfova et al., 1992). 

OD 610 


100 

10 

1 

0 

2.00 

1.50 

1.00 

0.50 

0.00 

0 12 24 36 48 60 

Times (h) 


A 

B 

Figure 6 summarizes the results of timecourse 

determination for growth and enzymatic 

profiles of P. siamensis sp. nov. Stationary phase 

growth of P. siamensis sp. nov. under low methanol 

supplement (0.75%) at its optimum temperature 


8 

6 

4 

2 

0 

4000 

3000 

2000 

1000 

0 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

0.5 

0.4 

0.3 

0.2 

0.1 

0.0 

0 12 24 36 48 60 

Times (h) 

Figure 6 Growth (A), methanol residue (B) and specific activities of alcohol oxidase (C), catalase (D), 

glutathione-dependent formaldehyde dehydrogenase (E) and formate dehydrogenase (F) of 

P. siamensis sp. nov. cultivated at various growth conditions. 37 °C and 0.75% (v/v) methanol, 

(▲); 37 °C and 2% (v/v) methanol, ( ); 40 °C and 0.75% (v/v) methanol, (■); 40 °C and 

2% (v/v) methanol, ( ) 

E 

F 

C 

D

130 

was observed soon after 24h cultivation. This 

obviously resulted in the decrease of methanol 

utilizing enzyme production and, probably, the 

incomplete methanol utilization under low 

methanol supplement. The highest specific AOD 

activity observed when the yeasts C. sithepensis 

sp. nov. and P. siamensis sp. nov. grown under the 

same growth condition, i.e., low methanol level 

and optimum temperature for growth were 

compared. Result obtained from P. siamensis sp. 

nov. (4.5 U/mg protein) was approximately half 

of that from C. sithepensis sp. nov. (7.5 U/mg 

protein). 

Determination of growth, methanol 

and specific enzyme activities of P. angusta 

cultivated at various growth conditions were also 

investigated in the same manner and results are 

summarized in Figure 7. Insignificant differences 

could be observed under all growth conditions 

studied. 

Study of alcohol oxidase gene(s) 

AOD fragment was amplified from yeast 

genomic DNA using AOD degenerated primers as 

follows: 

AODFp 

5’-AAYCCSTGGGTSTAYYTSCCNGG-3’ 

AODRp 

5’-GGRTAYTCNTRRAARTGRAACAT-3’ 

This fragment was then cloned into E. 

coli prior to the partial nucleotide sequencing. 

Yeast genomic DNA was digested with various 

restriction enzymes and subjected to Southern 

analysis using PCR amplified AOD fragment as a 

probe. Results of Southern analysis are shown in 

Figure 8. DNA band numbers were determined 

between genomic DNA samples of individual 

yeast isolates digested with the same set of 

restriction enzymes. Obviously, C. sithepensis sp. 

nov. contained more than one AOD gene because 

Southern analysis of genomic DNA digested with 


BamHI, EcoRI, XbaI and XhoI showed two bands 

whereas P. siamensis sp. nov. possessed only one 

AOD gene as seen from a single band when the 

genomic DNA was digested with the same set of 

restriction enzymes. These indications were 

coincident with the results obtained from partial 

nucleotide sequence indicating that C. sithepensis 

sp. nov. possessed two different AOD genes 

whereas P. siamensis sp. nov. had only one AOD 

gene. Alcohol oxidase genes have been isolated 

and characterized from P. angusta (Ledeboer et 

al., 1985), P. pastoris (Koutz et al., 1989), C. 

boidinii (Sakai and Tani, 1992) and P. methanolica 

(Nakagawa et al., 1996). Among these yeast 

species, P. pastoris and P. methanolica have been 

reported to have two methanol oxidase genes. 

These results were therefore the first evidence of 

two AOD genes existed in the genus Candida of 

methanol-utilizing yeasts. Alignment of deduced 

amino acid sequences of AOD of C. sithepensis 

sp. nov. and P. siamensis sp. nov. compared to the 

database (Figure 9) suggested that the 

C. sithepensis sp. nov.- AOD1 had 90% homology 

to P. methanolica- MOD1 whereas the 

C. sithepensis sp. nov.- AOD2 had 84% homology 

to P. methanolica- MOD2. The P. siamensis sp. 

nov.- AOD was found to be 97% homology 

to the P. angusta- MOX. Similarity of alcohol 

oxidase of the two methylotrophic yeasts 

is summarized in Figure 10 using UPGMA 

analysis. 

As a result of two AOD genes exist in 

C. sithepensis sp. nov., further studies on the 

regulation of these two genes at transcriptional 

level should be performed. In addition, 

an existence of alcohol oxidase isozymes will 

also be investigated under different growth 

conditions. 

CONCLUSIONS 

Optimum and maximum temperatures 

for growth on methanol of the two thermotolerant

OD 610 


100 

10 

1 

0 

2.00 

1.50 

1.00 

0.50 

0.00 

0 12 24 36 48 60 

Times (h) 


A 

B 


8 

6 

4 

2 

0 

4000 

3000 

2000 

1000 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

0.5 

0.4 

0.3 

0.2 

0.1 

0.0 

0 12 24 36 48 60 

Times (h) 

Figure 7 Growth (A), methanol residue (B) and specific activities of alcohol oxidase (C), catalase 

(D), glutathione-dependent formaldehyde dehydrogenase (E) and formate dehydrogenase 

(F) of P. angusta cultivated at various growth conditions. 34 °C and 0.75% (v/v) methanol, 

(▲); 34 °C and 0.75% (v/v) methanol, ( ); 37 °C and 0.75% (v/v) methanol, (■); 37 °C and 

1.5% (v/v)methanol, (×); 37 °C and 2% (v/v) methanol, ( ); 40 °C and 0.75% (v/v) methanol, 

(●); 40 °C and 2% (v/v) methanol, ( ) 

0 

E 

F 

C 

D

132 


Figure 8 Southern blot analysis of genomic DNA of the methylotrophic yeast C. sithepensis sp. nov. 

(S, left panel) and P. siamensis sp. nov. (FS, right panel) with partial AOD gene(s). Genomic 

DNAs were digested with various restriction enzymes as indicated. 

methylotrophic yeasts, Candida sithepensis sp. 

nov. and Pichia siamensis sp. nov., isolated from 

Thai fruits and soil were 34°C and 37°C for 

C. sithepensis sp. nov. and 37°C and 40°C for 

P. siamensis sp. nov., respectively. The µ max of the 

two yeasts was affected by methanol supplement. 

At its optimum temperature for growth (34°C), 

increasing methanol concentration within the 

range of 0.25 to 2% (v/v) decreased the µ max of C. 

sithepensis sp. nov. from 0.16 h -1 to 0.13 h -1 . When 

cultivated at its maximum temperature (37°C), the 

µ max of C. sithepensis sp. nov. drastically 

decreased within the same ranges of methanol 

concentration studied. Methanol supplement 

appeared to have slight effect to P. siamensis sp. 

nov. at its optimum temperature for growth. 

Reasonably high level of specific AOD 

activity was observed during the favorable growth 

of C. sithepensis sp. nov. at its optimum 

temperature. Lower level of specific AOD activity 

was shown at its maximum temperature as for an 

inferior growth appeared. Formaldehyde 

dehydrogenase production was contrast to AOD 

production since higher specific FLD activity 

profile was observed when the yeast C. sithepensis 

sp. nov. grown at its maximum temperature 

compared to the methylotrophic growth at its 

optimum temperature. Similar pattern, although 

in the lesser extent, was shown for FDH 

production. Insignificant differences on specific 

CAT activities could be observed. In case of P. 

siamensis sp. nov., stationary phase growth under 

low methanol supplement at its optimum 

temperature appeared soon after 24h cultivation. 

C. sithepensis sp. nov. showed two AOD 

genes designated as AOD1 and AOD2 whereas P. 

siamensis sp. nov. revealed only one AOD gene 

within the genome. AOD1 of C. sithepensis sp. 

nov. had 90% homology to MOD1 of P. 

methanolica whereas the AOD2 of C. sithepensis 

sp. nov. had 84% homology to the MOD2 of P. 

methanolica. High similarity (97% homology) of 

P. siamensis sp. nov. - AOD to the P. angusta- MOX 

was exhibited.


S-AOD1 NPWVYLPGVYPRNMRLDSKTATFYTARPSPHLNGRRAIVPCANILGGGSSINFLMYTRAS 

PmMOD1 NPWVYLPGVYPRNMRLDSKTATFYSSRPSPHLNGRRAIVPCANILGGGSSINFLMYTRAF 

FS-AOD NPWVYLPGVYPRNMRLDSKTATFYTSRPSKALNGRRAIVPCANILGGGSSINFLMYTRAS 

HpMOX NPWVYLPGVYPRNMRLDSKTATFYSSRPSKALNGRRAIVPCANILGGGSSINFLMYTRAS 

S-AOD2 NPWVYLPGVYPRNMRLDSKTATFYNSRPSPHLNGRRAIVPCANILGGGSSINFMMYTR-- 

PmMOD2 NPWVYLPGVYPRNMRLDSKTATFYSSRPSPHLNGRRAIVPCANILGGGSSINFMMYTRGF 

CbAOD1 NPWVYLPGIYPRNMRLDSKTATFYNSRPSKHLNGRRAIVPQANILGGGSSINFMMYTRAS 

PpAOX1 NPWVYLPGIYPRNMKLDSKTASFYTSNPSPHLNGRRAIVPCANVLGGGSSINFMMYTR-- 

PpAOX2 NPWVYLPGIYPRNMKLDSKTASFYTSNPSPHLNGRRAIVPCANILGGGSSINFMMYTRGF 

********:*****:******:**.:.** ********* **:*********:**** 

S-AOD1 --ASDYDDWEMEGWKTDDLLPLMKKLETYQRPCNNRDLHGFDGPIKVSFGNYTYPHCQDF 

PmMOD1 TSASDYDDWESEGWTTDELLPLMKKIETYQRPCNNRELHGFDGPIKVSFGNYTYPNGQDF 

FS-AOD --ASDYDDWESEGWTTDELLPLIKKIETYQRPSNNRDLHGFDGPIKVSFGNYTYPTCQDF 

HpMOX --ASDYDDWESEGWSTDELLPLIKKIETYQRPCNNRDLHGFDGPIKVSFGNYTYPTCQDF 

S-AOD2 ASASDYDDWETEGWTTDDLLPLMKKLETYQRPSNNRELHGFDGPIKVSFGNYTYPNCQDF 

PmMOD2 TSASDYDDWESEGWTTDELLPLMKRLETYQRPCNNPDLHGFDGPIKVSFGNYTYPNCQDF 

CbAOD1 --ASDYDDWESEGWTTDELLPLMKKFETYQRPCNNRDVHGFDGPIKVSFGNYTYPQCQDF 

PpAOX1 GSASDYDDFQAEGWKTKDLLPLMKKTETYQRACNNPDIHGFEGPIKVSLGNYTYPVCQDF 

PpAOX2 TSASDYDDFEAEGWKTKDLLPLMKKTETYQRACNNPEIHGFEGPIKVSFGNYTYPVCQDF 

******:: ***.*.:****:*: *****..** ::***:******:****** *** 

S-AOD1 L--RAAESQGIPFVDDAEDLTTAHGAEHWLKWINRDLGRRSDSAHAYIHPTMRNKQNLYL 

PmMOD1 FTIRAAESQGIPFVDDAEDLKCSHGAEHWLKWINRDLGRRSDSAHAYIHPTMRNKQNLFL 

FS-AOD LR--AAESQGIPVVDDLEDFKTSHGAEHWLKWINRDLGRRSDSAHAYIHPTMRNKQSLFL 

HpMOX LR--AAESQGIPVVDDLEDFKTSHGAEHWLKWINRDLGRRSDSAHAYVHPTMRNKQSLFL 

S-AOD2 L--RAAESQGIPFVDDAEDLTTSHAAEHWLKWINRDLGRRSDAAHAYIHPTMRNKQNLYL 

PmMOD2 FTLRAAESQGIPFVDDAEDLKTSHASQHWLKWINRDLGRRSDAAHAYIHPTMRNKSNLYL 

CbAOD1 LR--ACETQGIPYVDDLEDLKTSHGAEQWLKWINRDFGRRSDTAHAFIHSTMRNKENLFL 

PpAOX1 LRASES--QGIPYVDDLEDLVTAHGAEHWLKWINRDTGRRSDSAHAFVHSTMRNHDNLYL 

PpAOX2 FTLRATESQGIPYVDDLEDLVTAHGAEHWLKWINRDTGRRSDSAHAFVHSTMRNHDNLYL 

: **** *** **: :*.:::******** *****:***::*.****:..*:* 

S-AOD1 ITST--KCDKVIIENGTAVAVKTVPMKPTGS-PKTAVARTYRARKQIVVSCGTISSPLVL 

PmMOD1 IFTTSTKCEKIIIENGVATGIKTVPMKPTGS-PKTQVARTFKARKQIIVSCGTISSPLVL 

FS-AOD ITST--KCDKVIIEDGKAVAVKTVPMKPLN--PKKPVSRTFRARKQIVISCGTISSPLVL 

HpMOX ITST--KCDKVIIEDGKAVAVRTVPMKPLN--PKKPVSRTFRARKQIVISCGTISSPLVL 

S-AOD2 ITST--KADKIIIEDGVAVAVKTVPTKPVGGASAKSHGRTYRAKRQIIISSGTISSPQIL 

PmMOD2 IFTTSTKADKVIIEDGVAAGIQVVPSKPLN--PEKPAAKIYKARKQIILSCGTISTPLVL 

CbAOD1 MTNT--KVDKVIIEDGRAVAVRTVPSKPIGD---SKVSRTFKARKQIVVSCGTVSSPMVL 

PpAOX1 ICNT--KVDKIIVEDGRAAAVRTVPSKPLNP--KKPSHKIYRARKQIVLSCGTISSPLVL 

PpAOX2 IFTCNTKVDKIIVEDGRAAAVRTVPSKPLNA--KKPTHKVYRARTQIVLSCGTISSPLVL 

: . * :*:*:*:* *..::.** ** . . : ::*: **::*.**:*:* :* 

S-AOD1 QR--SGIGAAHKLRQVGIKPIVDLPGVGLNFQDHHCFFTPYYVKPDVPTFDDFVRGDKAV 

PmMOD1 QRSFTGIGSAHKLRQVGIKPVVDLPGVGMNFQDHYCFFTPYHVKPDTPSFDDFVRGDKAV 

FS-AOD QR--SGIGAAHHLRSVGVKPIVDLPGVGENFQDHYCFFTPYHVKPDVPTFDDFVRGDPVA 

HpMOX QR--SGIGAAHHLRSVGVKPIVDLPGVGENFQDHYCFFTPYYVKPDVPTFDDFVRGDPVA 

S-AOD2 ERS--GVGSAHKLRQAGIKPIVDLPGVGENFQDHYCYFVPYHVKPDTPSFDDFVRGDKET 

PmMOD2 QRSGFTIGSAHKLRQAGIKPIVDLPGVGMNFQDHYCFFTPYHVKPDTPSFDDFARGDKAV 

CbAOD1 QRS--GIGEPSKLRAAGVKPIVELPGVGRNFQDHFCYFVPYRIKQDSESFDAFVSGDKEA 

PpAOX1 QRS--GFGDPIKLRAAGVKPLVNLPGVGRNFQDHYCFFSPYRIKPQYESFDDFVRGDAEI 

PpAOX2 QRSGFTFGDPIKLRAAGVKPLVNLPGVGRNFQDHYCFFSPYRIKPQYESFDDFVRGDANI 

:* .* . :** .*:**:*:***** *****.*:* ** :* : :** *. ** 

S-AOD1 QKS--AFDQWYANKDGPLTTNGIEAGVKIRPTEEELATADDEFRAAYDDYFGSKPDKPLM 

PmMOD1 QKSAFTFDQWYANKDGPLTINGIEAGVKIRPTEEELATADDEFRAAYDDYFGNKPDKPLM 

FS-AOD QKS--AFDQWYSNKDGPLTTNGIEAGVKIRPTDEELATADEDFRQGYADYFENKPDKPLM 

HpMOX QKA--AFDQWYSNKDGPLTTNGIEAGVKIRPTEEELATADEDFRRGYAEYFENKPDKPLM 

S-AOD2 QTAAFN--QWYANKDGPLTTNGIEAGVKIRPTEEELSTADDDFKDGWFSYFENKPDKPLM 

PmMOD2 QKSAFFTDQWYANKDGPLTTNGIEAGVKIRPTAEELATADEDFQLGYASYFENKPDKPLM 

CbAOD1 QKS--AFDQWYATGAGPLATNGIEAGVKIRPTEAELATADKAFQQGWESYFENKPDKPLM 

PpAOX1 QKRLFD--QWYANGTGPLATNGIEAGVKIRPTPEELSQMDESFQEGYREYFEDKPDKPVM 

PpAOX2 QKKVFFTDQWYANGTGPLATNGIEAGVKIRPTPEELSQMDESFQEGYREYFEDKPDKPVM 

*. ***:. ***: ************ **: *. *: .: .** .*****:* 

S-AOD1 HYSVISG--FFGDHTKIPNGKFMTMFHFLEY 

PmMOD1 HYSLIFTSGFFGDHTKIPNGKYMCMFHFLEY 

FS-AOD HYSVISG--FFGDHTKIPNGKFMTMFHFLEY 

HpMOX HYSVISG--FFGDHTKIPNGKFMTMFHFLEY 

S-AOD2 HYSLISG--YFGDHTKIPHGKYMTMFHFLEY 

PmMOD2 HYSLISFTGFFGDHTKIPNGKYMTMFHFLEY 

CbAOD1 HYSVISG--FFGDHTRLPPGKYMTMFHFLEY 

PpAOX1 HYSIIAG--FFGDHTKIPPGKYMTMFHFLEY 

PpAOX2 HYSIIAFTGFFGDHTKIPPGKYMTMFHFLEY 

***:* :*****::* **:* ******* 

Figure 9 Alignment of deduced amino acid sequences of partial AOD gene(s) of the methylotrophic 

yeast C. sithepensis sp. nov. (S-AOD1 and S-AOD2) and P. siamensis sp. nov. (FS-AOD) 

with that of Pichia methanolica (PmMOD1, PmMOD2), P. pastoris (PpAOX1, PpAOX2), P. 

angusta (HpMOX) and Candida boidinii (CbAOD1). Identical and conserved residues are 

shown as *

134 


This work was supported by the JSPS- 

NRCT Kasetsart University and Yamaguchi 

University Core University Program (Microbial 

Resources), Kasetsart University Research and 

Development Institute (KURDI) and National 

Research Council of Thailand (NRCT). 


Bergmeyer, H.U. 1955. Zur Messung von 

Katalase-Aktivitaten. Biochem. Z. 327: 255- 

258. 


Figure 10 Similarity of the AOD gene(s) of the methylotrophic yeast C. sithepensis sp. nov. and P. 

siamensis sp. nov. The tree was constructed by UPGMA analysis. 

Bradford, M.M. 1976. A rapid and sensitive 

method for the quantification of microgram 

quantities of protein utilizing the principle of 

protein-dye binding. Anal. Biochem. 72: 248- 

254. 

Cereghino, J. L. and J. M. Cregg. 2000. 

Heterologous protein expression in the 

methylotrophic yeast Pichia pastoris. FEMS 

Microbiol. Rev. 24: 45-66. 

Demain, A. L., H. J. Phaff, C. P. Kurtzman. 1998. 

The industrial and agricultural significance of 

yeasts, pp. 13-20. In C.P. Kurtzman and J. W. 

Fell (eds.). The Yeasts: A Taxonomic Study, 

4 th ed. Elsevier, Amsterdam.

Douma, A. C., M. Veenhuis, W. de Koning, M. 

Evers and W. Harder. 1985. Dihydroxyacetone 

synthase is localized in the peroxisomal matrix 

of methanol grown Hansenula polymorpha. 

Arch. Microbiol. 143: 237-243. 

Faber, K. N., W. Harder, G. Ab and M. Veenhuis. 

1995. Review: Methylotrophic yeasts as 

factories for the production of foreign 

proteins. Yeast. 11: 1331-1344. 

Fukui, S., S. Kawamoto, S. Yasuhara and A. 

Tanaka. 1975. Microbody of methanol-grown 

yeasts. Eur. J. Biochem. 59: 561-566. 

Gellissen, G. 2000. Heterologous protein 

production in methylotrophic yeasts. Appl. 

Microbiol. Biotechnol. 54: 741-750. 

Goodman, J. M. 1985. Dihydroxyacetone synthase 

is an abundant constituent of the methanolinduced 

peroxisome of Candida boidinii. J. 

Biol. Chem. 260: 7108-7113. 

Houard, S., M. Heinderyckx and A. Bollen. 2002. 

Engineering of non-conventional yeasts for 

efficient synthesis of macromolecules: the 

methylotrophic genera. Biochime. 84: 1089- 

1093. 

Koutz, P., G.R. Davis, C. Stillman, K. Barringer, 

J. Cregg and J. Thill. 1989. Structure 

comparison of the Pichia pastoris alcohol 

oxidase genes. Yeast 5: 167-177. 

Ledeboer, A.M., L. Edens, J. Maat, C. Visser, J.W. 

Bos and C.T. Verrips. 1985. Molecular cloning 

and characterization of a gene coding for 

methanol oxidase in Hansenula polymorpha. 

Nucleic Acids Res. 13: 3063-3068. 

Limtong, S., N. Srisuk, W. Yongmanitchai, H. 

Kawasaki, H. Yurimoto, T. Nakase and N. 

Kato. 2004. Three new thermotolerant 

methylotrophic yeasts, Candida krabiensis sp. 


nov., Candida sithepensis sp. nov., and Pichia 

siamensis sp. nov., isolated in Thailand. J. 

Gen. Appl. Microbiol. 50: 119-127. 

Nakagawa, T., T. Mizumura, H. Mukaiyama, T, 

Miyaji, H. Yurimoto, N. Kato, Y. Sakai and 

N. Tomizuka. 2002. Physiological role of the 

second alcohol oxidase gene MOD2 in the 

methylotrophic growth of Pichia methanolica. 

Yeast 19: 1067-1073. 

Nakagawa, T., T. Uchimura and K. Komagata. 

1996. Isozymes of methanol oxidase in a 

methanol-utilizing yeast, Pichia methanolica 

IAM 12901. J. Ferment. Bioeng. 81: 498- 

503. 

Sakai, Y. and Y. Tani. 1992. Cloning and 

sequencing of the alcohol oxidase-encoding 

gene (AOD1) from the formaldehydeproducing 

asporogenous methylotrophic 

yeast, Candida boidinii S2. Gene 114: 67-73. 

Schutte, H., J. Flossdorf, H. Sahm and M.R. Kula. 

1976. Purification and properties of 

formaldehyde dehydrogenase and formate 

dehydrogenase from Candida boidinii. Eur. 

J. Biochem. 62: 151-160. 

Tani, Y., Y. Sakai and H. Yamada. 1985. Production 

of formaldehyde by a mutant of methanol 

yeast, Candida boidinii S2. J. Ferment. 

Technol. 63: 443-449. 

Veenhuis, M. and W. Harder. 1987. Metabolic 

significance and the biogenesis of microbodies 

in yeasts, pp. 436-458. In M. D. Fahimi and 

H. Sies (eds), Peroxisomes in Biology and 

Medicine. Springer-verlag, New York. 

Volfova, O., Z. Zizka and M. Anderova. 1992. 

Effect of increasing methanol concentrations 

on physiology and cytology of Candida 

boidinii. Folia Microbiol. 37: 413-420.


Characterization of Grass Degrading Bacteria Active on 

β-1,3-1,4-D-glucans from Bacillus subtilis GN156 Potential 

Use for Grass Silage-Making 

Jirawan Apiraksakorn, Tonglian Buwjoom and Sunee Nitisinprasert 

ABSTRACT 

One hundred and sixty-one bacterial isolates were screened for (i) the stability of CM-cellulase 

at high temperature of 60°C as primary screening, (ii) the stability of pH and temperature of 3-7 and 

30-60°C, respectively and (iii) the activities of pH and temperature range following stability study. The 

isolate GN156 showed high stability of CM-cellulase activity at the pH and temperature of 3.7 - 7.2 and 

30 – 70°C, respectively. Based on physical and biochemical properties, this isolate was identified as 

Bacillus subtilis. The enzyme system study revealed various hydrolytic enzymes of CM-cellulase, 

dextrinase, cellobiase, xylanase, polygalacturonase, polymethylgalacturonase, but, β-1,3-1,4-glucanase 

was the most effective enzyme. Therefore, optimum pH and temperature of β-1,3-1,4-glucanase were 

further studied. Interestingly, its activities appeared at wide range of pH and temperature of 5.5-9 and 

40-60°C, respectively. The profile of growth and enzyme production indicated that β-1,3-1,4-glucanase 

produced by B. subtilis GN156 was associated with cell growth. Induction of β-1,3-1,4-glucanase 

production by 1% of CM-cellulose, pectin and xylan revealed an increment of activities of 47, 41 and 

11-folds, respectively. When various concentrations of CMC were taken into account, the CMC 

concentration of 0.8% (w/v) provided the maximum β-1,3-1,4-glucanase production. 

Key words: Bacillus subtilis, β-1,3-1,4-glucanase, silage, hydrolytic enzyme 

INTRODUCTION 

Since dairy product and premium meat 

quality demand have been increased, livestock 

farming in Thailand was therefore growing rapidly. 

However, lacking of grass roughage in dry period 

is a big problem. Thus, both good quality and large 

amount of grass commonly used as roughage are 

needed. Therefore, grass storing in a form of silage 

has been proposed to be benefit in both 

preservation and nutritional improvement. 

Silage is a fermented feed under 

anaerobic condition. During ensiling, water– 

soluble carbohydrates (WSC) in the crop is 

consumed by lactic acid bacteria to produce lactic 

acid and some acetic acid that causes the pH of 

ensiled material decreased. Consequently, spoilage 

microorganisms in silage are also inhibited 

(Henderson, 1993). The major problem of grass 

ensiling in Thailand is lack of sufficient amount 

of suitable WSC. A shortage of sugar causes poor 

fermentation due to lactic acid reduction, increase 

in pH, ammonia and also other fermentation 

products (Stetälä, 1988). Generally, the tropical 

Department of Biotechnology, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fagisnn@ku.a.c.th 


grass contains mainly cellulose, hemicellulose and 

other carbohydrates which can not be metabolized 

by lactic acid bacteria. To improve fermentability, 

it is necessary to breakdown carbohydrates into 

soluble and consumable form. Addition of 

enzymes to forage is becoming more common for 

the purpose of enhancing the WSC in the ensiling 

process (Stefanie et al., 2000). Application of 

hydrolytic enzymes from fungi have been 

previously studied. Jaakkola (1990) found that 

addition of enzyme mixture of cellulase, 

hemicellulase and glucosidase from different 

fungal sources did not affect cell wall degradation. 

However, when more cellulase concentration was 

added, the degadation activity increased. In 

contrast, commercial hydrolytic enzymes from 

Acremonium sp. applied to various varities of grass 

in Thailand did not improve silage quality 

(Ohmomo, 1995). It is possible that the proper 

enzyme mixture has the potential to improve 

degradation activity. 

Therefore, the objective of this study was 

to isolate the effective hydrolytic enzymes 

producing bacteria and to characterize the physical 

and chemical properties of interesting enzyme 

being important to grass degradation. 


Sample sources 

The samples for screening of hydrolytic 

enzyme producing microorganisms obtained from 

various sources in North, Northeast and the Middle 

of Thailand were grass, silages, soil and cow feces, 

Screening technique 

A 10 % (w/v) sample in normal saline 

(0.85% NaCl) was 10-fold serially diluted and 

spread on NA medium containing 1% 

Carboxymethyl cellulose (CMC). After incubation 

at 37°C for 24 h, colonies with clear zone were 

primarily screened and later grown individually 

in NB medium containing 1% CMC under aerobic 


conditions at 150 rpm for 18-20 h at 37°C. Cell 

free culture containing enzyme components (CFC) 

was further tested for high temperature stability. 

Enzyme sample was incubated at 60°C for 60 min. 

Each of 2 ul treated enzyme or untreated one was 

transferred as a droplet on 1.5 % agar plate 

containing 1% CMC and then let it dry. The 

enzyme reaction was performed by incubating at 

50°C for 20 min. Then, the reaction plate was 

rinsed with 1 M Tris –HCl pH 7.5, stained with 

1% Congo red for 10 min and destained with 1 M 

NaCl. The stability was determined as equal clear 

zone appearance of temperature treated and 

untreated sample. 

Secondary screening was investigated 

according to enzyme activity and stability in a wide 

range of pH and temperature from selected isolates 

of primary screening. The effects of pH on enzyme 

activity and stability were studied in citrate 

phosphate buffer pH 3-7, while the effects of 

temperature were from 30-60°C. 

Identification of selected isolate 

Morphology of 24 h cell culture on NA 

plate was studied by a light microscope (Olympus 

Co., Japan). Gram staining was followed by the 

method of Beisheir (1991). Spore-forming was 

determined by the method described by Gerhardt 

et al. (1981). 

Biochemical test was performed by using 

API 20 E and API 50 CHB Kit, the results were 

analysed by APILAB Plus program version 2.1 

(Biomericux Sa Co., France). 

Enzyme production 

The selected isolate was grown in 50 ml 

NB medium in 250 ml flask under aerobic 

conditions at 150 rpm for 18-20 h at 37°C. 1% 

innoculum (v/v) was transferred into 100 ml of 

NB, which contained 1% (w/v) inducer. After 24 

h of incubation, the culture was centrifuged at 

4°C, 11,000g for 15 min and the supernatant was 

stored at -20°C for further study.

138 

Determination of enzyme activities 

Carboxymethyl cellulase activity (CMcellulase) 

was detemined by the modified method 

of Okeke and Obi (1995) by performing the 

reaction mixture of 0.1 ml of sample and 0.1 ml 

of 1 % (w/v) CMC (Sigma) in 50 mM citrate 

phosphate buffer pH 5.5 at 50°C for 20 min. The 

amount of reducing sugar released was determined 

by Dinitrosalicylic acid (DNS) method (Miller, 

1959). One unit of enzyme was defined as the 

amount of enzyme that released glucose equivalent 

to 1 µmol of glucose per min. 

β-1,3-1,4-glucanase activity was 

measured using the same procedure as 

carboxymethyl cellulase activity. However, 1 % 

(w/v) of barley β-glucan (Sigma) was used as 

substrate. One unit of enzyme was defined as 

mentioned above 

Xylanase activity was measured by the 

same method described by the modified method 

of Okeke and Obi (1995), but 1% (w/v) of oat spelt 

xylan (Sigma) was used as substrate. One unit of 

enzyme was defined as mentioned above. 

Pectin methylesterase activitiy was 

measured using the modification method of Huang 

and Mahomey (1999) by titration of carboxyl 

groups released from pectin (Sigma) at 50°C for 

20 min. Approximately 0.5 ml of enzyme sample 

was mixed with 4.5 ml of 0.5% pectin in 50 mM 

cirtrate phosphate buffer pH 5.5. The acid released 

was titrated with 0.01 N NaOH. One unit of pectin 

methylesterase was defined as the amount of 

enzyme which released 1 mmole of carboxyl 

groups per min under the assay condition. 

Pectin lyase activitiy was determined 

according to the modification method of Huang 

and Mahomey (1999). 0.1 ml enzyme sample was 

mixed with 0.4 ml of 0.5 % pectin in 50mM cirtrate 

phosphate buffer pH 5.5, incubated at 50°C for 20 

min. 4 ml of 0.01 M HCl was added and 

absorbance was measured at 235 nm. One unit of 

pectin lyase was defined as the amount of enzyme 

which released 4-5 unsaturated trans-elimination 


products showing absorbance of 0.2 at 235 nm. 

Pectate lyase activitiy was determined by 

the same method of pectin lyase activity 

determination, except that 0.5% polygalacturonic 

acid (PGA) (Sigma) was used as substrate. One 

unit of pectate lyase was defined as the amount of 

enzyme which released unsaturated transelimination 

products showing absorbance of 0.2 

at 235 nm. 

Polygalacturonase activity was measured 

by the method of pectate lyase activity assay as 

mensioned above, except that 1% pectin was used 

as a substrate. One unit of enzyme was defined as 

the amount of enzyme that released galactose 

equivalent to 1 µmol of galacturonic acid per min. 

Polymethylgalacturonase activity was 

measured by the method of carboxymethyl 

cellulase activity assay mensioned elsewhere, 

except that 1% PGA was used as substrate. One 

unit of enzyme was defined as the amount of 

enzyme that released reducing sugar equivalent 

to 1 µmol of galacturonic acid per min. 

Protein assay 

Protein concentration was determined by 

the method of Lowry et al. (1951) using bovine 

serum albumin as a standard. 

Determination of grass degradation activity 

Cell-free culture fluid (CFC) containing 

enzyme components was filtrated with 0.2 micron 

filter membrane. One ml of CFC was added to the 

tube containing 1 g of grass powder prepared by 

milling and sieving (mesh number 120). The 

mixture was incubated for 24 h at experimental 

temperatures. Then the extract solution was 

prepared using the method of Ohmomo et al. 

(1995). In brief, 10 ml of distilled water was added 

into the reaction mixture, shaked at 100 rpm for 2 

h and then filtrated through Whatman No.4 paper 

filter. To remove all residual soluble carbohydrates, 

the retentate was washed with 10 ml distilled water 

twice. All filtrates were pooled and adjusted to the

final volume of 25 ml. The complete mixed 

solution was further analyzed of reducing sugar 

content by DNS method (Miller, 1959). 


Screening of thermostable CM-cellulase 

producing bacteria 

A total of 161 islolates were primarily 

screened from 15 samples in the areas of North, 

Northeast and the Middle of Thailand. Only 14 

isolates of AM110, AN114, DN138, EN215, 

GN156, GN231, GN232, HN164, HN166, M10, 

MM252, ON271, RN278 and SN280 showing 

thermostable CM-cellulase activity were selected 

for further study. It seemed that all effective 

isolates distributed to every parts of Thailand 

studied. 

Since grass ensiling process occurred at 

the wide ranges of both pH and temperature, 

investigation of CM-cellulase which was active 

at the pH and temperature of 3 – 7 and 30 - 60°C, 

respectively was decided for secondary screening. 

The results are shown in Table 1. Only 7 isolates 

of AN114, EN215, GN156, GN232, HN166, M10 

and SN280 could produce enzymes which showed 

relative activity higher than 60% at temperature 

ranges of 30-60°C. However, only 5 isolates, 

AN114, GN156, HN166, M10 and SN280 were 

stable at the same temperature range. 

To determine the effect of pH to the 

enzyme activity of these 5 isolates, citrate 

phosphate buffer pH 3, 4, 5 and 7 were used to 

perform enzyme reaction. However due to high 

pH of CFC, the pH of reaction mixture changed 

from pH 3, 4, 5 and 7 to 3.7, 4.5, 5.3 and 7.1, 

respectively. Among these 5 isolates, only the 

enzyme from AN114, GN156 and HN166 showed 

higher relative activities than 60%. However, 

GN156 could show higher relative activity for 70 

% at 4 pH levels. Moreover, it was stable with 

relative activity higher than 85 % at the same pH 

range for 1 h. Thus, the isolate GN156 was selected 


for further characterization. 

Considering the sources of 5 isolates 

exhibiting high activity at wide temperature 

ranges, they were isolated from different sample 

sources: nitrifeed, grass, corn silage and rust fungi. 

However, only one isolate, GN156, found 

produced high enzyme activity at wide pH ranges. 

The pH of corn silage usually decreased to 3.5 at 

the end of ensiling. Therefore, changes of pH 

condition during ensiling to extreme pH might 

provide an advantage to obtain an effective strain. 

It was quite common to find CMcellulase 

producing B. subtilis, however, their 

activities were not active at high temperature 

(Malburg et al., 1992). B. subtilis GN156 growed 

well at low temperature of 37°C but being able to 

produce thermostable cellulase. Therefore, it 

would be a promising strain for further study. 

Identification of the isolate GN156 

Morphology 

The colony of GN156 on NA was 

circular, raised, translucent and smooth. It was 

white and butyrous. The edge was undulate. The 

24 h cell culture was Gram positive, rod shape 

and motility. Elliosoidal endospore was found in 

the center of the cell. This indicated that the isolate 

GN156 belonged to a group of rod, spore-forming 

bacteria based on Bergey’s Manual of Systematic 

Bacteriology (Holt et al., 1986). 

Biochemical test 

The biochemical properties and ability 

to utilize various carbon sources were determined 

using API 20 E and API 50 CHB kit, respectively. 

The isolate GN156 could produce betagalactosidase, 

acetoin, catalase, oxidase and 

gelatin hydrolysis. This strain could not use citrate 

as a carbon source and could not reduce nitrate 

neither which disagreed to B. subtilis type strain 

based on Bergey’s Manual of Systematic 

Bacteriology (Holt et al., 1986). Therefore, 

utilization of various carbon sources was further

140 

Table 1 Optimum and stability of pH and temperatures of CM-cellulase from bacterial isolates screened from various sources in Thailand. 

Source Isolate Relative activity of CM-cellulase (%) 

Temperature pH 

Optimum Stability Optimum Stability 

30 40 50 60 30 40 50 60 3.7 4.5 5.3 7.1 3.7 4.5 5.3 7.1 

1 AM110 57.3 74.9 100 100 93.6 91.2 87.7 64.9 39.7 77.3 100 97.2 78.6 100 100 97.1 

AN114 65.9 94.3 100 89.8 102.3 96.6 72.3 60.2 69.9 75.9 95.2 100 72.4 115.9 105.1 95.2 

DN138 100 81.0 0 0 84.5 81.0 0 0 95.7 100 0 0 95.5 95.7 0 0 


2 EN215 87.5 81.0 100 81.0 0 0 0 81.0 0 100 100 100 0 0 114.3 0 

3 GN156 62.4 83.5 100 68.2 72.9 70.6 80.0 60.0 73.0 88.9 100 85.7 91.3 92.9 92.1 114.8 

GN231 56.3 75.0 100 97.3 87.2 87.6 85.1 46.2 14.6 66.9 100 91.04 92.3 102.9 94.9 98.8 

GN232 83.9 72.6 100 75.8 0 0 75.8 0 91.3 100 91.3 100 109.5 91.3 119.0 91.3 

4 HN164 51.5 73.8 100 94.9 92.2 90.8 89.9 58.9 51.53 73.8 100 94.9 96.3 96.4 91.0 101.1 

HN166 82.3 100 96.8 75.8 79.0 90.3 85.5 82.3 68.7 100 92.5 80.6 113.0 80.6 96.8 100 

5 M10 96.2 88.7 100 92.5 96.2 96.2 96.2 88.7 0 0 100 95.5 0 0 109.1 0 

MM252 54.9 76.5 100 99.1 90.6 92.5 93.4 52.9 11.6 60.2 100 87.9 104.6 100.7 101.1 107.2 

6 ON271 57.1 71.4 100 96.4 82.1 81.3 81.7 52.7 26.5 71.3 100 90.1 87.5 106.2 102.2 109.0 

7 RN278 0 0 100 92.5 88.7 92.5 88.7 84.2 0 100 100 100 0 100 100 100 

SN280 70.8 78.1 71.9 100 104.2 97.9 102.1 88.54 50.6 85.5 100 87.9 114.3 102.8 87.9 97.3 

+ Possitive result - Negative result 

1: Nitrifeed, Nakornrachasima Province 2: Corn silage, Rachaburi Province 

3: Corn silage, Nakornrachasima Province 4: Grass, Rachaburi Province 

5: Chaing Mai Province 6: Grass, Nakornpratom Province 

7: Rust fungi

carried out and the result showed that the isolate 

GN156 produced acid from glycerol, L-arabinose, 

ribose, D-xylose, glucose, fructose, mannose, 

inositol, manitol, sorbitol, α-methyl-D-glucoside, 

amygdalin, arbutin, esculin, salicin, cellobiose, 

maltose, melibiose, sacchrose, trehaose, raffinose, 

starch, glycogen and gentiobiose. The 

identification made by API data base correlation 

indicated 96.3 % similarity to Bacillus subtilis. 

Grass degradation at different temperatures 

The cell-free culture fluid (CFC) of 

GN156 exhibited grass degradation activities at 

various temperatures as shown in Figure 1. The 

activities increased 14, 7 and 6-folds at 37°C, 50° 

C and 60°C, respectively comparing to nonenzyme 

added control. The maximum activity 

occurred at 50°C that would be an appropriate 

condition for its enzymatic reaction. 

Enzyme system of Bacillus subtilis GN156 

The CFC from isolate GN156 showed 

clearly the increment of grass degradation. To 

investigate its hydrolysis action, the activities 

against some of polysaccharides, and disaccharides 

substances were performed. The results showed 

the highest activity of 4.61 U/ml on barley βglucan 

(Table 2), while the activities on xylan, 

dextrin, CMC and cellobiose were 0.56, 0.57, 0.05 

and 0.05 U/ml, respectively. However, no activity 

against laminarin was found. Considering pectic 

Reducing sugar (mg/ml) 

4 

3 

2 

1 

0 

37C 

50C 

60C 

control GN156 

Figure 1 Reducing sugars released from grass 

hydrolysis at different temperatures. 


substance degradation, it showed no activities of 

pectin methylesterase, pectin lyase and pectate 

lyase, but, only low activities of polygalacturonase 

and polymethylgalacturonase were detected at the 

concentrations of 0.15 and 0.14 U/ml, respectively. 

The tropical grass components compose of about 

30-40% cellulose, 30-40% hemicellulose, 1-5% 

starch and 1-2% pectin (Statälä, 1989). Various 

activities appearance of CM-cellulase, cellobiase, 

β-glucanase, xylanase, dextrinase and pectinase 

supported grass degradation very well. 

The CFC showed high activity to 

β-glucan having 1,3 and 1,4 glycosidic linkage, 

while it did low activity to β-1,4 glycosidic linkage 

alone of CMC or no activity against β-1,3 

glycosidic linkage alone of larminarin.. Therefore, 

it could be concluded that the CFC of this strain 

was specific to only β-1,4 glycosidic linkages and 

preferred to the structure of 1,3 and 1,4 linkages 

which was defined as action of β-1,3-1,4glucanase 

(Dixon and Webb, 1979). Hence, 

degradation of barley β-glucan was an action of 

β-1,3-1,4-glucanase, but there was no synergistic 

action of 1,4 and 1,3 glycolysis. 

Many bacterial enzymes showed both of 

β-1,3-1,4-glucanase and β-1,3-glucanase activities 

Table 2 Activities of enzyme from isolate 

GN156 on various substrate. 

Substrate Activity 

(U/ml) 

Barley β-1,3-1,4-glucan 4.61 

CMC 0.05 

Xylan 0.56 

Laminarin 0 

Dextrin 0.57 

Cellobiose 0.05 

Pectin methylesterase 0 

Pectin lyase 0 

Pectate lyase 0 

Polymethylgalacturonase (PMG) 0.15 

Polygalacturonase (PG) 0.14

142 

for example recombinant E. coli JM101 (Louw 

and Reid, 1993), recombinant Streptococcus bovis 

JB1 (Ekinci et al., 1997), Bacillus sp. BE1, 

Bacillus sp. FE1, Pseudomonas PE1 and 

Pseudomonas PE2 (Kitamura et al., 2002). B. 

subtilis GN156 hydrolyzed only β-1,4 linkages 

which may left the structure of β-1,3 linkage 

behind. This structure was not altered by the acid 

used for extraction (Muller et al., 1997). The 

oligosaccharides obtained might be useful for 

further prebiotic application in future study. 

Characterization of β-1,3-1,4-glucanase 

The high β-1,3-1,4-glucanase activity 

played a role in hydrolysis of β-1,3-1,4-glucan 

which was a major hemicellulose in grass. 

Therefore, its physical and chemical properties of 

pH and temperature were more interesting to study. 

The effect of pH on β-1,3-1,4-glucanase 

The pH optima of β-1,3-1,4-glucanase 

was determined by conducting the activity assays 

on β-glucan at 50°C in various pH values of 3-12. 

The results are shown in Figure 2. The activity 

curve showed optimal pH of 7 with the secondary 

inflection in the pH range of 8-9. Moscatelli et al. 

Relative activity (%) 

100 

80 

60 

40 

20 

0 


optimal pH 

pH stability 

(1961) found similar results of 2 inflections in 

curve from B. subtilis at pH 6.5-6.6 and 7.1-7.3. 

In addition, higher optimal pH was shown by crude 

enzymes of B. brevis (Louw and Reid, 1993), of 

recombinant Escherichia coli and of Clostridium 

thermocellum (Schimming et al., 1991) at pH 8- 

10. β-glucanase from B. subtilis GN156 similar 

to other Bacillus sp. strains showed pH optimal at 

neutral to alkaline pH with secondary inflection 

shifting to alkaline pH. β-1,3-1,4-glucanase was 

completely inhibited at extreme pH of 3 and 12. 

Considering pH stability, it was found 

that β-1,3-1,4-glucanase was stable at various pH 

ranges from 3-11 at 4°C for 24 h. However, at high 

pH of 12, β-1,3-1,4-glucanase was completely 

inhibited. 

The effect of temperature on β-1,3-1,4glucanase 

The optimum temperature for β-1,3-1,4glucanase 

was observed by incubating at 

temperature of 20-90°C for 20 min. The 

temperature optima was 60°C as shown in Figure 

3. On either side of this optimum temperature, the 

activity declined sharply. Considering its stability 

at 1 h, β-1,3-1,4-glucanase was stable at 20-50°C, 

0 2 4 6 8 10 12 14 

Figure 2 The pH effect on β-1,3-1,4-glucanase activity from isolate GN156 was carried out at 50°C 

for 20 min. (citrate phosphate buffer pH 3-5.5, phosphate buffer pH 7-9, glycine-NaOH buffer 

pH 10-12). pH stability of the enzyme was treated at various pH at 4°C for 24 h. The remaining 

activity was performed by standard condition at pH 5.5, 50°C for 20 min. 

pH

while the enzyme was inhibited at higher 

temperature of 60°C and completely inhibited at 

70-90°C. Moscatelli et al. (1961) found that β- 

1,3-1,4-glucanase from B. subtilis had optimum 

temperatures at 50-60°C but it showed significant 

loss of activity in 5 min at 60°C. However, Louw 

and Reid (1993) reported a thermostable β-1,3- 

1,4-glucanase from B. brevis showing optimum 

temperatures at 65-70°C, while crude enzyme from 

recombinant Escherichia coli containing β-1,3- 

Relative activity (%) 

100 

80 

60 

40 

20 

0 


1,4-glucanase gene from Clostridium 

thermocellum showed the highest activity at 80° 

C (Schimming et al., 1991). It clearly concluded 

that β-1,3-1,4-glucanase from GN156 was not 

thermostable. 

Effect of inducers on β-1,3-1,4-glucanase 

production 

Polysaccharides of CMC, pectin and 

xylan introduced individually to nutrient broth 

0 10 20 30 40 50 60 70 80 90 100 


optimal temp 

temp stability 

Figure 3 Effects of temperature on β-1,3-1,4-glucanase activity from isolate GN156 carried out in 

citrate phosphate buffer pH 5.5, temperature stability of the enzyme treated at various 

temperature for 1 h, the remaining activity performed by standard condition at 50°C for 20 

min. 

Beta-1,3-1,4 glucanase activity (U/ml) 

10 

8 

6 

4 

2 

0 

Activity 

Spf act 

Control CMC Pectin Xylan 

Figure 4 Effects of inducers on β-1,3-1,4-glucanase production. 

10 

8 

6 

4 

2 

0 

Beta-1,3-1,4 glucanase specific activity (U/mg)

144 

were studied for β-1,3-1,4-glucanase induction. 

The results are shown in Figure 4. CMC and pectin 

supplement induced high β-1,3-1,4-glucanase 

activity of 8.5 and 7.6 U/ml which were higher 47 

and 41 folds than the control, respectively, while 

xylan activity was only 10.8 folds. El-Helow and 

El-Ahawany (1999) studied 5 inducers, 

galactomannan, lichenan, pectin, starch and xylan, 

supplementation with pectin enhanced maximum 

β-1,3-1,4-glucanase production. These results 

supported high production of β-1,3-1,4-glucanase 

from the cultivation of strainGN156 having pectin 

as an inducer. Therefore, both pectin and CMC 

were effective inducers for β-1,3-1,4-glucanase 

production. 

To investigate the effect of CMC level 

on β-1,3-1,4-glucanase production, various CMC 

concentrations of 0, 0.2, 0.4, 0.6, 0.8 and 1% of 

nutrient broth (w/v) were studied. The results are 

shown in Figure 5. The increments of CMC 

concentration of 0.2, 0.4, 0.6 0.8 and 1 % induced 

enzyme activities of 2.87, 4.54, 5.19 5.82 and 5.55 

U/ml respectively. At the concentration ranges of 

0-0.8%, higher CMC concentration induced higher 

Beta-1,3-1,4 glucanase activity (U/ml) 

7 

6 

5 

4 

3 

2 

1 

0 

Activity (U/ml) 

Spf act (U/mg) 


0.0 0.2 0.4 0.6 0.8 1.0 

Conc. of CMC (%) 

enzyme activity. However, enzyme activity started 

to decrease when high concentration of 1% was 

used. It was clearly shown that the induction of β- 

1,3-1,4-glucanase production by B. subtilis GN156 

depended on the amount of CMC. The 

concentration of 0.8 % (w/v) CMC provided the 

highest activity. 

Effect of grass on the growth and β-1,3-1,4glucanase 

production 

The growth and enzyme production of 

Bacillus subtilis GN156 in NB medium alone (NB) 

and NB with other carbon sources of grass (NBG) 

and CMC (NBC) were perfomed. The samples 

were taken every 2 h for 24 h to analyse cell 

number and β-1,3-1,4-glucanase activity. The 

results are shown in Figure 6. Cell growth on NBG 

was not different from the NBC and NB during 

10 h. They all reached stationary phase after 4 h. 

However, cell growth from the NB treatment 

started to decline during 12-24 h while the ones 

from both NBG and NBC still grew during 24 h 

and reached 9.4 × 10 9 and 3.5 × 10 9 cfu/g at 24 h, 

respectively. 

Figure 5 Effects of CMC concentration on β-1,3-1,4-glucanase production byBacillus subtilis GN156. 

7 

6 

5 

4 

3 

2 

1 

0 

Beta-1,3-1,4 glucanase Specific activity (U/mg)

0.1-0.39 U/ml of β-1,3-1,4-glucanase 

activity could be detected when the growth reached 

stationary phase at 6 h. Stulke et al.(1993) reported 

that β-1,3-1,4-glucanase is expressed when cells 

enter stationary stage in response to nutrient 

limitation. This aspect was supported by the study 

of Tang et al.(2004) which proposed that βglucanase 

produced by B. subtilis ZJF-1A5 was 

associated partially with cell growth. 

The induction of β-1,3-1,4-glucanase 

activity by CMC exhibited 1.78, 5.12, 5.78 and 

5.96 U/ml while the one by grass were 0.79, 2.91, 

5.64 and 6.48 U/ml after 12, 16, 20 and 24 h, 

respectively. Interestingly, the control exhibited 

only low activities of 0.66 – 0.88 U/ml. It was 

noticed that β-1,3-1,4-glucanase production from 

NBC was higher than NBG during 16 h. However, 

grass induction treatment reached maximum 

production of 6.48 U/ml after 24 h. Both CMC 

and grass induced the higher β-1,3-1,4-glucanase 

activity 6.73 and 7.36 folds. than the control 

Therefore, this kind of induction would increase 

grass degradation, especially, when silage 

fermentation started during 24 h. 

Cell number (CFU/ml) 

1.E+11 

1.E+10 

1.E+09 

1.E+08 

1.E+07 

1.E+06 

1.E+05 

1.E+04 

1.E+03 

1.E+02 

1.E+01 

1.E+00 


0 4 8 12 16 20 24 

Time (h) 

CONCLUSION 

To solve the problem of lacking of 

sufficient amount of WSC in grass ensiling 

process, addition of effective hydrolytic enzymes 

to structural carbohydrates was proposed. GN156 

was screened for the stability of CM-cellulase at 

high temperature and the activity at wide ranges 

of pH and temperature. It was identified as B. 

subtilis. The activities of grass degradation clearly 

increased at various temperatures of 37, 50 and 

60° C. Study on the enzymatic system of B. subtilis 

GN156, exhibited the activities of β-1,3-1,4glucanase, 

CMCase, xylanase, dextrinase, 

cellobiase, polymethylgalacturonase and 

polygalacturonase. Among these enzymes, β-1,3- 

1,4-glucanase provided the highest activity with 

an optimal pH and temperature of 7 and 60°C and 

stability of 3-11 and 20-50°C, respectively. Both 

CMC and grass used as inducers could enhance 

β-1,3-1,4-glucanase production 6.73 and 7.36 

folds comparing to the control. The stability of 

wide ranges of both pH and temperature would be 

an advantage for natural ensiling process. 

Figure 6 Effect of various inducers on growth and -1,3-1,4-glucanase production. 

● Control, ■ CMC, ▲ Grass, —Cell number and ---- Activiy. 

8 

7 

6 

5 

4 

3 

2 

1 

0 

Activity (U/ml)

146 

ACKNOWLEDGEMENT 

This work was supported by Commission 

on Higher Education, Ministry of education, Royal 

Thai Government and Research Grant of Graduate 

School, Kasetsart University, Bangkok, Thailand. 


Beisheir, L. 1991. Microbiology in Practice. 6 th 

ed., Harper Collins Publisher. Inc. New York. 

524 p. 

Collin, C.H., P.M. Lyne and J.M. Grange. 1989. 

Microbiology Method. 6 th ed., Butterworth 

& Co., (publishers) Ltd., Great Britain. 409 

p. 

Dixon, M. and E. C. Webb. 1979. Enzyme. 3 rd ed. 

New York, Academic Press. 950 p. 

Ekinci,M. S., S. I. McCrae and H. J. Flint. 1997. 

Isolation and overexpression of a gene 

encoding extracellular b-1,3-1,4-glucanase 

from Streptococcus bovis JB1. Appl. 

Environ. Microbiol. 63: 3752-3756. 

El-Helow, E.R. and A.M. El-Ahawany. 1999. 

Lichenase production by catabolite 

repression-resistant Bacillus subtilis mutants: 

Optimization and formulation of an agroindustrial 

by-product. Enzyme Microb. 

Technol. 24: 325-331. 

Gerhardt, P., R.G.E. Murray, R. N. Costilow, E. 

W. Nester, W.A. Wood, N. R. Krieg and G.B. 

Phillips. 1981. Manual of Methods for 

General Bacteriology. Washinton: American 

Society for Microbiology. 415 p. 

Henderson, N. 1993. Silage additives. Anim. Feed 

Sci. Technol. 45: 35-56. 

Holt, J.G., P.H.S. Sneath, N.S. Mair and M. E. 

Sharpe. 1986. Bergey’s Manual of 

Systematic Bacteriology. Vol 2. Baltimore : 

Williams & Wilkins. 1599 p. 

Huang, L.K. and R. R. Mahoney. 1999. 

Purification and characterization of an endopolygalacturonase 

from Verticillium albo- 


atrum. J. Appl. Microb. 86: 145-156. 

Jaakkola, S. 1990. The effect of cell wall degrading 

enzymes on the preservation of grass and on 

the silage intake and digestibility in sheep. J. 

Agri. Sci in Finland. 62: 51-62. 

Kitamura, E., H. Myouga and Y. Kamei. 2002. 

Polysaccarolytic activities of bacterial 

enzymes that degrade the cell walls of Pythium 

porphyrae, a causative fungus of red rot 

disease in Porphyra yezoensis. Fisheries Sci. 

68: 436-445. 

Louw, M.E. and S.J. Reid. 1993. Characterization, 

cloning and sequencing of a thermostable 

endo-(1,3-1,4) β-1,3-1,4-glucanase-encoding 

gene from an alkalophilic Bacillus brevis. 

Appl. Microbiol. Biotechnol. 38 : 507-513. 

Lowry, O. H., N.J. Rosembrough, A.L. Farr and 

P.L. Randall. 1951. Protein measurement with 

the folin phenaol reagent. J. Biotechnol. 

Chem. 193: 265-275. 

Malburg, L.M.,Jr., J.M.T. Lee and C.W. Forsberg. 

1992. Degradation of cellulose and 

hemicelluloses by rumen microorganisms, pp. 

127-159. In G. Winkelmann (ed.). Mirobial 

Degradation of Natural Products. VCH 

Publishers, Inc., New York. 

Miller, G.L. 1959. Use of dinitrosalicylic acid 

reagent for determination of reducing sugar. 

Anal. Chem. 31: 426 – 428. 

Moscatelli, E. A., E. A. Ham and E. L. Riches. 

1961. Enzymatic properties of a β-glucanase 

from Bacillus subtilis. J. Biol. Chem. 236: 

2858-2862. 

Müller, A., H. Enaley, H. Pretus, R. McNamee, E. 

Jones, E. McLaughlin, W. Chandley, W. 

Browder, D. Lowman and D. Williams. 1997. 

The application of various protic acids in the 

extraction of (1→3)-β-D-glucan from 

Saccharomyces cerevisiae. Carbohydr. Res. 

299: 203-208. 

Ohmomo, S., N. Katayama, W. Potacharoen, O. 

Tanaka, S. Sirianuntapaiboon and P. 

Atthasampunna, 1995. Screening of lactic acid

acteria suitable for silage-making in tropical 

regions. JARQ. 29: 251-256. 

Okeke, B. C. and S.K.C., Obi. 1995. 

Saccharification of agro-waste materials by 

fungal cellulases and hemicellulases. Biores. 

Technol. 51: 23-27. 

Schimming, S., W.H. Schwarz and W.L. 

Staudenbauer. 1991. Properties of a 

thermoactive b-1,3-1,4-glucanase (lichenase) 

from Clostridium thermocellum express in 

Escherichia coli. Biochem. Biophysic. Res. 

Commun. 77: 447-452. 

Stefanie, J.W.H., O. Elferink, F. Driehuis, J.C. 

Gottschal and S.F. Spoelstra. 2000. Silage 

fermentation process and their manipulation. 

Eletronic Conference on Tropical Silage. 

Available. http://www.fao.org/waicent/ 

faoinfo/agricult/agp/gp/silage/contents. 

HTML. May 21, 2000. 


Stetälä, J. 1988. Enzymes in grass silage 

production. Food Biotechnol. 21: 211-225. 

Stokes, M.R. 1992. Nutrition, feeding, and calves 

: effects of an enzyme mixture, an inoculant, 

and their interaction on silage fermentation 

and dairy production. J. Dairy Sci. 75: 764- 

773. 

Stulke, J., R. Hanschke, M. Hecker. 1993. 

Temporal activation of b-glucanase 

synthesisIn Bacillus subtilis is midiated by the 

GTP pool. J. Gen. Microbiol. 139: 2041- 

2045. 

Tang, X.-J., G.-Q. He, Q.-H. Chen, X.-Y. Zhang 

and M.A.M. Ali. 2004. Medium Optimization 

for the production of thermal stable βglucanase 

by Bacillus subtilis ZJF-1A5 using 

response surface methodology. Bioresource 

Technol. 93: 175-181.


Thermal Ageing of Thermoplastic Elastomeric Natural Rubber-Low 

Density Polyethylene Blends 

ABSTRACT 

Wiwat Suaysom and Wirunya Keawwattana* 

The effect of amount and type of antioxidants on the mechanical properties in the 65/35 NR/ 

LDPE blend was investigated using thermal aging both unstressed and under tensile elongation. The 

ageing of a family of thermoplastic elastomers from blends of natural rubber and low density polyethylene 

was studied using thermal treatments at 60°C for 3 days, and 70°C for 5 days. Thermal ageing of the 

blends of two polymers caused the tensile properties to deteriorate, especially at longer times or higher 

temperatures of ageing. When an antioxidant for rubber or heat or light stabilizer for polyethylene was 

added, thermo-oxidative stability was increased. Mechanical properties also indicated changes due to 

ageing. Samples were prepared with dicumyl peroxide crosslinking agent. The antioxidants used were 

(i) N, N’-diphenyl-1,4-phenylenediamine; DPPD, (ii) Tetrakis[methylene(3,5-di-tert-butyl 

hydroxyhydrocinnamate)]methane; Irganox, and (iii) Tris(2,4-di-tert-butylphenyl) phosphate; Irgafos. 

All of the antioxidants produced considerable improvement in the thermal-resistance. The most effective 

one was the combination of Irganox and Irgafos even though the continuous phase in these blends was 

low-density polyethylene. 

Key words: natural rubber, low-density polyethylene, thermal ageing 

INTRODUCTION 

Blending of two or more existing 

polymers or copolymers is nowadays a widely 

accepted method in order to respond to the demand 

of new materials. Blending of two polymers 

usually gives better properties than those of a 

single polymer. 

The emergence of thermoplastic 

elastomers (TPEs) is one of the important 

developments in the field of polymer science and 

technology in recent years. TPEs are a new class 

of materials which combine the properties of 

vulcanized rubbers with the ease of processability 

of thermoplastics. TPEs based on rubber-plastic 

Department of Chemistry, Faculty of Science, Kasetsart University, Bangkok 10903. 

* correspondent author, e-mail : wirunyak@yahoo.com 

blends have potential for many applications in 

engineering and consumer goods (Walker, 1979 

and Holed et al., 1996). The outdoor applications 

such as automotive window seals and footwear 

are necessary to develop satisfactory stabilization 

systems to ensure prolonged life. Stabilizers are 

required to protect the compounds from thermal 

degradation (at the elevated temperatures used in 

processing) and photo-degradation caused by the 

ultraviolet irradiation (UV) in sunlight (Adam et 

al., 1991 and Cataldo, 2001). Although some 

stabilizers have a wide application and can be used 

with many different polymers, some additives act 

as a stabilizer with one polymer and as a prodegradant 

for another. Therefore, when choosing 


a stabilizing system for a polymer blend, it is 

necessary to know the effect of the stabilizer on 

both polymeric components. Similarly, any 

possible pro-degradant action of any other 

additives must be considered (Bhowmick et al., 

2002 and Sulekha et al., 2004). 

Ideally a stabilizer will have a beneficial 

effect on both component polymers but in many 

cases it will be present to protect one of the 

components only. In this case it is unfortunate 

that it will usually become distributed into both 

polymeric components during compounding and 

subsequent fabrication. This is a common problem 

for all polymer blends, not for the thermoplastic 

elastomers. 

As rubber and plastics are generally 

stabilized with different types of additives. These 

thermoplastic elastomeric compositions consist 

of continuous plastics matrix and dispersed 

rubber domains, it is worth while to investigate 

which type of antioxidants would be most 

efficient in protecting these materials. The effect 

of commercial antioxidants on the thermal 

degradation behavior of a family of TPE based on 

natural rubber/low-density polyethylene blends 

was reported. There are several antioxidants that 

provide the good protection against oxidation 

extending outdoor lifetime. 

The object of this experiment was to 

study the effect of amount and type of antioxidants 

on the mechanical properties in the 65/35 NR/ 

LDPE blen under thermal ageing. 


Materials 

Natural rubber (NR) supplied by 

Sengmui, Thailand had molecular mass 780×10 3 

g/mol, intrinsic viscosity (benzene 30°C m 3 /kg [η] 

= 0.44) and Wallance plasticity 59.0. The low 

density polyethylene (LDPE) supplied by Thai 

Petrochemical Industrial had Melt Flow Index 

(MFI) 40 g/10min. Dicumyl peroxide (DCP) as a 


crosslinking agent was supplied by Aldrich, 

Germany. Pigment (green) was supplied by Lucky 

Four, Thailand. Compounds based on the 

homopolymers but containing DCP were prepared 

to compare the effect of DCP on the individual 

components of the blends. The antioxidants 

selected for investigations were: 

1. DPPD; N, N’-diphenyl-1,4-phenylenediamine, 

an antioxidant for rubber, supplied 

by Aldrich, germany. 

2. Irganox; Tetrakis[methylene(3,5-ditert-butyl-4-hydroxyhydrocinnamate)] 

methane, a 

high molecular weight phenolic antioxidant (1178 

gmol -1 ) as a thermal (processing) stabilizer for 

plastics, supplied by Ciba Specialty Chemicals, 

Switzerland. 

3. Irgafos; Tris (2,4-di-tert-butylphenyl), 

a phosphate stabilizer for protection against 

discoloration and change of physical properties 

caused by excessive heat exposure for 

plastics, supplied by Ciba Specialty Chemicals, 

Switzerland. 

Mixing and vulcanization procedures 

The blends of natural rubber and low 

density polyethylene with compositions were 

given in Table 1. The blends were prepared in a 

Brabender Plasticorder by melt mixing the plastic 

and the rubber at 130°C at rotor speed of 60 rpm 

(Cam Blade) for approximately 7 minutes. The 

polyethylene was melted first and the rubber was 

then added and blended. The antioxidants and 

pigment were mixed at this stage. The curative 

DCP was added at a level of 0.7 part per hundred 

of resin (phr) and the mixing continued until the 

torque increased by 3-4 units. The compounds 

based on the homopolymers containing DCP (NR/ 

DCP) were also prepared. The mixes were sheeted 

with the two-roll mill and kept at room temperature 

for a day. 

To vulcanize the blend, the mixes were 

compressive molded using a hydraulic hot 

press at 170°C, under pressure 17 MPa. The

150 

vulcanization times were calculated from a 

decomposition half-life of DCP. In the present 

study, at the cure temperature of 170°C, the 

decomposition half-life of DCP was approximately 

1.7 min. Consequently, the cure time of about 12 

min was used to achieve approximately 99.2 % 

cure. 

Ageing 

Accelerated thermal ageing tests were 

followed in the present investigation. It was well 

known that natural rubber without any stabilizer 

degraded to a liquid at high temperature or long 

ageing times. The thermal ageing experiment was 

performed in a closed oven using various times 

and temperatures; 60° for 3 days, and 70°C for 5 

days. The aged samples were allowed to rest at 

room temperature for 30 min and the physical 

properties were then measured. 

Mechanical properties testing 

Tensile testing 

The vulcanized samples were cut into 

tensile specimens using the punching machine. 


Table 1 Compositions of the rubber-polyethylene blends. 

Sample Code NR:LDPE DCP DPPD Irganox (phr) Irgafos (phr) 

(phr) (phr) (phr) 

N100C 100:00 0.7 - - - 

N65C 65:35 0.7 - - - 

65N 1 65:35 0.7 0.5 - - 

65N 2 65:35 0.7 0.75 - - 

65N 3 65:35 0.7 1 - - 

65I 1 65:35 0.7 - 0.5 - 

65I 2 65:35 0.7 - 0.75 - 

65I 3 65:35 0.7 - 1 - 

65N 1I 65:35 0.7 0.25 0.75 - 

65N 2I 65:35 0.7 0.5 0.5 - 

65N 3I 65:35 0.7 0.75 0.25 - 

65I 1F 1 65:35 0.7 - 0.5 0.1 

65I 1F 2 65:35 0.7 - 0.5 0.25 

65I 1F 3 65:35 0.7 - 0.5 0.5 

Note All blends contain 0.5 phr of pigment 

The cutting die punched the sample into dumbbellshape 

(Figure 1). Testing was carried out on a 

universal testing machine (Instron model 5569) 

in accordance with ASTM D412-92 

The testing crosshead speed of 500 mm/ 

min was used with a full scale load cell at 1 kN. 

At least 5 specimens were used for each 

measurement. The following tensile properties 

were measured: 100% modulus, 300% modulus, 

tensile strength and elongation at break. 

Calculation 

Moduli were calculated from the equation, 

σ = F/A (1) 

where σ = stress (MPa) 

F = observed force (N) 

A = cross-sectional area of 

unstretched specimen (mm 2 ) 

1. 100 % modulus = stress at 100% 

elongation 

2. 300 % modulus = stress at 300% 

elongation 

3. Tensile strength = stress at rupture of 

specimen

4. The percentage of elongation at break 

The percentage of elongation at break was 

calculated from the equation 

Percentage of elongation = (l-l o)/l o×100 (2) 

where l = observed distance between the 

grips extensometer on the 

stretched specimen (mm) 

l o = original distance between the 

extensometer (mm) 

Hardness 

The hardness of the specimen was 

measured using Shore A hardness tester (Wallace). 

The 6 mm thick specimen was placed on a test 

platform. The indenter was held in a vertical 

position to provide indentations at least 12 mm 

from any edge of the specimen. Five 

measurements were made at different positions on 

the test piece at least 6 mm apart. An average of 

the five measurements was taken as the hardness 

value of the test sample. 

Abrasion 

The 12 mm thick specimens about were 

detected the density by Densimeter. The weight 

of the specimens was measured before and after 

the test friction by DIN abrasion (ZWICK). The 

equation for loss volume is as follows: 


Loss volume = m 1 - m 2 / D (3) 

where m 1 = weight of the specimen 

before testing 

m 2 = weight of the specimen after 

testing 

D = density of the specimen 

Compression set 

The original thickness of the specimens 

was measured. The test specimens were placed 

between the plates of the compression device with 

the spacers on each side, allowing sufficient 

clearance for the bulging of the rubber when 

compressed (Figure 2). 

The specimens were held at 70°C for 22 

hr, then, rested on a poor thermally conducting 

surface, such as wood, for further 30 min before 

making the measurement of the final thickness. 

The calculation of compression set as 

follow: 

C B = [(t o- t i) / (t o- t n)] × 100 (4) 

where: C B = compression set expressed as 

percentage of the original 

deflection 

t o = original thickness of specimen 

t i = final thickness of specimen 

t n = thickness of the spacer bar used 

(4.5 mm) 

Figure 1 Tensile test specimens. Figure 2 Device for compression set test under 

constant deflection.

152 


Mechanical properties 

The results of the tensile tests on the 

conditioned samples containing different amounts 

and types of antioxidants were shown in Figures 

3 and 4, respectively. The effect of DPPD content 

on the tensile properties of the 65/35 NR/LDPE 

blend was shown in the Figure 3. It was noticed 

that the stress-strain curves decreased as the DPPD 

increased. The blend with 0.5 phr of DPPD (65N 1) 

showed the highest values of tensile strength and 

elongation at break, which can be explained that 

the higher levels of antioxidant may retard the cure 

and reduce the efficiency of peroxide. However, 

the use of DPPD as an antioxidant was not 

appropriate for the 65/35 NR/LDPE blend because 

lots of bubbles existed on the sample including 

the change of the color from green to dusky green 

similar to those found in the case of 70/30 NR/ 

LDPE blend (Bhowmick et al., 2001). 

TS (MPa) 

12 

10 

8 

6 

4 

2 

0 


Figure 3 showed the effect of Irganox 

content on the tensile properties of the 65/35 NR/ 

LDPE blend. It was observed that the blend with 

the lowest amount of Irganox (0.5phr, 65I 1) 

showed the highest value of tensile and elongation 

at break. It was then obviously confirmed that the 

tensile properties of blends depended strongly on 

the concentration of Irganox. Comparing to a 

combination of DPPD and Irganox (Figure 4), the 

tensile properties was decreasing with increasing 

DPPD content in the blend. This is accordance with 

what Bhowmick et al. observed in studying the 

effect of stabilizers in photodegradation of 70/30 

NR/LDPE (Bhowmick et al., 2001). 

Figure 4 showed the plot of stress versus 

strain for the 65/35 NR/LDPE blend with and 

without additives, i.e., DPPD and Irganox. It can 

be seen that the tensile strength of the blends 

decreased with the addition of antioxidants. 

Obviously, the blend with 1 phr of DPPD (65N 3) 

showed the lowest value of tensile strength, which 

DPPD Irganox 

0 0.5 0.75 1 

Amount of antioxidant (phr) 

Figure 3 Tensile strength versus the amount of antioxidant in the 65/35 NR/LDPE blend with various 

types of antioxidant.

could be owing to too many bubbles inside the 

sample as discussed previously. But, the elongation 

at break value of the blend increased with 

increasing the amount of Irganox. It should be 

noted that the type of antioxidant had strong effect 

on the tensile properties of the 65/35 NR/LDPE 

blend. 

As a result, it is of interest to say that the 

performance of tensile properties depended 

strongly on the amount and type of antioxidant in 

which all blends with N,N¢-diphenyl-1,4phenylenediamine 

(65N 1, 65N 2, 65N 3, 65N 1I, 

65N 2I, and 65N 3I) presented low tensile strength 

and elongation at break due to too many bubbles 

on the sample. Hence, Irganox was considered to 

be a suitable antioxidant for the 65/35 NR/LDPE 

blend. This is accordance with what Bhowmick 

et al. observed in studying the effect of etabilizer 

inphotodegradation of 70/130 NR/LDPE 

(Bohowmck et al., 2001) 

Table 2 depicted the mechanical 

properties i.e., hardness, compression set, and 

abrasion of the blend with various amounts and 

TS (MPa) 

12 

10 

8 

6 

4 

2 

0 


types of antioxidant. It was found that the 

mechanical properties decreased with increasing 

antioxidant loading. Thus, the concentration of 

antioxidant had strongly effect on the efficiency 

of peroxide as a curing agent in a blend of NR 

with LDPE. However, the antioxidant was required 

to protect the blend from UV, natural sunlight and 

thermal degradation. So, the suitable type and 

amount of antioxidant considered from the 

mechanical properties and the appearance on the 

sample was Irganox with the lowest loading 

(0.5phr.) into the blends. Hence, the 65/35 NR/ 

LDPE blend with 0.5 phr Irganox will be fixed to 

further study in the next section. 

It should be noted here that one of the 

outstanding properties of N100C (pure NR), 

shown in Table 2, was its high tensile strength due 

to strain-induced crystallization. Therefore, the 

tensile strength of natural rubber can not be 

detectable, whereas that of all blends can be 

improved as given in Table 2. Moreover, it can be 

seen that the pure natural rubber had low value of 

compression set, indicating that the rubber will 

0/1 0.25/0.75 0.5/0.5 0.75/0.25 1/0 

DPPD/Irganox (phr) 

Figure 4 Tensile strength versus the antioxidant combination of DPPD and Irganox in the 65/35 NR/ 

LDPE blend.

154 


Table 2 Mechanical properties of various types and amounts of the 65/35 NR/LDPE blend. 

Sample Code Wt% DPPD Irganox Stress at Elongation 100% 300% Hardness Compression Loss volum 

NR (phr) (phr) max load at modulus modulus (shore A) set (%) (cm3 ) 

(MPa) break (%) (MPa) 

N100C 100 - - - - - - 29.16 8.27 - 

N65C 65 - - 9.92 595.30 1.88 3.28 66.52 45.30 0.22 

65N1 65 0.50 - 10.43 718.22 1.74 2.08 63.13 52.50 0.28 

65N2 65 0.75 - 9.23 670.66 1.60 2.76 62.68 52.95 0.28 

65N3 65 1 - 6.06 589.20 1.84 2.94 62.05 55.08 0.29 

65I1 65 - 0.50 10.32 621.32 1.83 3.22 63.98 51.51 0.23 

65I2 65 - 0.75 8.62 595.74 1.74 3.01 63.05 50.63 0.24 

65I3 65 - 1 10.00 624.98 1.77 3.01 62.85 47.93 0.25 

65N1I 65 0.25 0.75 8.89 608.90 1.67 2.93 62.25 44.68 0.25 

65N2I 65 0.50 0.50 8.94 619.78 1.56 2.84 60.77 38.58 0.31 

65N3I 65 0.75 0.25 6.91 563.00 1.55 2.80 56.45 44.54 0.29

quickly regain its original shape after being 

deformed, while poor hardness. However the 

addition of 35wt% LDPE with various amounts 

and types of antioxidant showed the opposite trend. 

Effect of Irgafos on mechanical properties and 

thermal resistance (Fixed 0.5 phr Irgafos) 

As well known, Irganox phenolic 

antioxidants function as oxygen-centered radical 

scavengers to provide processing and long-term 

thermal stability. Irgafos phosphites function as 

hydroperoxide decomposers to provide processing 

stability and color control as depicted in Figure 5. 

The blends with fixed Irganox content 

(0.5 phr) were prepared with various amounts of 

Irgafos (0.1, 0.25, and 0.5 phr). In order to study 

the effect of Irgafos on the mechanical properties 

and thermal resistance in the blend of 65/35 NR/ 

LDPE, tensile properties, hardness, compression 

Figure 5 Schematic of auto-oxidation. (http//www.cibasc.com) 


set, and abrasion were determined and compared 

with the blend without Irgafos (65I 1). Mechanical 

testing under both unaged and thermal ageing was 

undertaken. 

Unageing 

After melt mixing, it was found that all 

blends with various amounts of Irgafos including 

the one without Irgafos (65I 1) showed the same 

appearance that was lack of tackiness, color green 

and no bubble on the sample. Table 3 illustrated 

the relative mechanical properties of the 65/35 NR/ 

LDPE blend with various amounts of Irgafos. In 

polymer blends with Irgafos, their mechanical 

properties, i.e., hardness, modulus, and 

compression set were better than the blend without 

Irgafos (65I 1). In other words, Irgafos was 

believed to be capable of improving the 

mechanical properties in the 65/35 NR/LDPE 

Table 3 Mechanical properties of various amount of Irgafos in 65/35 NR/LDPE blend. 

Sample Irgafos Stress at Elongation 100% 300% Hardness Compression Loss 

Code (phr) max load at break modulus modulus (shore A) set (%) volume 

(MPa) (%) (MPa) (MPa) (cm3 ) 

65I1 - 10.32 621.32 1.83 3.22 63.98 51.51 0.23 

65I1F1 0.1 9.12 524.78 2.02 3.74 67.56 45.55 0.22 

65I1F2 0.25 9.64 575.53 1.91 3.38 68.64 48.66 0.23 

65I1F3 0.5 8.26 554.62 1.82 3.24 67.82 47.06 0.23

156 

blend with 0.5 phr Irganox. However, the 

concentration of Irgafos had no significant affect 

on their mechanical properties. 

Thermal ageing 

For outdoor applications, the prolonged 

exposure (air, sunlight, rain, etc.) resulting in the 

change in elastomer molecule, cannot be avoided. 

These changes were accelerated by oxidation from 

ozone and oxygen in the atmosphere, ultraviolet 

rays, temperature variations, and other 

environmental factors. Nevertheless, the effect of 

thermo-oxidative degradation on mechanical 

properties was determined in this study by varying 

temperatures and times; 60°C for 3 days and 70° 

C for 5 days. 

The effect of Irgafos content on 

mechanical and tensile properties in the 65/35 NR/ 

LDPE blend with and without ageing was given 

in Table 4 and Figure 6, respectively. It was found 

that the tensile properties of the blends decreased 

gradually with increasing the period of ageing time 

and temperature, possibly due to the degradation 

of the polymers, which also led to the reduction in 

the 100%, and 300% modulus, similar trend was 

observed in the case of 70/30 NR/LDPE blend 

(Bhowmick et al., 2002). The hardness of the aged 

blends was higher than that of the unaged. This 

may be attributed to the crosslink density which 

increased due to the thermal ageing, indicating the 

post-vulcanization reaction during ageing 

(Bhowmick et al., 2002). However, it was 

confirmed by the increase of rebounding value 

after thermal ageing. 

Figure 6 depicted the effect of Irgafos 

content on the tensile properties in the blends after 

thermal ageing. It was discovered that the amount 

of Irgafos acting as a secondary antioxidant 

showed no significant effect on the thermooxidative 

stability of the blends. Because the 

thermal degradation was prevented by primary 

antioxidant, not secondary antioxidant. However, 

it was expected to be capable of protecting the 


Table 4 Effect of Irgafos content on mechanical properties in 65/35 NR/LDPE blend with and without ageing. 

Sample Irganox Temp Temp Stress at Elongation at 100% 300% Hardness Compression Loss volume 

Code (phr) (°C) (°C) max load break (%) modulus modulus (shore A) set (%) (cm3 ) 

(MPa) (MPa) (MPa) 

65I1F1 0.1 - - 9.12 524.78 2.02 3.74 67.56 45.55 0.22 

60 3 8.49 547.14 1.88 3.38 68.43 - - 

70 5 8.66 561.92 2.13 3.44 71.74 33.09 0.23 

65I1F2 0.25 - - 9.64 575.53 1.91 3.38 68.64 48.66 0.23 

60 3 9.66 532.84 1.98 3.85 68.60 - - 

70 5 9.45 563.64 2.13 3.67 70.74 32.73 0.22 

65I1F3 0.5 - - 8.26 554.62 1.82 3.24 67.82 47.06 0.23 

60 3 9.00 551.85 1.92 3.46 68.74 - - 

70 5 9.49 573.42 2.11 3.54 72.20 30.89 0.18 

Note All blends contain 0.5 phr of Irganox

TS (MPa) 

12 

10 

8 

6 


4 

0 0.1 0.2 0.3 0.4 0.5 0.6 

Amount of Irgafos (phr) 

Figure 6 Plot of tensile strength versus the amount of antioxidant in the 65/35 NR/LDPE blend with 

and without ageing. 

blends from UV, heat and sunlight. As seen in Table 

4 the values of composition set in the blends with 

Irgafos reduced compared with those without 

Irgafos. That is to say Irgafos was quite necessary 

for the blend of NR and LDPE in order to improve 

the rebound of the blends in both unaged and 

thermal aged blends. 

CONCLUSIONS 

The effect of amount and type of 

antioxidants on the mechanical properties in the 

65/35 NR/LDPE blend was investigated. It was 

found that the type of antioxidants had strongly 

effect on the efficiency of mechanical properties. 

That is, the mechanical properties of the blends 

decreased with increasing the amount of DPPD. 

The blends containing only Irganox showed the 

similar trend. Furthermore, the improvement of 

the mechanical properties and thermal resistance 

can be obtained in the blend with the combination 

of Irganox and Irgafos. 


Adam, C., J. Lacoste and J. Lemaire. 1991. Photooxidation 

of polyisoprene. J Polym Degrad 

unaged 

ageing at 60 C, for 3 days 

ageing at 70 C, for 5 days 

Stab. 32: 51-69. 

Bhowmick, A.K., J. Heslop and J.R. White. 

2001. Effect of stabilizers in photodegradation 

of thermoplastic elastomeric rubberpolyethylene 

blends-a preliminary study. 

Polym. Degrad. Stab. 74: 513-521. 

Bhowmick, A.K., J. Heslop and J.R. White. 2002. 

Thermal, UV- and sunlight ageing of 

thermoplastic elastomeric natural rubberpolyethylene 

blends. J of Material Science 

37: 5141-5151. 

Cataldo, F. 2001. On the ozone protection 

of polymers having non-conjugated 

unsaturations. J. Polym Degrad Stab. 72: 

287-296. 

Holed, G., N.R. Legge, R.P. Quirk and H.E. 

Schroeder. 1996. Thermoplastic Elastomers. 

2nd ed. Hanser/Gardner Pub, Inc., Ohio. 620 p. 

Sulekha, P.B., R. Joseph and K.B. Manijooran. 

2004. New Oligomer-Bound Antioxidants in 

Natural Rubber/ Polybutadiene Rubber and 

Natural Rubber/Stylene –Butadiene Rubber 

Blends. J. Appl Polym Sci. 93: 437-443. 

Walker, B.M. 1979. Handbook of Thermoplastic 

Elastomers. Van Nostrand Reinhold Ltd., 

New York. 45 p.


Morphology and Haemolymph Composition Changes in Red 

Sternum Mud Crab (Scylla serrata) 

Jintana Salaenoi 1 *, Anchanee Sangcharoen 2 , 

Amara Thongpan 3 and Mingkwan Mingmuang 2 

ABSTRACT 

Characteristic differences between the abnormal, red sternum mud crab compared to those of 

normal one were clearly seen. A red sternum mud crab had soft carapace, red chelae and joint, pale 

hepatopancreas and gill, loose muscle, while its haemolymph contained milky unclotted substance. 

SDS-PAGE analysis of protein in haemolymph showed an intense band of oxyhemocyanin (~75 kDa) 

in normal crab, but none in red sternum crab as also confirmed by the absence of 340 nm absorbance of 

oxyhemocyanin. As for trace elements, the content of copper and zinc in haemolymph was found to be 

three and four times higher than those in normal crab but calcium, magnesium and iron were higher in 

red sternum crab. These results indicated that red sternum mud crab had lost oxyhemocyanin and some 

important elements in haemolymph which were necessary to maintain the normal morphological features 

of this mud crab. 

Key words: mud crab, red sternum, oxyhemocyanin, haemolymph, trace elements, morphology 

INTRODUCTION 

Mud crab (Scylla serrata), an economic 

aquatic animal, has been well-accepted as high 

quality food both locally and internationally 

because of their tasty meat and nutritious value. 

Hence, the demand for this aquatic animal 

increases accordingly, especially a soft shell crab 

which has higher price than the regular one and 

more subsequent establishments of several softshell 

crab farms. Changes of enzyme activities, 

epidermal components and trace elements during 

molting stages of mud crabs were investigated to 

understand their molting behavior (Salaenoi, 2004; 

Salaenoi et al., 2004). Collecting a large number 

of mud crab from soft-shell crab farms, we have 

up to 10% of abnormal crabs with red sternum 

among the normal population have been found. 

The high number of these red sternum crab results 

in the low productivity of mud crab, while the 

cause of it is still unknown. There are several 

investigations on the disease of other crustacea 

having the outer appearances similar to the red 

sternum crab. They indicated the possible cause 

of the disease which might may be a parasitic 

dinoflagellate, Hematodinium sp. (Stentiford et 

al., 2002; Pestal et al., 2003). These symptoms 

were collectively called “pink crab disease” (PCD) 

since the haemolymph and muscle of these 

affected crabs assumed pink coloration, while the 

1 Department of Marine Science, Faculty of Fisheries, Kasetsart University, Bangkok 10900, Thailand. 


3 Department of Genetics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, email: ffisjid@ku.ac.th 


meat was irregular in texture. The similar features 

have been ascribed to Hematodinium infections 

in tanner crabs (Chionoecetes bairdi and C. opilio), 

where the term “bitter crab disease” (BCD) was 

called due to its bitter taste when cooked (Meyers 

et al., 1987). Aside from the outer appearance of 

these symptoms, there is no report on the changes 

in composition of trace elements in these sick 

animals. Some trace metals, i.e., copper, zinc, 

manganese, magnesium and iron are essential 

elements that play various physiological roles. The 

lack or excess of these elements in animals cause 

a stress that could be manifested in sublethal 

responses or in causing death (Soegianto et al., 

1999). Hemocyanin was reported to be a major 

protein in the haemolymph of normal crab 

(Terwilliger et al., 1999). The change of blood 

color and characteristic as seen in red sternum crab 

is possible due to the hemocyanin composition. 

However, there is no evidence to conclude that 

red sternum syndrome seen in mud crab is similar 

in all aspects to bitter crab disease or to pink crab 

disease. Hence, the study on changes in 

biochemical components of haemolymph of these 

red sternum crabs which contributed to the 

symptoms could tell more about their conditions 

and how to prevent the disease occurring in crab 

farming. 


Animal preparation 

Red sternum mud crabs, Scylla serrata 

and the normal ones were collected from a 

soft-shell crab farm in Klung District, Chanthaburi 

Province. The animals were transferred to the 

laboratory and kept in an individual aquarium 

containing 24 ppt salinity seawater. Observations 

on their morphology were recorded. Prior to 

experimentation, the animals were anaesthetized 

in cold water at 4 °C for 1 min. Haemolymph 

samples were withdrawn from the sinus at 

the base of the pereiopods and 10% tri-sodium 


citrate was used as anti-coagulant at the ratio of 

5:1. All tissue samples were kept in ice. 

Haemolymph samples were kept at -20 o C for 

further analysis. 

SDS-PAGE 

The diluted blood samples of red sternum 

and normal mud crab were mixed with loading 

buffer (200 mM Tris-HCl pH 6.8, 4% (w/v) SDS, 

25% glycerol, 0.04 %(w/v) bromphenol blue, 12% 

b-mercaptoethanol) at the ratio of 1:2, then boiled 

for 5 min and subjected to 12.5% SDS-PAGE. The 

electrophoresis was performed at constant voltage 

of 100V at room temperature. The protein bands 

were stained with Coomassie Brilliant Blue R-250. 

Oxyhemocyanin absorbance 

The diluted blood samples of both red 

sternum and normal mud crab were measured for 

their absorbances at the wavelength range of 310- 

700 nm using UV-visible spectrophotometer 

(Jasco V550) and subtracted with baseline 

spectrum. The scanning rate was 100 nm/min, 

having bandwidth of 1 nm. 

Determination of trace elements 

Trace elements (Mg, Fe, Cl, Cu, Mn 

and Zn) were determined using the method of 

AOAC (1980). The analysis was done by wet 

ashing, acid hydrolysis under vacuum condition. 

Approximately 0.2 g of tissue sample was 

dissolved in 10 ml of acid mixture (conc. HNO 3, 

conc. H 2SO 4 and conc. HClO 4 at the ratio of 5:1:2) 

in a 75 ml test tube. The sample tube was set in a 

digesting apparatus under a fume hood and heated 

at 180-200 °C until the clear solution appeared. 

After cooling, the solution was diluted with 

deionized distilled water to make a total volume 

of 50 ml. It was thoroughly mixed and left for 

precipitation to occur. The supernatant was 

collected and kept in a 100 ml polyethylene bottle 

with a tight cover. Calcium content was determined 

by using atomic absorption spectrophotometer

160 

(AA-680 ShiMADZU, Atomic Absorption/Flame 

Emission Spectrophotometer, flame : AIR/C 2H 2). 


Morphological observation 

Red sternum and normal mud crabs 

collected from soft shell crab farm were observed 

(Figure 1). Their characteristics are shown and 

described in Table 1 and Figure 1(A-H). 

The observed characters of red sternum 

and normal mud crabs were clearly different. 

Sternum, chelae and joints of red sternum crab 

were red in color while those of the normal ones 

were white to pale yellow (Figure 1 B-D). On the 

contrary, hepatopancreas and gills of the abnormal 

crabs were pale, soft and unshaped but those of 

the normal crabs were rigid (Figure 1 E-F). The 

haemolymph of red sternum crabs was not clotted 

but formed a milky-like substance, while that of 

the normal crabs was colorless or pale blue (Figure 

1 H). The shells of red sternum crabs were softer 

than those of normal crabs, while the muscle was 

loose and had slow motility. They preferred to 

remain on land and hard surface and their eating 


ability was also reduced. Diseased crabs displayed 

signs of acute morbidity, such as drooping limbs 

and mouthparts. When the red sternum crabs 

became severely deteriorated, their legs and 

chelipeds were paralyzed. Molting could not be 

proceeded as usual and they finally died. 

Not only the taste of red sternum mud 

crabs was found to be bitter than the normal ones 

but the symptoms found in mud crab (Scylla 

serrata) were also similar to those of infected crab 

of PCD and BCD. Although Hematodinium sp. 

was found to be the cause of several diseased crabs, 

i.e., the edible crab (Cancer pagurus) (Stentiford 

et al., 2002), tanner crab (Chionoecetes bairdi) 

(Meyers et al., 1987), snow crabs (Chionoecetes 

opilio) (Pestal et al., 2003), the Norway lobster 

(Nephrops norvegicus) (Field and Appleton, 

1995), the velvet swimming crab (Necora puber) 

(Wilhelm and Mialhe, 1996) and blue crab 

(Callinectes sapidus) (Messick, 1994), but other 

parasites, ciliates, barnacles, virus and yeast were 

also reported to be the potential pathogens of crabs 

as well (Stentiford et al., 2003). 

Similar to both PCD and BCD, red 

sternum mud crabs were found more often in 

Table 1 Characteristics and behaviors of red sternum and normal mud crab. 

Organ, tissue, and Normal mud crab Red sternum mud crab 

observed behaviors 

Carapace Hard Soft 

Chelae and joint White Red 

Sternum Hard, white to pale yellow Soft, red 

Haemolymph Clotted after leaving at room temp. for 

1 min, no color to pale blue 

Unclotted, form a milky-like substance 

Hepatopancreas Fresh and rigid Pale and soft 

Gill Fresh, rigid Pale, soft 

Muscle Rigid Loose 

Strength Strong, aggressive Calm, gentle 

Locomotive activity Highly active Limited movement 

Eating behavior Having good appetite Lose appetite 

Preferred habitat Mostly stay in water, only sometimes Only stay on land and hard 

on land surface


Figure 1 A) A soft-shell crab farm in Klung District, Chanthaburi Province B) A normal mud crab (Scylla serrata) C) A red sternum mud crab showing 

the first sign of red sternum symptom with pink sternum D) The late phase of red sternum showing dark red sternum, chelae, and joint E) 

and F) Comparing the inner organs of normal (E) and red sternum mud crab (F) Showing deterioration of the integument, gills, and hepatopancreas 

G) Collecting haemolymph from the base of pereiopods H) Normal transparent haemolymph and milky unclotted haemolymph of a red 

sternum mud crab

162 

warm, relatively shallow and high salinity water 

(Burnett, 1992). The ventral sides of their limbs 

were opaque, bright white color, rather than the 

normal translucent whitish-gray color. Dorsally, 

the carapace was slightly pinkish as opposed to 

the normal orange-tan color. In the case of PCD 

and BCD, Hematodinium consumes oxygen from 

the crab’s blood and tissues which cause it to 

become weak and lethargic (Meyers et al., 1996). 

The haemolymph of PCD and BCD infected crab 

is milky white in color rather than the normal 

translucent light gray. When cooked, the crabmeat 

had a chalky texture and a bitter aspirin-like flavor 

(Meyer et al., 1987). Stentiford et al. (2001) 

reported on the composition of Hematodiniuminfected 

in the Norway lobster (Nephrops 

norvegicus) tissues and suggested that disruptions 

in the normal carbohydrate and amino acid profiles 

of these tissues might cause the bitter taste of 

the meat. However, PCD and BCD had no impact 

on people who ate infected crab, but the parasites 

could be detected under a microscope 

(Meyers et al., 1996). Once crab got infected, 

Hematodinium grew rapidly inside the crab (up to 

10 6 parasites/ml of blood) over the course of 3 

to 6 weeks, the crab’s blood changed to a milkywhite 

color and lost its clotting ability (Stentiford 

et al., 2002). Although the outer appearances of 

red sternum, PCD and BCD were quite similar, it 

could not be conclude that the causes of these 

symptoms would be the same. Further 

investigation to clarify its cause needs to be done. 

SDS-PAGE 

The blood protein of red sternum mud 

crab was compared to that of normal mud crab 

using SDS-PAGE analysis. There were several 

intense protein bands found in the ranges of 66- 

97 kDa in the normal crab haemolymph, while the 

haemolymph of red sternum mud crab had no band 

at all in this range (Figure 2). Since the molecular 

weight of oxyhemocyanin is 75 kDa and this 

substance was the main component of blood 


Figure 2 Lane A) intense protein band of 75 kDa 

in normal haemolymph Lane B) 

disappearance of 75 kDa band in 

haemolymph of late phase red sternum 

symptom Lane C) molecular markers 

Lane D) less intense band of 75 kDa in 

haemolymph of mud crab having first 

sign of red sternum symptom. 

protein found in white shrimp (Penaeus vannamei) 

(Figueroa-Soto and Barca, 1997), the distinct band 

of 66-97 kDa suggested the existence of 

oxyhemocyanin in normal mud crab haemolymph 

and the gradual disappearance of this band (Figure 

2) could be the clear marker of changing in blood 

component of this symptom as shown by the 

unclotted and milky-like substances. 

To confirm this result, spectroscopic 

analysis was used to reveal the different spectra 

of haemolymph of red sternum and normal 

mud crab. In Figure 3, the spectrum of normal 

crab haemolymph showed the maximum 

absorbance at 340 nm representing the absorbance 

of oxyhemocyanin (Terwilliger et al., 1999) 

but it was absent in the red sternum crab 

haemolymph. 

Haemocyanin (Hc) is a copper-

containing protein, capable of transporting more 

than 80% of the O 2 delivered to the tissue (Ainslie, 

1980). It presents at high concentrations in 

crustaceans, comprising 90-95% of the total 

plasma protein in rockpool prawn (Palaemon 

elegans) and white shrimp (Penaeus vannamei) 

(Cariolou and Flytzanis, 1993). Besides 

transporting oxygen, hemocyanin also involves in 

the transport of metals and amino acids (Weeks 

and Rainbow, 1992). Therefore, the lack of 

hemocyanin in the haemolymph of red sternum 

mud crab affects the function of oxygen binding 

and the activities of metals. That is why red 

sternum mud crab prefers to stay on land for 

acquiring more oxygen. Taylor and Spicer (1987) 

investigated that aquatic animals experienced 

hypoxic conditions more often than air breathers 

and would adapt themselves by moving away 

from that environment or raise posture with the 

anterior part of the body above the water surface. 


Figure 3 Spectrum profile of protein absorbance at 310 nm - 660 nm. The maximum absorbance at 

340 nm of normal hacmolymph indicated the existence of oxyhymocyanin compared to the 

absence of this compound in red sternum crab. 

Trace elements 

Trace elements in haemolymph of red 

sternum mud crab were compared to those of 

normal mud crab by using atomic absorption 

spectrophotometer. The differences of major 

elements in both crabs are shown in Table 2. 

Most of the elements measured in both 

crabs were quite different except chloride. The 

contents of calcium, magnesium and iron in red 

sternum crab haemolymph were higher than that 

in normal crab. In contrast to the content of copper, 

manganese and zinc in the red sternum crab were 

lower. Although the increase in calcium content 

was only 8.99 %, magnesium was found to 

increase at 58.68 % in the red sternum mud crab. 

Calcium and magnesium are involved in several 

physiological processes, i.e., nerve conduction, 

muscle contraction and blood coagulation as well 

as activate alkaline phosphatase which is important 

for calcification and mineralization during molting

164 

cycle in crustacean (Chen et al., 2000). Since 

calcium is an important constituent of skeletons 

and many other rigid mechanical structures, the 

nerve of animals received excessive level of 

calcium will make them move slowly, which 

explains the limited movement found in red 

sternum crabs. However, the overall balance of 

ion concentration in cells is important, the excess 

of magnesium content and the increase in calcium 

in red sternum mud crab may affect the function 

of hemocyanin. 

The iron content in red sternum mud crab 

haemolymph was found to be four times higher 

than that in normal crab. Iron is a constituent of 

several intracellular enzyme systems, notably as 

the cytochromes (Rainbow, 1997). Although the 

total amount of iron in organism is not really much, 

the excess iron in haemolymph of red sternum mud 

crab can possibly hinder the enzyme activities. On 

the other hand, copper content in red sternum mud 

crab was three times lower than that in normal crab. 

It suggested that the decrease of copper led to the 

reduction of metabolic function of hemocyanin 

(Mangum, 1992). 

Manganese is important for the 

development of bones. It also functions as the 

activator of enzyme systems, but the connection 

with the deficiency symptoms in crustacean is not 

entirely clear (Rainbow, 1997). Manganese 

content in red sternum crab was found only half 

of that in normal crab. This effect might result in 

the non-rigid carapace of red sternum crab. On 


Table 2 Trace elements content in haemolymph of normal and red sternum mud crab. 

Type of element Trace elements content in haemolymph (ppm) 

Normal crab Red sternum crab 

Calcium 342.15 372.90 

Magnesium 252.68 400.95 

Iron 2.46 10.80 

Chloride 715 728 

Copper 122.55 37.44 

Manganese 6.76 3.44 

Zinc 23.76 5.78 

the other hand, zinc content was four times lower 

in red sternum than in normal crab. The decrease 

of zinc could result in the decrease of enzyme 

activities because zinc is a constituent of many 

important enzymes, including carbonic anhydrase 

and several peptidases which play important roles 

in CO 2 exchange (Bottcher and Siebers, 1993). 

The inefficient functions of these enzymes also 

explained the preferable behavior of red sternum 

mud crabs to live on land. In addition, one of the 

most quantitatively important ions in animals is 

chloride. The results showed that chloride content 

in red sternum mud crab remained at the same level 

as that of normal crab indicating its importance in 

keeping the osmotic balance for the animals. 

CONCLUSION 

The morphological characteristics of red 

sternum mud crabs were distinctively seen by the 

red color of sternum, the milky-like haemolymph 

and the deteriorations of gill, integument and 

hepatopancreas as well as the bitter taste of meat 

and the preferable habitat on land. The 

disappearance of hemocyanin in red sternum mud 

crab was illustrated by SDS-PAGE and confirmed 

by the absence of 340 nm spectrum. Changes of 

trace elements in the red sternum haemolymph 

could be divided into 2 groups. The increase in 

calcium, magnesium and iron and the decrease in 

copper, manganese and zinc, all of which 

contributed to the changes in physiological

functions and apparent behaviors of mud crab. 

This was the first investigation on red 

sternum mud crab. There was still no conclusive 

evidence that the causes of symptoms found in 

mud crab were the same as those of pink crab 

disease (PCD) or bitter crab disease (BCD). 

However, the changes in biochemical components 

of haemolymph and other tissues of red sternum 

crab could help elucidate these symptoms. Further 

investigations on lactate, enzymes, pigments, 

carbohydrate, protein and glycoprotein in red 

sternum crabs would be of great interest. 



Kasetsart University Research and Development 

Institute (KURDI) under the project 

“Biochemical composition of red sternum mud 

crab” (project no. 04113613). 


Ainslie, R.C. 1980. The quantitative role of 

hemocyanin in the respiration of abalone 

(genus Haliotis). J. Exp. Zool. 211: 87-99. 

Association of Official Analytical Chemists. 1980. 

Official Method of Analysis. 13 th ed. 

Association of Official Analytical Chemists. 

Washington D.C. 

Bottcher, K. and D. Siebers. 1993. Biochemistry, 

localization and physiology of carbonic 

anhydrase in the gills of euryhaline crabs. J. 

Exp. Zool. 265: 397-409. 

Burnett, L.E. 1992. Integrated function of the 

respiratory pigment hemocyanin in crabs. 

Am. Zool. 32: 438-446. 

Cariolou, M.A. and C.N. Flytzanis. 1993. Sexspecific 

gene expression in distinct tissues of 

the shrimp Penaeus vannamei. Comp. 

Biochem. Physiol. 106B: 705-716. 

Chen, Q.X., W.Z. Zheng, J.Y. Lin, Y. Shi, W.Z. 

Xie and H.M. Zhou. 2000. Effect of metal 


ions on the activity of green crab (Scylla 

serrata) alkaline phosphatase. Int. J. 

Biochem. Cell Biol. 32: 879-885. 

Field, R.H. and P.L. Appleton. 1995. A 

Hematodinium-like dinoflagellate infection of 

the Norway lobster Nephrops norvegicus: 

observations on pathology and progression of 

infection. Dis. Aquat. Org. 22: 115-128. 

Figueroa-Soto, C.G. and A.M.C. Barca. 1997. 

Purification of hemocyanin from white shrimp 

(Penaeus vannamei Boone) by immobilized 

metal affinity chromatography. Comp. 

Biochem. Physiol. 117B (2): 203-208. 

Mangum, C.P. 1992. Advances in Comparative 

and Environmental Physiology, Blood and 

Tissue Oxygen Carriers. Springer-Verlag, 

Berlin. 

Messick, G.A. 1994. Hematodinium perezi 

infections in adult and juvenile blue crabs 

Callinectes sapidus from coastal bays of 

Maryland and Virginia, USA. Dis. Aquat. 

Org. 19: 77-82. 

Meyers, T.R., J.F. Morado, A.K. Sparks, G.H. 

Bishop, T. Pearson, D. Urban and D. Jackson. 

1996. Distribution of bitter crab syndrome in 

tanner crabs (Chionoecetes bairdi, C. opilio) 

from the Gulf of Alaska and the Bering Sea. 

Dis. Aquat. Org. 26: 221-227. 

__________, T.M. Koeneman, C. Botelho and S. 

Short. 1987. Bitter crab disease: a fatal 

dinoflagellate infection and marketing 

problem for Alaskan tanner crabs 

Chionoecetes bairdi. Dis. Aquat. Org. 3: 

195-216. 

Pestal, G.P., D.M. Taylor, J.M. Hoenig, J.D. 

Shields and R. Pickavance. 2003. Monitoring 

the prevalence of the parasitic dinoflagellate 

Hematodinium sp. in snow crabs 

Chionoecetes opilio from Conception Bay, 

Newfoundland. Dis. Aquat. Org. 53: 67-75. 

Rainbow, P.S. 1997. Ecophysiology of trace metal 

uptake in crustaceans. Est. Coastal Shelf Sci. 

44: 169-175.

166 

Salaenoi, J. 2004. Changes of enzymes activities 

and epidermal components during molting 

stages of mud crab (Scylla serrata Forskal 

1775). Dissertation. Graduate School, 

Kasetsart University, Bangkok. 

_________, M. Mingmuang, A. Engkagul, P. 

Tabthipwon and A. Thongpan. 2004. 

Chitinase and carbonic anhydrase activities 

during molting cycle of mud crab (Scylla 

serrata Forskal 1775). Kasetsart J. (Nat. 

Sci.) 38: 74-82. 

Soegianto, A., M. Charmantier-Daures, J.P. Trilles 

and G. Charmantier. 1999. Impact of copper 

on the structure of gills and epipodites of the 

shrimp Penaeus japonicus (Decapoda). J. 

Crust. Biol. 19: 209-223. 

Stentiford, G.D., E.S. Chang, S.A. Chang and 

D.M. Neil. 2001. Carbohydrate dynamics 

and the crustacean hyperglycaemic hormone 

(CHH): effects of parasitic infection in 

lobsters. Gen. Comp. Endocrinol. 121: 13- 

22. 

____________, M. Evans, K. Bateman and S.W. 

Feist. 2003. Co-infection by a yeast-like 

organism in Hematodinium-infected European 

edible crabs Cancer pagurus and velvet 

swimming crabs Necora puber from the 


English Channel. Dis. Aquat. Org. 54: 195- 

202. 

____________, M. Green, K. Bateman, H.J. 

Small, D.M. Neil and S.W. Feist. 2002. 

Infection by a Hematodinium-like parasitic 

dinoflagellate causes Pink Crab Disease 

(PCD) in the edible crab Cancer pagurus. J. 

Invertebr. Pathol. 79: 179-191. 

Taylor, A.C. and J.I. Spicer. 1987. Metabolic 

responses of the prawns Palaemon elegans 

and P. serratus (Crustacea: Decapoda) to acute 

hypoxia and anoxia. Mar. Biol. Berlin. 95: 

521-530. 

Terwilliger, N.B., L. Dangott and M. Ryan. 1999. 

Cryptocyanin, a crustacean molting protein: 

Evolutionary link with Arthropod 

hemocyanins and insect hexamerins. Proc. 

Natl. Acad. Sci. 96: 2013-2018. 

Weeks, J.M. and P.S. Rainbow. 1992. The effect 

of salinity on copper and zinc concentrations 

in three species of Talitrid amphipods 

(crustacean). Comp. Biochem. Physiol. 

101C: 399-405. 

Wilhelm, G. and E. Mialhe. 1996. Dinoflagellate 

infection associated with the decline of 

Necora puber crab populations in France. Dis. 

Aquat. Org. 26: 213-219.


A Comparison of Rearing Pacific White Shrimp 

(Liptopenaeus vannamei Boone, 1931) in Earthen Ponds 

and in Ponds Lined with Polyethylene 

ABSTRACT 

Onanong Prawitwilaikul 1 , Chalor Limsuwan 1 , 

Wara Taparhudee 2 * and Niti Chuchird 1 

Pacific white shrimp (Liptopenaeus vannamei) were reared in three earthen ponds and three 

ponds lined with polyethylene of the same size. The larvae at post larval stage 15 were stocked at a 

density of 75 shrimp/m 2 . Shrimp were fed with commercial pellet feed only. The water salinity level 

during culturing period was 3-5 ppt. Culturing period was 112 days. The average yield in the earthen 

ponds was 1,470.8±61.7 kg/rai, while in the polyethylene-lined ponds was 1,737.5±208.6 kg/rai. The 

levels of total suspended solids and total ammonia nitrogen of earthen ponds were significantly greater 

than those of the polyethylene-lined ponds (P

168 


This research was conducted at the 

private shrimp farm in Bansang district, 

Prachinburi province that comprised four, 3 rai 

reservoirs and six, 2.5 rai rearing ponds which 

were three earthen ponds and three high-density 

polyethylene lined ponds. An average depth of all 

the ponds was 1.50 m. 

Pond preparation 

Before culturing, all ponds were cleaned 

by plowing and adjusted for suitable bottom level. 

Lime (CaCO 3) was applied at a rate of 200 kg/rai 

in two ponds, while the other two ponds were lined 

with 0.15 mm thick high-density polyethylene 

sheet on the bottom and dikes. The pond water 

was adjusted to a salinity of 5 ppt and an 

application of 1 mg/l Trichlorfon removed the 

disease carriers such as shrimp and crab. One week 

later, 15 kg of molass and 20 l of superphosphate 

fertilizer was applied in each pond in order to 

promote plankton bloom. 

Larval stocking 

Prior to stocking, post larvae 15 (PL 15) 

proved free from Taura syndrome virus (TSV) and 

infectious hypodermal and hematopoietic necrosis 

virus (IHHNV) using polymerase chain reaction 

(PCR) technique. PL 15 were stocked at a density 

of 75 PL/m 2 or 120,000 PL/rai and fed a 35% protein 

of feed. Feed was applied four times a day and feed 

quantity was adjusted following Chanratchakool’s 

method (Chanratchakool et al., 1994). 

Aeration and water exchange 

Four long-armed paddlewheel aerators 

and two spiral type aerators were used in each 

pond. The aerators were powered by four, 11 hp 

diesel motors, and water exchange was started after 

50 days of culture period. Water exchange was 

done at around 10% of pond water volume every 

week. 


Water quality analysis 

Water quality parameters i.e. 

temperature, pH and dissolved oxygen were 

monitored twice (6.30-7.00 and 14.00-15.00) 

everyday using thermometer, pH meter (ORION 

Model sa520, USA) and Do Meter (YSI model 

51B, USA), respectively. Water salinity, alkalinity, 

hardness, total ammonia-nitrogen, nitrite-nitrogen, 

total suspended solids and chlorophyll a were 

measured every week. Water salinity and 

conductivity were measured with a salinometer 

(YSI 30/10 FT, USA). Alkalinity and hardness 

were tested using the Strickland and Parsons 

(1972) method. Total suspended solids and 

chlorophyll a were done following standard 

methods (APHA et al., 1992). 

Plankton Abundance and Identification 

Four liter of water samples were 

collected from each experimental pond at 30 cm 

depth of 2 two opposite positions at 11.00 am. 

Water was mixed and filtrated through a 20 µ 

plankton net. Then water sample was kept in a 250 

ml plastic bottle and preserved with 4% formalin. 

Plankton identification was done following the 

Shirota (1996), Wongrat (1998) and Wongrat 

(1999) using Sedwick-Rafter counting cell. 

Data analysis 

Means and standard deviations of 

production, body weight gain, FCR, survival rate, 

production costs, water quality parameters and the 

amounts of plankton were calculated. Differences 

(P

than that in the earthen ponds (P0.05) except for total 

ammonia nitrogen and total suspended solids of 

earthen ponds which were significantly higher 

(P

170 

DISCUSSION 

The average total ammonia nitrogen was 

high in earthen ponds with some shrimp mortalities 

(10% of total number) during the 5 th - 9 th week of 

culturing. High total suspended solids in earthen 

ponds was due to high erosion associated with 

aeration and rain. The result of less plankton in 

the earthen ponds was a consequence of high total 

suspended solids and therefore limited light 

penetration for photosynthesis (Boyd, 1990). Moss 

(1995) revealed that L. vannamei grew 1.5 to 1.9 

times faster if the water had a high algae content, 

compared to clean water. A high level of ammonianitrogen 

and a low level of natural nutrient were a 

major cause of lower production in earthen ponds 

compared to lined ponds. This present study result 

is similar to the study of Moss (1995) and Pruder 

et al. (1992), who reported that L. vannamei 


Table 3 Water quality of rearing pacific white shrimp in earthen ponds and ponds lined with 

polyethylene. 

Parameter Mean ± S.D. 

Earthen pond HDPE lined pond 


Morning 29.5 ± 0.9 30.0 ± 0.8 

Afternoon 30.7 ± 1.0 31.3 ± 1.1 

pH 

Morning 7.7 ± 0.1 7.7 ± 0.1 

Afternoon 8.1 ± 2.1 8.3 ± 0.1 

Dissolved oxygen (mg/l) 

Morning 4.8 ± 0.7 4.8 ± 0.8 

Afternoon 5.9 ± 1.1 6.3 ± 0.8 

Salinity (ppt) 4.4 ± 1.0 3.7 ± 0.9 

Conductivity (mmhos/cm) 8.1 ± 1.9 6.9 ± 1.6 

Alkalinity (mg/l) 142.8 ± 30.3 134.1 ± 28.5 

Total ammonia-nitrogen (mg/l)* 0.44 ± 0.40 0.35 ± 0.20 

Nitrite-nitrogen (mg/l) 0.61 ± 0.60 0.74 ± 0.64 

Hardness (mg/l) 1,113.8 ± 215.1 947.3 ± 224.5 

Total suspended solids (mg/l)* 262.0 ± 197.0 98.6 ± 36.0 

Chlorophyll a (mg/l) 218.4 ± 137.4 285.3 ± 144.5 

Plankton (cell/l)* 2.6×10 6 ± 2.5×10 6 1.6×10 7 ± 3.6×10 4 

* It is significantly different at P40% of clay) as lining pond with polyethylene is 

rather costly, which is estimated at 30,000 Baht/rai.


Authors would like to thank the Thai 

National Research Institute for supporting the 

fund. 


Anonymous. 2001. HDPE line pond farm 

technology. Technical Breakthrough for 

Enhanced Consistent production in Tiger 

Prawn Farming in Asia. Available source : 

http://www.todayaque.com/HDPE 

(frontpate), May 7,2003. 

American Public Health Association (APHA), 

American Water Works Association (AWWA) 

and Water Pollution Control Federation 

(WPCF). 1992. Standard Method for the 

Examination of Water and Waste Water. 18 

th ed. American Public health Association, 

Washington, D.C. 

Boddeke, R. 1983. Survival strategies of Penaeid 

shrimps and their significance for shrimp 

culture, pp.514-523. In G.L. Rogers, R. day 

and A. Lim (eds.). Proceeding of the First 

International Conference on Warm Water 

Aquaculture-Crustacea. Brigham Young 

University Haweii Campus, Laie, HI. 

Boyd, C.E. 1990. Water Quality in Ponds for 

Aquaculture. Alabama, Agricultural 

Experiment Station, Auburn University, 

Auburn, Alabama. 

Chanratchakool, P., J.F. Turnbull and C. 

Limsuwan. 1994. Health Management in 

Shrimp Ponds. Aquatic Animal Health 

Research Institute, Kasetsart University 

campus, Bangkok. 

Kiertpinyo, P. 2001. Practical Technology for 

Litopenaeus vannamei Culture. Muang 

Kaset Magazine publishing, SamutPrakarn. 


López M., C. Adams, J.C. Cato and D. Sweat. 

2002. Cost and Returns Budgets for an 

Intensive Zero Water-Exchange Shrimp 

Culture Demonstration Project in 

Nicaragua, 2001. Florida Sea Grant College 

Program, University of Florida, Gainesville, 

FL. 

Moss, S.M. 1995. Production of growth-enhancing 

particles in a plastic-lined shrimp pond. 

Aaquculture 132: 253-260. 

Motoh, H. 1981. Studies on the Fisheries Biology 

of the Giant Tiger Prawn Penaeus monodon 

in the Phillipines. Aquaculture Department. 

Southeast Fisheries Department Centre. 

Tigbauan, Iloilo. 

Pruder, G.D., E. O. Duerr, W. A. Walsh, A. L. 

Lawrence and W. A. Bray. 1992. The technical 

feasibility of pond liners or rearing Pacific 

white shrimp (Penaeus vannamei) in terms of 

survival growth, water exchange rate and 

effluent water quality. Aquacultural 

engineering 11(3): 183-201. 

Shirota, A. 1996. The Plankton of South 

Vietnam: Freshwater and Marine 

Plankton. OTCA, Japan. 

Strickland, J.D.H. and T.R. Parsons. 1972. A 

Practical Handbook of Seawater Analysis. 

Fisheries Research Board of Canada, Ottawa. 

Wassenberg, T.J. and B.J. Hill. 1994. Laboratory 

study of the effect of light on the emergence 

behavior of eight species of commercially 

important adult penaeid prawn. Aust. J. Mar. 

Freshwat. Res. 1: 43-50. 

Wongrat, L. 1998. Zooplankton. Faculty of 

Fisheries, Kasetsart University, Bangkok. 

787p. 

Wongrat, L. 1999. Phytoplankton. Faculty of 

Fisheries, Kasetsart University, Bangkok. 

851p.


Application of Near Infrared Spectroscopy 

to Predict Crude Protein in Shrimp Feed 

Jirawan Maneerot 1 , Anupun Terdwongworakul 2 *, 

Warunee Tanaphase 3 and Nunthiya Unprasert 4 

ABSTRACT 

The preliminary study of using near infrared spectroscopy (NIRS) to predict crude protein in 

shrimp feed was carried out. The calibration set of 41 shrimp feed samples was laboratorily prepared 

against the validation set of 39 samples with similar distribution for the chemical content. The samples 

were scanned for the NIR spectra in the wavelength region of 1100 nm to 2500 nm. Both multiple linear 

regression (MLR) and partial least square regression (PLS) were performed to analyze data to create the 

best calibration equations. The results indicated that absorbances at wavelengths of 1686 nm, 1778 nm 

and 2486 nm were related with protein in the samples. At 1686 nm the absorbance was found to 

coincide with that of protein analytically found the most in the prepared shrimp feed. Following the use 

of wavelengths selected in MLR for constraining the wavelength region as input to PLS, the calibration 

equations were shown to improve in prediction accuracy. 

Key words: near infrared spectroscopy (NIRS), protein, shrimp feed 

INTRODUCTION 

Aquatic animals are those of economic 

animals which earn a large amount of exporting 

incomes to Thailand. Thailand was the world 

number one in exporting aquatic animals 

particularly shrimp in 1992. Since 1988, shrimp 

farms have been expanded by 90% (Fishery 

Economics Division, 2002). In shrimp farming 

the feed accounts is the main cost accounting for 

70-80% of the total investment. The feed as a 

result has been produced in an increasing amount 

every year. Since there are a large number of 

shrimp feed products available in the market, 

the quality control must be exercised to standardize 

the feed quality by the concerned authority. 

In quality control of aquatic animal 

feed, the chemical analyses must be applied 

inevitably. The analyses have to be carried out in 

laboratory rooms using a number of instruments 

and both local and imported chemicals. The results 

obtained are accurate and internationally accepted 

but the analyses take time, need a lot of spending 

and employ several personnel. Besides the 

waste from the laboratory room is regarded as 

pollutants. 

1 Department of Fishery Products, Faculty of Fisheries, Kasetsart University, Bangkok, 10900, Thailand. 

2 Department of Food Engineering, Faculty of Engineering at Kamphaeng Saen, Kasetsart University, Kamphaeng Saen, Nakhon 

Pathom 73140, Thailand. 

3 Kasetsart Agricultural and Agro - Industrial Product Improvement Institute (KAPI) ::Kasetsart University, Bangkok, 10900, 

Thailand. 

4 Department of Fisheries, Ministry of Agriculture and Coorperatives, Bangkok, 10900, Thailand. 

* Corresponding author, e-mail: fengant@ku.ac.th 


Near Infrared Spectroscopy (NIRS) has 

been used efficiently in predicting the chemical 

contents in animal feed (De Boever et al., 1994) 

and the Total Digestible Nutrients (TDN) in animal 

feed (Amari and Abe, 1997). De Boever et al. 

(1994) used partial least square regression (PLSR) 

to obtain the calibration model in the range of 1100 

nm to 2500 nm for predicting the protein in feed 

for cow. The accuracy of the prediction was shown 

with correlation coefficient, R = 0.96 and root 

mean square error of prediction (RMSEP) = 1.4. 

Amari and Abe (1997) studied the use of NIRS 

technique to determine the content of TDN in 

silage animal feed and showed that the first term 

in the model developed using multiple linear 

regression (MLR) related to the absorbance at 

2149 nm. Edney et al. (1994) investigated the 

quality of barley to use as animal feed using NIRS. 

They found that in the range 400 – 1800 nm the 

model with second derivative treated spectra could 

be used to predict the animal feed with R = 0.97 

and SEP = 0.31. Iwamoto et al. (1984) reported 

the use of MLR model to predict the protein 

content in wheat flour. In the industry of the 

aquatic animal feed, the NIRS has been found to 

be used the most compared with other industries 

(Warunee et al., 2001). The use of NIRS would 

reduce the chemical expense and shorten the 

analysis time. However the NIRS technique is 

not well-known and still new to Thai research 

community thus necessitating further study and 

development. 

This study was aimed to preliminarily 

investigate the application of NIRS in prediction 

of protein content in the laboratorily prepared 

shrimp feed. The shrimp feed was prepared so as 

to have the protein content varying in similar range 

to that in the commercial feed. The prepared 

shrimp feed with known proportion of raw material 

was used as samples for this initial study. Provided 

the results were promising the extension to the 

commercial feed with large variation of raw 

material will be carried out in the future. 



Shrimp feed for the tests 

Each individual raw material for 

formulating the shrimp feed such as wheat, ground 

fish, for instance was prepared and chemically 

analyzed for protein content. The obtained protein 

values were then used to calculate the proportion 

of each raw materials for preparing 80 samples of 

the shrimp feed with a range of protein between 

12 to 50% i.e. covering the protein range of the 

feed in the market. The protein content in the feed 

was prepared in such a way that the content was 

increased from 12 to 50% in even increment. The 

feed was then dried in the oven at 50°C for 16 

hours and kept at ambient temperature so that the 

moisture content was approximately not over 10%. 

Each sample of the shrimp feed was then ground 

by the mortar grinder to ensure the homogeneity 

of the sample. The sample was kept until the 

temperature was 25°C prior to being subjected to 

scanning by the spectrometer. After that the 

sample was chemically analysed for the protein 

content. Following the scanning and chemically 

analysing all 80 samples, the chemical data was 

used to divide the data into a calibration set of 41 

samples and a validation set of 39 samples with 

both sets having similarity in even distribution of 

the protein content. The samples were sorted in 

descending order of the protein content and each 

sample was assigned alternatively to both 

calibration set and validation set. The calibration 

set contained the first and the last data in the sorted 

order. As a result the calibration set has the protein 

content in a larger range than the validation set. 

Measurement of the spectra 

The samples of the shrimp feed were 

scanned in reflectance mode with NIR 

spectroscopy (Bran&Luebbe InfraAlyzer 500) in 

the near infrared region from 1100 nm to 2500 

nm at 2 nm increment. The samples were packed 

in the standard close cup and each sample was

174 

repacked and scanned three times for the average 

value. The NIR absorbance spectra were stored 

for subsequent development of the calibration 

equation with the reference i.e. the protein data. 

Chemical analyses 

The protein contents of the feed were 

determined by means of Kjeldahl method. The 

measurement was repeated three times for each 

sample and the measured values were averaged. 

The chemical values of the protein in the shrimp 

feed were shown in Table 1. 

Data analyses 

The data analyses were performed based 

on the assumption that other constituents in the 

shrimp feed would have the absorbance in other 

wavelength region than the protein content. 

The samples from the calibration set 

were used to derive the calibration equation. The 

spectra were initially pretreated with second 

derivative in comparison with the multiplicative 

scattering correction (MSC) prior to analyses in 

order to compensate for the particle size and the 

scattering effect. The calibration equations or 

models were developed using multiple linear 

regression (MLR) analysis and partial least square 

regression (PLSR) for comparison. The models 

were validated with data in the validation set. The 

optimum model was the one that yielded the lowest 

value of standard error of prediction (SEP) and 

low value of bias. 

For the multiple linear regression model, 

the software used was the Near Infrared Spectral 

Analysis Software (NSAS) which gave the 

calibration equation with four terms of the spectral 


variables at maximum as follows: 

%Protein = K 0 + K 1F 1 + K 2F 2 + K 3F 3 + K 4F 4 

where K i = coefficients of each term 

F i = function of absorbance at the wavelength i 

Regarding the partial least square 

regression model, the UNSCRAMBLER software 

package was used to derive the PLS models 

relating the spectral bands to the protein content. 

The spectral bands were varied and analyzed for 

the PLS model with minimized errors in 

prediction. The software suggested the number 

of factors, which were optimum in describing all 

variance in the spectra. 


The NIR absorbance spectra of the 

shrimp feed in the calibration with protein content 

between 12.08 and 52.80% are illustrated in 

Figure 1. Clearly, the absorbance were affected 

by variation in particle size as the whole were 

shifted at all wavelengths. Additionally, scattering 

effect was apparent which tilted the spectra to have 

higher asorbance at longer wavelengths. 

The particle size and the scattering effect 

imposed in the absorbance (Figure 1) were 

compensated for by second derivative or 

multiplicative scattering correction as shown in 

Figure 2 and Figure 3 respectively. 

MLR model 

The sample absorbance at each 

wavelength were correlated with the protein 

content and the resulted correlation coefficients 

were plotted against wavelength. The correlation 

plot in Figure 4, used as the tool in selection of 

Table 1 Protein contents in the prepared shrimp feed. 

Protein content 

Calibration set Validation set 

Number of samples 41 39 

A range of protein content 12.80% – 52.80% 13.32% – 50.57% 

Standard deviation 11.68% 11.19%

the first wavelength in the MLR model, showed 

high correlation between the treated absorbance 

and the protein content at various wavelengths at 


1513, 1686, 1750, and 1840 nm for instance. 

To assure the selection of the optimum 

absorbance band, the spectra of the protein type 

Figure 1 Original spectra of shrimp feed for the calibration set having protein content ranging from 

12.08 to 52.80%. 

Figure 2 MSC pretreated spectra of shrimp feed showing reduction of particle size and scattering 

effect.

176 

found the most in the shrimp feed was also 

examined. Representatively the shrimp feed 

samples with the lowest value and the highest 

value of the protein contents (i.e. 19 and 45% 

respectively) were analyzed by the high 

performance chromatography (HPLC). The 

analysis results would indicate the type of protein 


Figure 3 Second derivative of the absorbance of shrimp feed. 

contributing to the maximum proportion. 

It was found from Table 2 that L- 

Glutamic acid was accounted for the maximum 

percentage in both samples. The shrimp feed either 

with low protein content (19%) or high protein 

content (45%) contained L-Glutamic acid as a 

protein type with the maximum percentage. The 

Figure 4 Second derivative correlation plot for wavelength selection of the first term in the MLR 

calibration model.

absorbance region of L-Glutamic acid would be 

useful in selection of the wavelength at which the 

absorbance was prominent. 

Subsequently, 99% L-Glutamic acid was 

measured to obtain the spectrum and its 

corresponding second derivative NIR spectra was 

derived as shown in Figure 5 and Figure 6 

respectively. From Figure 5 the band of the 

spectrum was apparently broad and made it 

difficult to choose for the significant bands of the 

spectrum. 

With the second derivative NIR 

spectrum, the bands were markedly resolved. It 

could be seen in Figure 6 that the important bands 

of the L-Glutamic acid were at 1686, 2150, 2284 

and 2390 nm and these wavelengths closely 

coincided with those proposed by William et al. 

(1990). 

The band at 1686 nm was therefore 

chosen as the first term in the MLR model. 


Analysis results suggested that the optimum MLR 

model included spectra at wavelengths 1686, 1778, 

and 2486 nm with R=0.99, Standard error of 

calibration (SEC) = 1.30 and SEP=0.99. The 

model used the MSC treated spectra that were 

found giving lower errors than the second 

derivative treated ones. In Figure 7 is shown the 

NIR calculated protein contents plotted against the 

actual values of the validation set. 

The absorbance at 1686 nm given by the 

correlation plot in the MLR model was also the 

wavelength of the significant band in the L- 

Glutamic acid. Consequently the best MLR model 

is as follows: 

%Crude protein = 581.39 + 744.02 

F1686 – 1172.06F1778 – 485.33F2486 

where Fi = MSC treated log(1/R) or absorbance 

at wavelength i 

R=Reflectance 

Table 2 The first three of protein types accounting for the maximum percentage in shrimp feed samples. 

Type of protein (w/w) Feed with 19% protein Feed with 45% protein 

L-Glutamic acid (%) 2.88 6.50 

L-Aspatic acid (%) 1.75 3.97 

L-Leucine (%) 1.39 2.82 

Figure 5 Absorbance spectrum of 99% L-Glutamic acid.

178 

PLS model 

The comparison between the PLS models 

established from variably different bands of spectra 

was made. The full band of the spectra was firstly 


Figure 6 Second derivative of NIR spectrum of 99% L-Glutamic acid. 

used to make the PLS model. Secondly the 

minimum and the maximum wavelengths in the 

obtained MLR model were used to define a range 

of wavelengths for the PLS model analysis as 

Figure 7 The plot of NIR calculated protein contents against the actual values of the validation set 

with SEP of 0.99 and bias of –0.31.

suggested by Saranwong et al., 2001). The PLS 

models developed from the above two selections 

of the band were compared. 

The PLS analyses in Table 3 showed that 

using the whole spectrum (1100 to 2500 nm) for 

establishing the PLS model gave the highest error 

in prediction (SEP=1.05) but the bias was low. 

This was probably due to the inclusion of 

interference or unwanted absorbance in the model. 

The second PLS model took the starting wavelenth 

at 1600 nm and the ending wavelength in the range 

at 2500 nm. Both terminal wavelengths were 

empirically taken from the first (1686 nm) and the 

last wavelengths (2486 nm) of the MLR model. 

The accuracy in prediction of the second PLS 

model was slightly improved (SEP = 1.00) over 

the first model. However the bias was poorer at 

–0.34. 

CONCLUSION 

Prediction of protein content in 

laboratorily prepared shrimp feed was achieved 

using NIR spectroscopy. The MLR model was 

optimal when pretreated with MSC and the MSC 

treated spectra at 1686 nm was involved in the 

model. The spectra band at 1686 nm was 

coincident with the dominant band of L-Glutamic 

acid which was the protein type accounting to the 

maximum proportion in the shrimp feed. 

Similarly, MSC treatment resulted in the optimum 

PLS model. Comparative investigation for the 

optimum spectral range for PLS analysis was 

conducted. The results indicated that the range 

between 1600 nm and 2500 nm was the optimum 

range in establishing the PLS model with the 


Table 3 The analysis results of PLS calibrations in comparison with MLR calibration for prediction of 

protein contents. 

Model Range of wavelengths (nm) Fa R SEC SEP Bias 

PLS 1100 - 2500 2 0.99 1.29 1.05 -0.19 

PLS 1600 – 2500 2 0.99 1.32 1.00 -0.34 

MLR 1686, 1778 and 2486 3 0.98 1.30 0.99 -0.31 

a Number of factors used in the PLS calibration or the number of terms in the MLR calibration. 

lowest error of prediction. 


Department of Fisheries, Ministry of 

Agriculture and Coorperatives is gratefully 

acknowledged for the support of the raw material 

used in the research. The authors are also grateful 

to Kasetsart Agricultural and Agro - Industrial 

Product Improvement Institute ( KAPI ) Kasetsart 

University, for provision of the instrument and the 

laboratory space. 


Fishery Economics Division. 2002. Fishery 

Statistics of Thailand, Department of 

Fisheries, Bangkok. 

Amari, M. and A. Abe. 1997. Application of 

near infrared reflectance spectroscopy to 

forage analysis and prediction of TDN 

contents. J. Jpn. Grassl. Sci. 31: 55 – 63. 

De Boever, J.L., B.G. Cottyn, J.M. Vanacker 

and Ch.V. Boucque’. 1994. The use of nirs 

to predict the chemical composition and 

the energy value of compound feeds for 

cattle. Anim. Feed Sci. Technol. 51: 243- 

253. 

Osborne, B.G., T. Fearn and P.H. Hindle. 1993. 

Practice NIR Spectroscopy with 

Applications in Food and Beverage 

Analysis. Longman Scientific & Technology. 

Essex, 227 p. 

Sirinapa, S., J. Sornsrivichai and S. Kawano. 

2001. Improvement of PLS Calibration for

180 

brix value and dry matter of mango using 

information from MLR calibration. J. Near 

Infrared Spectrosc. 9: 287-295. 

Warunee, T., A. Terdwongworakul and V. 

Haruthaithanasan. 2001. Present conditions 

of R & D for nondestructive techniques 

including NIR spectroscopy in Thailand. 


pp. 36. In Proc. of the Symposium on 

Nondestructive Techniques the 17 th . Held 

at Tsukuba University, Japan. 

Williams, P. and K. Norris. 1990. Near – 

infrared Technology in the Agricultural 

and Food Industries. American Association 

of Cereal Chemists, Inc. Minnesota, 330 p.


The Effect of Peptidoglycan on Immune Response 

in Black Tiger Shrimp (Penaeus monodon Fabricius) 

Watchariya Purivirojkul 1 *, Nontawith Areechon 2 and Prapansak Srisapoome 2 

ABSTRACT 

The effect of peptidoglycan on immune response in Black Tiger Shrimp (Penaeus monodon 

Fabricius) was conducted under laboratory conditions. P. monodon (13±2 g) were fed for 7 days with 

diets containing graded levels of PG (0, 0.06, 0.12 and 0.18 g/kg feed). The immune responses were 

measured by phenoloxidase, superoxide anion, bactericidal activity, clearance ability and resistance 

against Vibrio harveyi. The results showed that shrimp which fed on diets containing PG 0.18 g/kg had 

these immune parameters at significantly higher (P

182 

approach to increase immunity to defend shrimp 

against diseases. Peptidoglycan (PG), derived from 

the cell wall of gram positive bacteria, was 

reported to be used as an immunostimulant in 

aquaculture. It was found to increase survival rate 

and enhance protection against pathogens in 

shrimp (Soderhall and Cerenius, 1992; 

Boonyaratpalin and Boonyaratpalin, 1995; Sung 

et al., 1996; Itami et al., 1998; Vargas-Albores et 

al., 1998) 

Although peptidoglycan can enhance 

immunity of shrimp, shrimp farmers need to use 

it wisely. If shrimp farmers use excess amount of 

PG supplement in feed, it may waste a lot of 

money. This research was conducted to determine 

the optimum concentration and time for effective 

use of PG which could be recommended for 

shrimp aquaculture. 


Experimental animals 

The test shrimps were obtained from a 

farm in Chachoengsao Province, Thailand. 

Shrimps with approximate weight of 11-15 g per 

animal were used. They were acclimatized in a 

45 × 60 × 45 cm 3 aerated aquarium system filled 

with 25 ppt chlorinated sea water and changing of 

water was made everyday. Twenty shrimps per 

replication were used. 

Experiment 1 Determination of optimum 

concentration 

Immune parameters 

Four treatments consisted of commercial 

shrimp pelleted feed mixed with PG at 0, 0.06, 

0.12 and 0.18 g/kg feed, top-dressing with squid 

oil, were used in the experiment. Shrimps were 

fed five times daily at 3% body weight per day. 

The immune parameters, consisted of 

phenoloxidase, superoxide anion, bactericidal 

activity and clearance ability were measured after 

7 days of the feeding trial. 


Experiment 2 Determination of optimum time 

of application 

The results from experiment 1 were 

evaluated to choose only one suitable 

concentration for this experiment. Four treatments 

consisted of control, feed with PG applied 3, 5 

and 7 days/week were used in the experiment. 

Shrimp were fed five times daily at 3% body 

weight per day. Shrimps were continuously fed 

with PG for 4 weeks, following by normal feed 

for 4 weeks and PG for 2 weeks, respectively. 

Immune parameters consisted of phenoloxidase, 

superoxide anion and bactericidal activity were 

measured every week for 10 weeks. 

All results were evaluated to establish the 

standard protocol for the application of 

peptidoglycan for Penaeus monodon culture 

including the concentration and the time of 

application. 


Preparation of hemolymph samples 

Blood sample of 0.5 ml from each sample 

was withdrawn from the base of 3 th walking leg 

of the shrimp by a syrynge containing 1.5 ml 

anticoagulant (K-199 + 5% L-cysteine). 

1) Phenoloxidase activity assay 

The method was modified from 

Supamattaya et al. (2000). 

After the blood was withdrawn, the 

hemocytes were washed three times with shrimp 

saline (1,000 rpm 4 °C 10 min). Hemocyte lysate 

(HLS) was prepared from hemocytes in a 

cacodylate buffer pH 7.4 by using the sonicator at 

30 amplitute for 5 second and the suspension was 

then centrifuged at 10,000 rpm., 4°C for 20 min. 

The supernatant was collected as HLS. Then 200 

µl of 0.1% trypsin in cacodylate buffer was mixed 

to the 200 ml HLS followed by 200 µl of Ldihydroxyphenylalanine 

(L-DOPA) at 4 mg/ml as 

the substrate. Enzyme activity was measured as 

the absorbance of dopachrome at 490 nm.

wavelength. Measurement of protein content in 

HLS was made by using the method of Lowry et 

al. (1951). The phenoloxidase activity was 

calculated as the increasing of optimum density 

(OD) per minute per mg of protein as: 

1 unit of phenoloxidase = ∆ OD 490 / 

min/ mg protein 

2) Superoxide anion (O 2 - ) 

The method was modified from 

Supamattaya et al. (2000) 

The O 2 - was detected by reduction of 

redox dye, nitroblue tetrazolium (NBT). By this 

method, hemocytes were washed 3 times in K- 

199 solution. Living cells were separated by using 

trypan blue solution and adjusted to 1× 10 7 cell/ 

ml suspension. 200 µl of the cell suspension of 

each sample was dropped into well of a 96 

microwell sterile plate. The plate was left for 45 

min at room temperature for incubation period. 

Unattached cells were washed out by K-199 

solution. 100 µl of the reaction mixture (0.5 mg 

zymozan in 0.5 ml serum + 20 mg NBT in 1 ml 

DMSO + K-199) was added to each well and the 

reaction mixture was incubated at 25 °C for 60 

min. NBT was reduced by O 2 - during incubation 

period into a water insoluable blue formozan 

(Supamattaya et al., 2000). The reaction was 

inhibited by putting 70% methanol into the 

samples for 3 min. and the samples were then 

allowed to air dry. 120 µl of 2M NaOH and 140 

µl of dimethyl sulfoxide (DMSO) were added to 

each well in order to dissolve the formozan. The 

concentration of the prussian-blue-colored solution 

was measured at 620 nm. KOH/DMSO was used 

as a blank control. The amount of O 2 - was 

indicated by the increasing in absorbance at 620 

nm of 0.001 from control. 

3) Clearance ability 

The method of study was modified from 

Martin et al. (1993). The pathogenic bacteria, V. 

harveyi, was subcultured in Tryptic soy agar (TSA) 


with 1.5% NaCl and incubated at 35°C for 24 hrs. 

A single colony of V. harveyi was suspended in 

1.5% NaCl sterile water. The solution with the OD 

of 0.1-0.15 measured by absorbance value at 640 

nm wavelength was used to count for the number 

of bacteria after cultured on TCBS agar by spread 

plate technique. Then 0.1 ml of the bacterial 

suspension with the counted number was injected 

to each tested shrimp while the control was 

injected with saline water. Three hours after 

injections, 0.5 ml of the blood from each shrimp 

was withdrawn to determine the number of 

bacteria by spread plate technique and statistically 

compared. 

4) Bactericidal activity 

Serum was separated from blood of each 

shrimp sample and diluted by 2.6% NaCl at 1:2, 

1:4, 1:8, 1:16 and 1:32. Then 0.5 ml of each serum 

dilution and 0.5 ml of NaCl as the control were 

used in the study. V. harveyi suspension of 0.5 ml 

(prepared from the method as in 3) was put into 

each serum dilution and the control. The treatments 

were incubated at room temperature for 3 hr before 

enumerating the number of bacteria by a spread 

plate technique. Recording of the results were 

made from the dilution that could decrease 50% 

V. harveyi compared to the control. 


Means were statistically compared by 

Analysis of Variance and Duncan’s New Multiple 

Range Test at p = 0.05. 

RESULTS 

Experiment 1 Determination of optimum 

concentration 


1) Phenoloxidase 

After 7 days of feeding, shrimp fed with 

PG 0.18 g/kg had phenoloxidase 358.26 ± 42.42 

unit/min/mg. protein which were significantly

184 

higher (P

The result from this experiment 

suggested that application of PG at 0.18 g/kg 

effectively enhanced the immunity of P. monodon. 

This concentration would be used in Experiment 

2 to determine the optimum time of application. 

Experiment 2 Determination of optimum time 

of application 

Shrimps were fed with PG at 0.18 g/kg 

for 3, 5 and 7 days/week for one month, normal 

unit/min/mg. protein 

500 

400 

300 

200 

100 

0 


feed for one month and PG for other two weeks. 

Study on immune parameters revealed the same 

trend in which during PG feeding period for 4 

weeks 5 and 7 days/week induced significant 

increase of immunity more than 3 days/week and 

control (P

186 

levels were significantly elevated again after two 

weeks of PG supplemented with the same protocol 

(5 and 7 days/week). 

DISCUSSION 

The results from this study indicated that 

PG 0.18 g/kg feed could effectively enhance 

immunity of Penaeus monodon consisting of 

phenoloxidase activity, superoxide anion, 

bactericidal activity and clearance ability. These 

values were significantly different (P

the crop. The recommended protocol will slightly 

increase the cost of production. 

CONCLUSION 

Penaeus monodon fed with 

peptidoglycan supplemented in feed at 0.18 g/kg 

feed for 7 days developed enhanced immunity 

including phenoloxidase activity, superoxide 

anion, bactericidal activity and clearance ability. 

Optimum time for peptidoglycan at 0.18 

g/kg feed was 5 days/week for 1 month, which 

could elevate phenoloxidase activity, superoxide 

anion and bactericidal activity. These elevated 

immune level lasted for other 2-4 weeks after 

shrimps were fed with normal feed and the 

immunity can be enhanced again after the same 

scheme of peptidoglycan application. 


This research was supported by National 

Research Council of Thailand. We thank Aqua 

Progress Co. Ltd. for peptidoglycan. 


Baticados, M.C.L., C.R. Lavilla-Pitogo, E.R. 

Crus-Lacierds, L.D. de la Pena and N.A. 

Sunaz. 1990. Studies on the chemical control 

of luminous bacteria V. harveyi and 

V. splendidus isolated from diseased Penaeus 

monodon larvae and rearing water. Dis. 

Aquat. Org. 91:133-139. 

Boonyaratpalin, S. and M. Boonyaratpalin. 1995. 

Effects of peptidoglycan (PG) on growth, 

survival, immune response, and tolerance to 

stress in black tiger shrimp, Penaeus 

monodon, pp. 469-477. In M. Shariff, J.R. 

Arthur and R.P. Subasinghe (eds.). Second 

Symposium Diseases in Asian Aquaculture. 

Phuket. 

Itami, T. 2002. Promising strategies against 

WSSV for kuruma shrimp in Japan. 


SEAFDEC Asian Aquaculture 24: 9-10. 

Itami, T., M. Asano, K. Tokushige, K. Kubono, 

A. Nakagawa, N. Takeno, H. Nishimura, 

M. Maeda, M. Kondo and Y. Takahashi. 1998. 

Enhancement of disease resistance of kuruma 

shrimp, Penaeus japonicus, after oral 

administration of peptidoglycan derived from 

Bifidobacterium thermophilum. Aquaculture 

164: 277–288. 

Lowry, O.H., N. J. Rosebrough, A. L. Farr and 

R. J. Randall. 1951. Protein measurement 

with the folin phenol reagent. J. Biol. Chem. 

193: 265-275. 

Martin, G. G., D. Poole, C. Poole, J. E. Hose, M. 

Arias, L. Reynolds, N. Mckrell and A. Whang. 

1993. Clearance of bacteria injected into the 

hemolymph of the penaeid shrimp, Sicyonia 

ingentis. J. Inver. Pathol. 62: 308-315. 

Moriarty, D.J.W. 1998. Control of luminous 

Vibrio species in penaeid aquaculture ponds. 

Aquaculture 164: 351-358. 

Soderhall, K. and L. Cerenius. 1992. Crustacean 

immunity. Annu. Rev. Fish Dis. 2: 3-23. 

Supamattaya, K., J. Pongmaneerat and 

T. Klowklieng. 2000. The effect of β–glucan 

(MacroGard ® ) on growth performance, 

immune response and disease resistance in 

black tiger shrimp, Penaeus monodon 

Fabricius. Songklanakarin J. Sci. Technol. 

22 : 677-688. 

Sung, H. H., Y. L. Yung and Y. L. Song. 1996. 

Enhancement of microbicidal activity in the 

black tiger prawn Penaeus monodon via 

immunostimulation. J. Crust. Biol. 16: 278- 

284. 

Vargas-Albores, F., J. Hernandez-Lopes, 

T. Gokkas-Galvan, K. Montano-Perez, 

F. Jimenes-Vega and G. Yepiz-Plascencia. 

1998. Activation of shrimp cellular defence 

functions by microbial products, pp. 161-166. 

In T.W. Flegel (ed.). Proceeding to the 

Special Session on Shrimp Biotechnology 

5 th Asian Fisheries Forum. Advances in 

Shrimp Biotechnology. Chiangmai.


Distribution and Early-life Development of Thai River Sprat 

Clupeichthys aesarnensis Wongratana, Larvae, in Pasak Jolasid 

Reservoir, Lop Buri Province, Thailand 

ABSTRACT 

Santi Poungcharean 

The distribution of Thai river sprat Clupeichthys aesarnensis Wongratana, larvae, in Pasak 

Jolasid Reservoir, Lop Buri Province during April 2003 to February 2004 was studied bi-monthly by 

plankton net towing in 24 sampling stations. The environmental conditions and water qualities were 

recorded and measured in situ. The abundance of larvae showed differential significance in months 

(P

spawning season. The fertilized eggs are 

suspended in water mass. The newly hatching 

larval have approximately 6.6 mm. in total 

length (TL), before developed from pre-flexion 

to flexion, post-flexion and juvenile stage, 

respectively. Thai river sprat is nocturnal fish, 

feeding on zooplankton (Whitehead, 1985 and 

Termvidchakorn, 2003). 

Distribution of Thai river sprat occurrs 

in the river and reservoir of Mekong basin in Laos, 

Cambodia and Thailand. In Thailand, Thai river 

sprats are distributed and abundant in the northeastern 

part. There were distribution records in 

Ubonratana (Khon Kaen province), Sirinthorn 

(Ubon Ratchathani province), Lompow (Kalasin 

province) and Pasak Jolasid reservoir (Lop Buri 

province) (Chookajorn et al., 1977 and Jutagate 

and De Silva, 2003). 

The aims of this paper were to study 

distribution and developing stages of Thai river 

sprat larvae in Pasak Jolasid reservoir. The fieldcollection 

materials and provided information 

were used on size and occurrence of larvae 

collected in sampling stations covering the whole 

reservoir by griding, that was able to understand 

early-life histories and some environmental 

conditions that required spawning and nursery. 

Finally, the outputs of this paper were be able to 

support Thai river sprat fishery management in 

Pasak Jolasid and others reservoir based on 

spawning-nursery ground and timing information. 


1. Study sites 

Pasak Jolasid reservoir is located in 

Pasak river at geological coordinated 14°50′ 32″ 

N 101°05′ 00″E. The dam has been operational 

since September 1999 with 960 cubic million 

meter at maximum level, 43 meters above the 

mean sea level, covering approximately 149 square 

kilometers. The dam receives most water from 

Pasak river, originating from Phetchabun mountain 


in the South of Loei province. 

The 24 sampling stations were studied 

to ensure the coverage of the whole reservoir by 

griding. The sampling stations were located in 4 

separated zones by different morphology of the 

reservoir as follows (Figure 1): 

Zone 1: Upstream zone (14°50′ N to 

15°11′ E) consisted of 4 stations along Pasak river. 

The both sides, especially the west were changed 

to floodplain flood season. 

Zone 2: River inlet zone (15°00′ N to 

15°05′ E) consisted of 8 stations in the upper part. 

It was shallow and widely area that received water 

from upstream with many of marginal grasses, 

litter and remnant woods visible especially when 

water level decreased. 

Zone 3: Transitional zone (14°55′ N to 

15 o 00′ E) consisted of 8 stations in the middle part. 

It was an open area with high wind and strong 

Figure 1 Sampling stations for Thai river sprat 

larvae in Pasak Jolasid reservoir, Lop 

buri province.

190 

currents. 

Zone 4: River outlet zone (14°50′ N to 

14°55′ E) consisted of 4 stations in the lower part. 

It was an open and deeply area with high wind 

and strong currents. 

2. Field surveys 

The quantitative collection of fish larvae 

in the study area was carried out by the larval net 

that was cylindrical shape, 120 centimeter in length 

with 650 microns in mesh size at the mouth part 

and 330 microns at the 60 centimeters of the cod 

end with opening of 60 centimeters in diameter. 

The flow meter was attached at the mouth part to 

enable the estimation of water volume movement 

through the net. Larval net was operated by towing 

about 5 minutes at 0.5 meters under subsurface 

level with boat speed approximately 1 knot. The 

environmental conditions and water qualities were 

recorded and measured in situ. The sampling 

stations were operated every bi-monthly. All 

samples were kept in 10% water formalin and after 

sorting, the larvae were preserved in 4% 

neutralized formalin solution for analysis. 

3. Identification and data analysis 

Thai river sprat larval specimens were 

identified under stereo-microscope by 

Termvidchakorn (2003)’s guidebook. The 

estimation of fish larvae density was evaluated and 

standardized in 1,000 cubic meters of water 


volume. The characteristics i.e. body shape, 

myomere, strict and position of pigmentation in 

every stage of fish larvae were described and 

illustrated by stereo microscope and camera lucida. 

The difference of larvae density in each month was 

tested by ANOVA analysis and Duncan Multiple 

Range Test (DMRT). 

RESULTS 

1. Distribution of Thai river sprat larvae 

Thai river sprats were collected from 6 

survey periods during April 2003 to February 

2004. Only 19 from 24 sampling stations in June 

2003 and August 2003 were collected (Station 8- 

12) because the water level was decreased by 

irrigation discharge (Table 1). 

Thai river sprat larvae were widely and 

irregularly distributed in Pasak Jolasid reservoir 

and Pasak river. The larval abundance were highest 

in June 2003 (46.87% of all surveys) and 

significantly different from the other months 

(P

water temperature was high in April 2003 

(31.5±0.8°C) and low in December 2003 

(23.0±0.3°C). The secondary data, averaged water 

level of reservoir was conducted from the Royal 

Irrigation Department’s reports and then, water 

volume and water surface area were estimated 

from water level at the dam site (Table 2). 

3. Development of Thai river sprat larvae 

There were 4,787 individuals from the 

total of 6 sampling periods and all of these 

consisted of series of different larval developing 

stages. The complete specimens in each series 

were selected for taxonomical illustrations and 

descriptions as follows: 

Yolk-sac stage (Figure 3a) 3.46 mm. in 

notochord length (mm.NL), body elongate with 

37 myomere, head round, upper jaw reaching the 

middle of eye, strength gut opening at 82% of total 

length, yolk sac at anterior part of gut, fins not 

developed, the end of notochord (urostyle) 

strength. 

Pre-flexion stage (Figure 3b) 4.11 mm. 

NL, body elongate with 39 myomere, head round, 

upper jaw beyond to anterior of eye, strength gut 

opening at 80% of total length, no yolk sac, fins 

not developed, the end of notochord descending 

upward to dorsal position. 

Flexion stage (Figure 3c) 7.35 mm.NL, 

body elongate with 39 myomere, head round, 


upper jaw beyond anterior of eye, strength gut 

opening at 78% of total length, Pectoral, dorsal, 

anal, caudal fin developed, posterior gut vertical 

striated. Posterior gut and brain covered with 

melanophore pigments. 

Post-flexion stage (Figure 3d) 11.61 mm. 

in total length (mm.TL), body elongate, strength 

gut opening at 78% of total length, fins developed 

except pelvic fin. Few melanophore pigments 

covered on brain, caudal fin base and spotted along 

anal-fin base. 

Juvenile stage (Figure 3e) 11.61 mm.TL, 

body deeply, gut opening at 63% of total length, 

fins developed, scutes at belly developed. (Figure 

3f) 20.83 mm.TL, body widely, gut opening at 59% 

of total length, similar to adult. 

DISSCUSSIONS 

There were very few studies on 

freshwater fish larvae in Thai reservoir for 

prediction spawning ground and season. Although 

there were several studies about spawning ground 

but methodology and some objectives were not 

different. However, the variety of Thai river sprat’s 

distribution data from this study can be used to 

indicate spawning ground and season in Pasak 

Jolasid reservoir. Because the fish in larval stage 

has a limited ability to movement in water volume 

especially in Pasak Jolasid reservoir which is 

Table 2 Environmental condition data which parallel measured and recorded with surveys in Pasak 

Jolasid reservoir during April 2003 to February 2004. 

Observations Dissolved oxygen Water Water level Water Water 

(ml/l; Mean±SD) temperature (m) surface area* volume 

(C; Mean±SD) (km 2 ) (million m 3 ) 

April 2003 6.85 ± 1.98 31.5 ± 0.8 39.50 121.08 474.21 

June 2003 5.51 ± 2.97 29.6 ± 2.1 37.58 93.37 271.94 

August 2003 4.76 ± 0.87 29.3 ± 0.2 36.88 72.89 235.77 

October 2003 8.27 ± 1.60 29.9 ± 1.2 41.64 143.98 763.43 

December 2003 9.78 ± 1.45 23.0 ± 0.3 42.36 156.63 755.53 

February 2004 8.77 ± 2.01 24.4 ± 0.1 40.45 144.57 616.00 

* estimated from Royal Irrigation Department (1999)’s operation rule curve

192 

standing waters, thus the occurrences of larvae in 

sampling stations indicated specific spawning in 

each area determined. 

The results showed that, spawning of 

Thai river sprat occurred throughout of the year 

according to Sirimonkonthawon (1994) and 

Chookajorn et al.(1997). The populations of Thai 

river sprat in northeast Thailand is a short-life span, 

spawning period is peaked in June to July or rainy 

season and low in December to February or cold 

season. Occurrence of larval population surveys 

in June and August which covered this period was 


regarded according to their spawning season. 

Abundance of larvae was high in June- 

October or rainy season because of the optimized 

of their environmental conditions such as water 

temperature of 29.3-31.5 °C, the optimum level 

for general tropical fishes was supported by 

Termvidchakorn (2003) who explained that low 

temperature in cold season was the factor that 

inhibited the gonad development in fish especially 

in tropical zone. In addition, notification of fixing 

spawning season and fishing gears from 16 May 

to 15 September in every year from Department 

Figure 2 Distribution of Thai river sprat in Pasak Jolasid reservoir during April 2003 to February 2004 

(size of dark cycling replicated density of Thai river sprat larvae per 1,000 cubic meter of 

water volume).

of Fisheries protects and provides parent stocks 

survival to spawn. On the contrary, Jutagate (2002) 

discussed that catches and efforts of Thai river 

sprat fisheries in Sirinthorn reservoir were low 

because of wind effect, which created turbulent 

water in northeast monsoon as well as Pasak 

Jolasid reservoir which received impact form 

northeast monsoon. Abundance of larvae was low 

in December to February. As well as the above 

reason, the water temperature in this period was 

as low as 23.0-24.1 °C and this period was an 

opening of fishing season, there were Thai river 

sprat’s fishery in Pasak Jolasid reservoir. In 

Figure 3 Larval development of Thai river sprat 

(a) Yolk-sac stage; 3.46 mm.NL (b) Preflexion 

stage; 4.11 mm.NL (c) Flexion 

stage; 7.35 mm.NL (d) Post-flexion 

stage; 11.61 mm.TL (e) and (f) Juvenile 

stage; 12.03 and 20.83 mm.TL. 


addition, rising of water level to the maximized 

level in October and slow deceasing till April 

(Royal Irrigation Department, 1999) provided Thai 

river sprat larvae’s favorableness as their habitat. 

It reduced chances of larval catchments by fish 

larvae net or other samplings because yield of 

larvae per water volume deceased. 

This study showed that Thai river sprat 

larvae were more abundant in offshore or river 

canal than near shore (Figure 2). Whitehead (1985) 

described that the mostly clupeids fish was a 

pelagic schooling fish and spawning in open area, 

the fertilized eggs and early-life stages were 

suspended in water volume as plankton. Lima and 

Lima (2004) explained that fish in early life stage 

had a passive movement, wind and current was a 

rule for their drift and distribution. The offshore 

of Pasak Jolasid reservoir was an area that strongly 

received impacts from wind and current as the 

same. 

Populations of Thai river sprat in Pasak 

Jolasid reservoir were in higher abundance than 

before the dam was constructed. This was because 

the environmental changed from running to 

standing water condition. This changing was 

conducted to provide Thai river sprat’s habitat. 

However, Thai river sprat were found in Pasak 

river before the dam was built by Electric 

Generating Authority of Thailand (1982) and 

Department of Fishery (1994), but there were only 

few numbers of populations. After the dam was 

constructed, populations of Thai river sprat were 

high and had production to the fishery level. From 

this survey, it was found that there were Thai river 

sprat fisheries in the East-Middle part of reservoir 

by breach seine. Additionally, Department of 

Fisheries (2002) and SUMAFISH (2003) reported 

that Thai river sprat were in high abundance after 

the dam was constructed. 

Development of Thai river sprat’s 

myomere in yolk-sac larvae were not complete. 

Figure 3a shows 37 developed myomere, with 

opened digestive gut at 82%NL in yolk-sac larvae

194 

and moving to 59%TL in juvenile (Figure 3b-f). 

Malanophore pigmentation occurred on brain, 

anal-fin base and caudal-fin base after flexion 

stage. The results were similar to those reported 

by Termvichakorn (2003). The other pigments or 

coloration might be lost from preservation by 

neutralized formalin solution especially silver 

write transverse-band, located from posterior gill 

opening to caudal-fin base not visible from 

preserved specimens. However, yellowish 

coloration on caudal-fin base was visible from 

some preserved specimens. In identification, Thai 

river sprat larvae may be misidentified among 

Boneo river sprat (Clupeiodes boneensis Bleeker, 

1851) especially before post-flexion stage 

specimens. It was a small clupeids fish found in 

Pasak Jolasid reservoir but not in abundance. 

However, Thai river sprat larvae could be 

separated by the last two anal-fin rays detached 

form the others (Whitehead, 1985 and Rainboth, 

1996). 

CONCLUSIONS 

Spawning of the Thai river sprat in Pasak 

Jolasid reservoir occurred throughout of the year 

but was high in early rainy season and low in cold 

season and irregularly in anywhere of the reservoir. 

Low water temperature and current affected 

spawning and larval drift to suitable habitat. 

Development of myomere in the yolk-sac larvae 

was not complete while the digestive gut moved 

form posterior into mid-portion of body in 

juvenile. 


I am very grateful to Dr. Prachit Wongrat, 

Dr.Apichart Termvidchakorn and Dr. Charumas 

Meksumpun, my thesis committee for their 

valuable advices. Thanks are extented to 

Mr.Suchart Piladech, Fisheries Resources 

Conservative Unit, Department of Fisheries leader 


and his officers, for their helpfulness and facilities 

in fieldwork. Finally, I also thanks the members 

in my laboratory for their helpfulness and 

friendships. 


Chookajorn, T., N. Sompasong, and S. Chotikul. 

1997. Thai river sprat fisheries in Northeast 

Thailand. Thai.Fish.Gaz. 30: 93-99. 

Department of Fisheries. 1994. Investigation of 

fisheries biology and environment in Pasak 

river before dam construction. Special 

technical paper. Freshwater Fisheries Division 

and Fisheries Environmental Division. 

Ministry of Agriculture and Cooperatives. 76 

p. 

Department of Fisheries. 2001. Ecological and 

Fish Population in the Pasak Jolasid 

Reservoir. Technical paper No.24. Fisheries 

Environmental Division. Ministry of 

Agriculture and Cooperatives. 40 p. 

Electricity Generating Authority of Thailand. 

1982. Upper Pasak project; environmental 

and ecological investigation (main report). 

Prepared by TESCO, Ltd. In association with 

Mahidol University and Chulalongkorn 

University. 21 p. 

Jutagate, T. 2002. Thai river sprat: Biology and 

management in Sirinthorn reservoir, 

Thailand. Ph.D thesis. Deakin University, 

Warrambool. 

Jutagate,T. and S. De Silva 2003. Yield, growth 

and mortality rate of the Thai river sprat, 

Clupeicthys aesarnensis, in Sirinthorn 

Reservoir, Thailand. Fisheries Management 

and Ecology 10: 221-231. 

Lima,A.C.D. and C.A.R.M.Araujo-Lima. 2004. 

The distribution of larval and juvenile fishes 

in Amazonian rivers of different nutrient 

status. Freshwater Biology 49: 787-800. 

Rainboth, W. J. 1996. Fishes of the Cambodian 

Mekong. FAO species identification field

guide for fisheries purposes. FAO, Rome. 

265 p. 

Royal Irrigation Department. 1999. The Pasak 

Jolasid Dam, Lop buri and Saraburi 

province. Contribution on the Opening 

Ceremony of the Pasak Jolasid Dam by His 

Majesty the King Bhumipol. 

Sirimongkolthawon, R. 1994. Some biological 

aspect of Clupeichthys aesarnensis 

(Clupeidae) in Ubonratana Reservoir, 

Khon Haen province. Technical paper No.44. 

Freshwater Fisheries Division, Department of 

Fisheries. 38 p. 

SUMAFISH. 2003. Strategies for Sustainable 

Management of Fishery Resources in the 

Pasak Jolasid Reservoir, Thailand through 


Ecological and Socio-economic Assessment 

(Final report). ASEAN Centre for 

Biodiversity Conservation and the European 

Commission. 298 p. 

Termvidchakorn, A. 2003. Freshwater fish 

larvae. Inland Fisheries Resources Research 

and Development Institute. Department of 

Fisheries. Bangkok. 130 p. 

Whitehead, P.J.P. 1985. Clupeoid Fishes of the 

World (Suborder Clupeoidei): Part 1. FAO, 

Rome. 303 p. 

Wongratana,T. 1983. Diagnoses of 24 new species 

and proposal of a new name for a species of 

Indo-Pacific clupeoid fishes. Jap.J.Ichthyol. 

29: 385-407.


Gonadal Development and Sex Inversion in Saddleback 

Anemonefish Amphiprion polymnus Linnaeus (1758) 

Sukjai Rattanayuvakorn 1 , Pisut Mungkornkarn 2 , Amara Thongpan 3 

ABSTRACT 

and Kannika Chatchavalvanich 4* 

Gonadal development and sex inversion of saddleback anemonefish, Amphiprion polymnus 

were described. One-month juveniles had sexually undifferentiated gonads with primordial germ cells 

aggregated in groups, while two- to three-month juveniles displayed immature hermaphroditic gonads 

containing early developmental stages of both male and female germ cells, namely spermatogonia, 

primary spermatocytes, oogonia, and primary oocytes in chromatin-nucleolus stage. Spermatogenesis 

began at 4 months having testicular tissue comprising of spermatogenic cells in all developmental stages 

but ovarian cavity was first seen later at 5 months. The male region of ovotestis was peripheral, whereas 

the female region was more centrally located. There was no connective tissue between ovarian and 

testicular areas. Six-to 11-month fish had slightly larger ovotestes than those at previous age. Protandric 

sex inversion first occurred at 12 months. Sex change was characterized by degeneration of male germ 

cells, deposition of yellow-brown pigment and the formation of vitellogenic oocytes. Before spawning 

activity began, their gonads contained female germ cells in all stages with numerous vitellogenic oocytes, 

whereas functional males had both ovarian and testicular tissues. Most females of breeding pairs had 

mature oocytes in their gonads and began to spawn when their ages reached 14 months. 

Key words: gonadal development, sex inversion, saddleback anemonefish 

INTRODUCTION 

Growth and development of juvenile fish 

to reach their mature stage are generally 

determined by the gonadal development. Male and 

female are distinctly different in their time required 

for sex differentiation and morphological changes. 

Most marine fishes are hermaphrodite, i.e., 

ambiguous in sexual organ identification 

especially during the juvenile stage (Warner, 1984; 

Garratt, 1986). Anemonefishes are known for 

socially controlled protandry with a monogamous 

mating system and live in colonies within clusters 

of sea anemones (Fricke and Fricke, 1979; Moyer 

and Nakazono, 1978; Ross, 1978a, b; Fricke, 1979, 

1983). The gonad of juveniles, subadults and 

functional males are ovotestes but those of females 

are ovaries (Moyer and Nakazono, 1978; Fricke, 

1979). Only the largest two individuals in each 

colony are reproductively active. The largest fish 

1 Institute of Marine Science, Burapha University, Bangsaen, Chonburi 20131, Thailand. 

2 Institute of Science, Rangsit University, Patumthani 12121, Thailand. 


4 Department of Zoology, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fsciknc@ku.ac.th 


is a functional female and the second largest one 

is a functional male (Fricke and Fricke, 1977; 

Fricke, 1979). These functional female and male 

live together as a pair in which the female 

dominates the male. If the functional female 

disappears or is removed from a colony, the 

functional male changes sex protandrically and 

becomes a reproductive female (Shapiro, 1992). 

Hattori (1991) reported on this type of socially 

controlled growth and sex change in anemonefish 

Amphiprion frenatus in Okinawa, Japan. 

Saddleback anemonefish Amphiprion 

polymnus Linnaeus (1758), is predominantly 

found in the Gulf of Thailand. It is one of the main 

species living on the coral reefs in this area. 

Unfortunately, their number is drastically 

decreased due to deterioration of natural habitat 

and the high commercial demand of them as 

ornamental fish. To replenish its population, 

attempts are made to raise the anemonefish in 

artificial conditions and mass production. The first 

approach to attain this goal was to understand its 

reproductive system, the stage of gonadal 

development and timing of their sex inversion. So 

far, there has been no record on Amphiprion 

polymnus germ cell morphology, gonadal 

development, sex inversion especially at 

histological level. This would render a better plan 

in controlling their population or inducing sex 

inversion at certain period and condition as well. 


Ten couples of sexually mature male and 

female saddleback anemonefishes, Amphiprion 

polymnus Linnaeus (1758) with the average 86 cm 

SL (standard length), 16.18 mg (SL range 55-113 

cm, body weight range 8.5-28.2 mg) were 

collected from the Gulf of Thailand. They were 

raised in the laboratory at Burapha University, 

Chonburi Province. The rearing conditions were 

set to mimic the saddleback natural habit. They 

were kept in a tank of 30 × 60 × 40 cm in size. The 


aerated sea water in the tank was controlled at 25- 

28 °C, having 30-32 ppt salinity, 0.02 ppm 

ammonia, 0.01 ppm nitrite, 10 ppm nitrate. The 

pH varied between 6.5 and 7.5. Egg-laying 

materials comprised of bivalve shell, bark rock, 

coral and seaweeds. Broken ceramic plates were 

also put in the tank as lining for home surrounding, 

while sea anemones Heteractis crispa and 

Stichodactyla haddon were added to mimic the 

natural environment of this fish. They were fed 

with brine shrimp, finely chopped fish, shrimp, 

clams and dry algae flakes ad libitum. F1 offspring 

were reared in a tank of 75 × 150 × 80 cm. When 

their ages reached 5 months, groups of three fish 

each were collected and reared in separate 

container of 30 × 60 × 40 cm until F1 offspring at 

the age of 12-14 months laid eggs. 

Sample preparation 

Five to ten juveniles starting from the age 

of 1-month were monthly sampled until they 

reached 14-month of age. They were anaesthetized 

with quinadine. Their gonads were excised and 

fixed. The gonads of specimen larger than 40 mm 

SL were removed and fixed in Bouin’s solution 

for 24 h. The whole body without head and tail of 

smaller than 40 mm SL were fixed in Bouin’s fluid, 

and decalcified for 3 days. They were processed 

using a standard paraffin method and serially 

sectioned at 6 ?m thick. Sections were stained with 

hematoxylin - eosin and examined under a light 

microscope. 

RESULTS 

Gonadal development of Amphiprion polymnus 

Since saddleback anemonefish 

Amphiprion polymnus was a protandrous 

hermaphrodite, its gonad was found to contain both 

male and female germ cells at the early stages of 

development. Primordial germ cells were found 

when they were 1 month old. These cells were 

aggregated in a group of 8-15 cells enclosed with

198 

thin membrane (Figure 1A, B). The primordial 

germ cells were spherical, about 6.1 µm in 

diameter, with large nucleus and scanty cytoplasm 

(Figure 1 B). Each aggregation was enclosed by 

thin connective tissue and attached to the 

abdominal cavity by the dorsal mesentery. Gonads 

at this stage of unidentified male or female were 

defined as indifferent gonads. 

When the juveniles reached 2-3 months 


old, gonadal lamellae were found parallel along 

the abdominal cavity beside the small intestine 

(Figure 1C). The gonads contained both male and 

female germ cells, namely spermatogonia, 

oogonia, some primary spermatocytes and primary 

oocytes in chromatin-nucleolus stage (Figure 1D, 

E, F). These spermatogonia were spherical, about 

6.4 µm in diameter, with round nuclei and clear 

cytoplasm. The spermatogonia were enclosed by 

Figure 1 The gonadal development of Amphiprion polymnus showing (A) the location of primordial 

germ cells, (B) group of primordial germ cells developed in gonads of 1-month juvenile, (C) 

and (D) ribbon-like gonads of 2-month juvenile, (E) and (F) gonads of 3-month juvenile. K, 

kidney; Oc, oocyte; Og, oogonia; PGC, primordial germ cells; Sc, spermatocytes; Sg, 

spermatogonia.

their supporting cells. The primary spermatocytes 

were about 5.6 µm in diameter; their nuclei were 

at early stage of prophase I. These germ cells 

gradually multiplied in number by mitosis and 

formed spermatocyst among stromal tissue. The 

oogonia, however, were about 10.7 µm in diameter 

with round or oval nuclei having distinct nucleoli. 

Primary oocytes were distinguished from oogonia 

by their larger size and more darkly stained 

cytoplasm. The germ cells of both sexes were 

found intermingled in the same area. 


Spermatogenesis began at 4 months 

(Figure 2A, B). Male area of a gonad contained 

spermatogenic cells in all stages, i.e. 

spermatogonia, primary spermatocytes, secondary 

spermatocytes, spermatids and spermatozoa. The 

development of these germ cells in spermatocysts 

was synchronous. As for female area at 4 months, 

the gonad consisted of oogonia, oocytes in 

chromatin-nucleolus stage and perinucleolus stage 

but lack discernable ovarian cavity (Figure 2B). 

Follicle cells which enclosed perinucleolar oocyte 

Figure 2 The gonadal development of Amphiprion polymnus showing (A) and (B) gonads of 4-month 

fish, (C) and (D) gonads of 5-month fish, (E) and (F) gonads of 6-month fish. OC, ovarian 

cavity; OT, ovarian tissue; Sc, spermatocytes, Sz, spermatozoa; TT, testicular tissue.

200 

were clearly seen. 

The presence of an ovarian cavity was 

first seen at 5 months (Figure 2 C). At this same 

period of time, the testicular tissue of ovotestis 

was peripherally located, while the ovarian tissue 

located more centrally (Figure 2D). There was no 

connective tissue between ovarian and testicular 

regions. Germ cell types of both sexes were the 

same as those found at the previous stage (4months). 

Ovotestes of 6-to 11-month fish were 

similar to those of 5-month but gonadal size was 

larger than those at previous stages. Testicular 

tissues and the number of male germ cells had 

remarkably increased. Ovarian tissues were 

composed of the same cell types as those found at 

previous stages but perinucleolar oocytes were 

increased in number in older fish. Vitellogenic 

oocytes were still not found. 

Sex inversion 

Male saddleback anemonefish began to 

have sex change when their ages were between 

12-to14-months. The sex change was characterized 

by the diminishing of male germ cells whereas 

the female germ cells were increased in number. 

The degeneration of male germ cells was also 

observed (Figure 3A, B, and C). Numerous 

pyknotic nuclei were observed in deteriorate 

spermatocyst (Figure 3 D). Late stage of 

degeneration resulted in the appearance of yellowbrown 

pigment which usually located at the 

periphery of gonads (Figure 3D). At the same time, 

vitellogenic oocytes that contained numerous yolk 

granules were developed in their gonads (Figure 

3E, F). However, advanced stage of vitellogenic 

oocytes was not found. At this age, some fish still 

had gonad containing numerous male germ cells 

in all stages and also female germ cells but oogenic 

activity did not proceed beyond perinucleolus 

stage. However, few yellow-brown pigments were 

found. When the fish reached 14 months, most 

females of breeding pairs began to have spawning 


activity. The third and usually the smallest fish of 

each colony, aged 5-to-14-months remained to be 

juvenile as long as breeding pairs still persisted. 

DISCUSSION 

Gonadal development of Amphiprion 

polymnus was found closely correlated with age. 

This differed from that of European eel which 

related more to body size (Colombo et al., 1984; 

Colombo and Grandi, 1996). The appearance of 

sexually differentiated germ cells, marking the 

steps of gonadal differentiation, corresponded to 

2-3 months of ages. The first sexually 

differentiated germ cells of anemonefish were 

hermaphrodite, having clones of both oogonia and 

spermatogonia, while those of the European eel 

were female only (Grandi and Colombo, 1997). 

Thus, in Amphiprion polymnus, gonadal 

development goes through a hermaphroditic phase 

before proceeding in two different ways. First, the 

development of ovotestis proceeds with 

spermatogenesis of the testicular zone, and the fish 

function as males. Second, some hermaphrodites 

change sex at the age of 12-14 months and function 

as females. 

In A. polymnus, the male zone of 

ovotestis was peripheral, while the female part was 

more centrally located but no connective tissue 

was found between ovarian and testicular regions. 

These were the same as those of A. frenatus and 

other anemonefish (Brusle-Sicard and Reinboth, 

1990), but were contrast to most protandric 

teleosts, in which the ovarian and testicular zones 

are distinctly separated by a well-developed 

connective tissue (Reinboth, 1962, 1970; Pollock, 

1985; Micale and Perdichizzi, 1994). Moreover, a 

typical sparid gonad was found to consist of a 

dorsal ovarian zone and a ventral testicular zone 

(Pollock, 1985; Micale and Perdichizzi, 1994). 

Gonads of 1-month juvenile of A. 

polymnus were indif ferent since they consisted of 

primordial germ cells, having spermatogonia and

oogonia at undifferentiated stages. Gonads of 2- 

3-month fish were immature hermaphrodite 

consisting of early stages of both male and female 

germ cells. This observation revealed that sex 

differentiation began at 2-3 months. Moreover, the 

development of male germ cells in a cyst was 

synchronous as those of Amphiprion frenatus 

(Brusle-Sicard and Reinboth, 1990). 

Gonads of 4-month fish had 


spermatogenesis but had no oogenic activity. This 

observation indicated that these hermaphroditic 

gonads stages were actually functional male. 

Oogenic activity of ovotestis does not proceed 

beyond the perinucleolus stage, as also seen in the 

protandric Amphiprion frenatus (Brusle-Sicard and 

Reinboth, 1990). Protandric sex inversion began 

in 14-month-old fish. However, if the rearing 

condition was suitable, the hermaphroditic gonad 

Figure 3 Protandric sex change of Amphiprion polymnus showing (A) degeneration of testicular tissue, 

(B) and (C) increasing of ovarian tissue, (D) pyknotic nuclei and yellow-brown pigment, (E) 

developing vitellogenic oocytes; (F), vitellogenic oocytes. D, degeneration of testicular tissue; 

Oc, ovarian cavity; OT, ovarian tissue; P, pigment; PN, pyknotic nuclei; TT, testicular tissue; 

VOc, vitellogenic oocytes; YG, yolk granules.

202 

could change to female at 6-12 months depending 

on several factors, i.e., the completion of rearing 

condition or the removal of functional female from 

a colony (unpublished data) which was similar to 

other anemonefishes (Fricke, 1979, 1983; Ochi 

and Yanagisawa, 1987; Ochi, 1989; Hattori and 

Yanagisawa, 1991; Brusle-Sicard et al., 1994). 

Environmental factors are claimed to play a major 

role in sex differentiation. High population 

densities and high temperature are also known to 

favor male differentiation (D’ Ancona, 1957). 

Temperature and steroid hormone may act on 

gonadal differentiation in fishes by inducing or 

inhibiting the production of H-Y antigen early in 

development, as also seen in chickens and sea 

turtles (D’Ancona, 1959). 

The observation of sex change in A. 

polymnus as characterized by degeneration of male 

germ cells, deposition of yellowish pigment and 

the formation of advanced vitellogenic stage 

oocytes was similar to those occurred in Diplodus 

sargus (Micale and Perdichizzi, 1994). These 

yellowish pigments were usually found at the 

periphery of gonads which was the location of the 

male germ cells, in which the male germ cells had 

been replaced by masses of yellowish pigment. 

However, gonads of some fish even at the age over 

12 months may still contain numerous male germ 

cells in all stages. This indicated that these fish 

might be functional male throughout their lives, 

but they also could be changed to female if the 

condition favored the inversion. 

CONCLUSION 

Histological evidents indicated that the 

hermaphroditic gonad of Amphiprion polymnus 

under rearing condition had spermatogenic activity 

for the first time at the age of 4 months and had 

protandric sex inversion at the age of 12-14 

months. Sex inversion in saddleback anemonefish 

was essential to reproduction. Gonadal 

development began with hermaphrodite and 


proceeded to functional male and functional 

female, respectively. Functional female developed 

from protandric sex inversion only. Sex inversion, 

however, depended on external factors and 

conditions which could be induced at proper stage 

of development in rearing condition. Without sex 

inversion, Amphiprion polymnus would remain 

hermaphrodite and functional male all their lives 

which obstructed their full productive cycle and 

lower the population in natural environment as 

well. 



Graduate School, Kasetsart University, and 

Institute of Marine Science, Burapha University. 


Brusle-Sicard, S. and R. Reinboth. 1990. 

Protandric hermaphrodite peculiarities in 

Amphiprion frenatus Brevoort (Teleostei, 

Pomacentridae). J. Fish Biol. 36: 383-390. 

Brusle-Sicard, S.R. Reinboth. and B. Fourcault. 

1994. Germinal potentialities during sexual 

state changes in a protandric hermaphrodite, 

Amphiprion frenatus Brevoort (Teleostei, 

Pomacentridae). J. Fish Biol. 45: 597-611. 

Colombo, G., G. Grandi. and R. Rossi. 1984. 

Gonad differentiation and body growth in 

Anguilla anguilla. J. Fish Biol. 24: 215-228. 

Colombo, G. and G. Grandi. 1996. Histological 

study of the development and sex 

differentiation of the gonad in the European 

eel. J. Fish Biol. 48: 493-512. 

D’Ancona, U. 1957. Nuove ricerche sulla 

determinazion sessuale dell’anguilla. II. Le 

influenze ambientail sul differenziamento 

della gonade. Arch. Oceano. Limnol. 11: 69- 

111. 

D’Ancona, U. 1959. Distribution of the sexes and 

environmental influence in the European eel.

Arch. Anat. Microsc. Morphol. Exp. 48: 61- 

70. 

Fricke, H.W. 1979. Mating system resource 

defense and sex change in anemonefish 

Amphiprion akallopisos. Z. Tierpsychol. 50: 

313-326. 

Fricke, H.W. and S. Fricke. 1977. Monogamy and 

sex change by aggressive dominance in coral 

reef fish. Nature, Lond. 226: 830-832. 

Fricke, H.W. 1983. Social control of sex: field 

experiments with the anemonefish 

Amphiprion bicinctus. Z. Tierpsychol. 61: 

71-77. 

Garratt, P.A. 1986. Protogynous hermaphoditism 

in the slinger, Chrysoblephus Puniceus 

(Gilchrist & Thompson, 1908) (Teleostei: 

Sparidae). J. Fish Biol. 42: 699-712. 

Grandi, G. and G. Columbo. 1997. Development 

and early differentiation of gonad in the 

European eel, Anguilla anguilla (L.) 

Anguilliformes, Teleostei: A cytological and 

ultrastrutural study. J. Fish Morphol. 231: 

195-216. 

Hattori, A. 1991. Socially controlled growth and 

size-dependent sex change in the anemonefish 

Amphiprion frenatus in Okinawa, Japan. Jpn. 

J. Ichthyol. 38: 165-177. 

Hattori, A. and Y. Yanagisawa.1991. Sex change 

of the anemonefish Amphiprion clarkii in a 

habitat of high host density: a removal study. 

Jpn. J. Ecol. 41: 1-8. 

Moyer, J.T. and A. Nakazono. 1978. Protandrous 

hermaphoditism in sex species of anemonefish 

genus Amphiprion in Japan. Jpn. J. Ichthyol. 

20: 85-93. 

Micale, V. and F. Perdichizzi. 1994. Further studies 

on the sexuality of the hermaphroditic teleost 


Diplodus sargus, with particular reference to 

protandrous sex inversion. J. Fish Biol. 45: 

661-670. 

Ochi, H. 1989. Mating behavior and sex change 

of the anemonefish Amphiprion clarkia in 

temperate waters of southern Japan. Env. Biol. 

Fish. 26: 257-275. 

Ochi, H. and Y. Yanagisawa. 1987. Sex change 

and social structure in anemonefish in 

temperature water, pp. 239-241. In Y. Ito., 

J.L. Brown and J. Kikkawa (eds.). Animal 

societies: Theories and facts. Japan 

Scientific Societies Press, Tokyo. 

Pollock, B.R. 1985. The reproductive cycle of 

yellowfin bream, Acanthopagrus Australis 

(Gunther), with particular reference to 

protandrous sex inversion. J. Fish Biol. 26: 

301-311. 

Reinboth, R. 1962. Morphologische und 

fuktionelle Zweigeschlechtlichkeit bei 

marinen Teleostiern ( Serranidae, Sparidae, 

Centracanthidae, Labridae). Zool. Jahrb. 

Abt. Allg. Zool. Physiol. 69: 405-480. 

Reinboth, R. 1970. Intersexuality in fishes. Mem. 

Soc. Endocrinol. 18: 515-543. 

Ross, R.M. 1978a. Reproductive behavior of the 

anemonefish Amphiprion melanopus on 

Guam. Copeia 1: 103-107. 

Ross, R.M. 1978b.Territorial behavior and ecology 

of the anemonefish Amphiprion melanopus on 

Guam. Z. Tierpsychol. 46: 71-83. 

Sharpiro, D.Y. 1992. Plasticity of gonadal 

development and protandry in fish. J. Exp. 

Zool. 261: 194-203. 

Warner, R.R. 1984. Mating behavior and 

hermaphroditism in coral reef fishes. Am. Sci. 

72: 128-136.


In sacco Degradation Characteristics of Crop Residues and Selected 

Roughages in Brahman-Thai Native Crossbred Steers 

ABSTRACT 

Songsak Chumpawadee 1 , Kritapon Sommart 2 , 

Thevin Vongpralub 2 and Virote Pattarajinda 2 

Three crop residues and five roughages were selected to evaluate nutritive value using nylon 

bag technique. Nylon bag technique was conducted in two rumen fistulae Brahman-Thai native crossbred 

steers. Steers were fed 0.5% BW of concentrate and rice straw ad libitum. The treatments were 1) water 

hyacinth, 2) kraphanghom, 3) corn stover, 4) cassava hay, 5) sugarcane top, 6) Chinese spinach, 7) rice 

straw and 8) cavalcade hay. The result indicate that the rapidly soluble fraction (a) of DM OM and CP 

were highest in Chinese spinach (P

(Chinh et al., 2000), rice straw (Agbagla-Dohnani 

et al., 2001) and corn stover (Hindrichson et al., 

2001). Crop residues have the advantage of being 

nutritionally valuable even in the dry season, when 

feeding of ruminants in tropical countries is most 

critical (Mgheni et al., 2001). However, crop 

residues have some drawbacks such as high 

content of cell wall, low protein and low utilization 

rate (Hindrichson et al., 2001). The addition of 

foliage from tree leaves or supplementation with 

seed meals or even urea can improve the utilization 

of low quality roughages, mainly through the 

supply of nitrogen to the rumen microbes. In 

addition, Wanapat et al. (1997) reported that 

cassava hay might be a good source of protein 

supplement for ruminants in dry season, 

particularly on a small-holder dairy farm. 

Kraphanghom (Paederia foetida Linn) is also 

available in dry season and has high protein 

content, making it a useful supplement for 

ruminants in dry season. 

The degree of nutrient degradation 

occurring in the rumen has major influence on the 

total utilization of nutrient in feedstuffs. In sacco 

nylon bag technique is often used to characterize 

rumen fermentation kinetics of nutrient in the 

rumen. In spite of numerous studies conducted on 

the used of crop residue as ruminant feed, limited 

information is available about ruminal 

degradation. There is also little research that 

characterizes individual feeds (DePeters et al., 

1997). Therefore, the aim of this study was to 

assess the chemical composition and ruminal 

degradation characteristics of crop residues and 

selected roughages using the in sacco nylon bag 

technique. 


Crop residues and selected roughages 

preparation and analysis 

The crop residues and selected 

roughages, 1) water hyacinth (Eichhorunia 


crassipes Solms, WH), 2) kraphanghom (Paederia 

foetida Linn, KH), 3) corn stover (Zea mays, CS), 

4) cassava hay (Manihot esculenta Crantz, CH), 

5) sugarcane top (Saccharum officinarum Linn, 

SCT), 6) Chinese spinach (Amaranthus viridis L., 

CP), 7) rice straw (Oryza sativa, RS) and 8) 

cavalcade hay (Centrosema pascuorum cv. 

Cavalcade, CC) were collected from Maha 

Sarakham province area in the North-East of 

Thailand. Fresh samples (1 kg.) were harvested 

by hand from three specimens. Duplicate fresh 

samples (0.5 kg. /replicate) were dried in a hot, 

dry air force oven at 65 °C for 72 h and weighed. 

The samples (Table1) were then ground to 

pass through a 1 mm screen for nylon bag 

incubation and chemical analysis. The samples 

were analyzed for dry matter (DM), crude 

protein (CP) and ash (AOAC, 1990). Neutral 

detergent fiber (NDF), acid detergent fiber 

(ADF) and acid detergent lignin (ADL) were 

assayed using the method proposed by Van Soest 

et al. (1991). 

In sacco degradation procedure 

Ruminal degradation measurement using 

the nylon bag technique was carried out in 

Brahman-Thai native cross bred after a two week 

adaptation period. The steer with an average body 

weight of 250±15 kg and fitted with a permanent 

rumen cannula were offerred rice straw ad libitum 

and received concentrate at 0.5 % BW (concentrate 

mixture: 49.8% cassava chip, 17.5% rice bran, 

14.6% palm meal, 7.0% soybean meal, 1.4% urea, 

0.4% salt, 1.0 % mineral mix and 8.3% sugarcane 

molasses). Approximately 3.0 g (as fed basis) of 

each test feed was accurately weighed into 

synthetic bag with a mean pore size of 45. Bag 

plus the sample were placed into the rumen of the 

beef steer, 30 min after the morning meal and 

retrieved after a period of 3, 6, 12, 24, 48 and 72 h 

(four bags of each feed for each period). After 

removal from the rumen, bags were rinsed in pipe 

line fresh water and washed by hand under tap

206 

water until the water became clear. After washing, 

the bags were placed into a hot dry air force oven 

at 65 °C for 48 h and weighed. To determine the 

content of water soluble material bags, 

representing 0 h degradation also underwent the 

same washing procedure as the incubated bags. 

Dried residues of each incubation time from each 

steer were pooled, for DM, OM and CP analyzed, 

then DM, OM and CP disappearance values were 

calculated for the difference between nutrient 

weight before and after incubation of each sample. 

The degradability data obtained for DM, OM and 

CP for each feed was fitted to the equation P= a+b 

(1-e -ct ) (∅rskov and McDonald, 1979). The 

effective degradability was calculated as 

ED=a+b{c/(c+k)}, where K = fractional passage 

rate (0.02/h) 


All data obtained were subjected to the 

analysis of variance (ANOVA) procedure 

according to the Complete Randomized Design 

using the general linear (GLM) of the SAS system 

(SAS, 1996). Treatment means were compared 

using Duncan’s New Multiple Range test. 

Probabilities less than 0.05 were considered to be 

significant. 



Chemical composition of crop residue and 

selected roughage 

The chemical compositions of crop 

residues and selected roughages are presented in 

Table 1. The crude protein content of the crop 

residues and selected roughages ranged from 3.0 

to 26.42 %. Rice straw had the lowest crude protein 

content, while the Chinese spinach had the highest 

crude protein content. Similar crude protein 

content were observed in krapanghom and cassava 

hay. When comparing crude protein content of 

krapanghom and cassava hay with alfalfa hay as 

reported by Alcaide (2000), it was found that the 

crude protein content of krapanghom, cassava hay 

and alfalfa hay were similar. Low protein content 

was observed in crop residues (corn stover, 

sugarcane top and rice straw). The result of study 

agrees with Hindrichson et al. (2001) who reported 

that most crop residues were low in protein and 

high in fibrous content. The crude protein content 

of rice straw was lower than that reported by Liu 

et al. (2002). However, the crude protein content 

of rice straw was similar to that reported by 

Fonseca et al. (1998) and Department of Livestock 

Development (2004). The crude protein content 

Table 1 Chemical composition of crop residues and selected roughages. 

Feedstuffs 1 DM (%) CP Ash NDF ADF ADL 

……………..%DM basis………………. 

WH 15.0 12.9 14.3 69.2 42.7 3.7 

KP 21.7 16.6 8.6 50.2 45.3 11.8 

CS 23.3 6.3 6.3 67.4 38.4 4.1 

CH 24.0 15.8 8.7 50.9 51.4 12.6 

SC 36.7 5.8 5.3 79.9 54.6 8.9 

CN 15.6 26.4 23.3 40.1 19.9 4.9 

RS 91.5 3.0 13.6 72.1 53.3 4.9 

CC 94.5 10.9 7.6 59.9 41.6 11.5 

DM = dry matter, CP = crude protein, NDF = neutral detergent fiber, 

ADF = acid detergent fiber and ADL = acid detergent lignin, 1 WH = water hyacinth, 

KH = kraphanghom, CS = corn stover, CH = cassava hay, SC = sugar cane top, 

CN = Chinese spinach, RS = rice straw, CC = cavalcade hay

of cassava hay was lower than that reported by 

Promkot and Wanapat (2004) and Wanapat et al. 

(1997). The difference of crude protein content 

was probably due to maturity level and leavesstem 

ratio of cassava hay. Normally, the crude 

protein content of cassava hay decreased as 

maturity increased (Wanapat, 2003). The crude 

protein content of sugarcane top was higher than 

that reported by Kawashima et al. (2002). In 

addition, crude protein content of corn stover was 

higher than those reported by Hindrichson et al. 

(2001) and Mgheni et al. (2001). There are many 

factors that affect crude protein content such as 

stage of growth (Promkot and Wanapat, 2004) 

maturity and species or variety and soil types 

(Baloyi et al., 1997). These factors may partially 

explain differences in crude protein content 

between our study and others. 

Ash content of crop residue and selected 

roughage was ranged from 5.27 to 23.31%. 

Sugarcane top had the lowest ash content while 

the Chinese spinach had the highest. The ash 

content of rice straw was lower than that reported 

by Department of Livestock Development (2004), 

but higher than that reported by Liu et al. (2001). 

The difference of ash content was probably due to 

variety of rice straw (Agbagla-Dohnani et al., 

2001) and soil type (Thu and Preston, 1999). 

However, ash content of rice straw was similar to 

that reported by Fonseca et al. (1998). In addition, 

the ash content of rice straw was similar to ash 

content of water hyacinth. The ash content of corn 

stover was lower than that reported by Hindrichson 

et al. (2001) and Magheni et al. (2001). Ash 

content of sugarcane top, water hyacinth and 

cassava hay were similar to previous reports (Thu 

and Preston, 1999; Kawashima et al., 2002). 

Neutral detergent fiber (NDF) content of 

crop residues and selected roughages ranged from 

40.06 to 79.93%. Chinese spinach had the lowest 

NDF content while sugarcane top had the highest. 

Similar NDF content was observed in water 

hyacinth, corn stover and rice straw. The NDF 


content of rice straw was higher than that reported 

by Department of Livestock Development (2004) 

and Fonseca et al. (1998). However, NDF content 

of rice straw was similar to that reported by Liu et 

al. (2002). The NDF content of corn stover was 

lower than that reported by Hindrichson et al. 

(2001). Neutral detergent fiber contents of 

sugarcane top, water hyacinth and krapanghom 

were all similar to previous reports (Thu and 

Preston, 1999; Kawashima et al., 2002; 

Department of livestock Development, 2004) 

Acid detergent fiber (ADF) content of 


19.96 to 54.61%. Chinese spinach had the lowest 

ADF content while the sugarcane top had the 

highest ADF content. Similar ADF content was 

observed in water hyacinth, krapanghom and 

cavalcade hay. The ADF content of rice straw was 

higher than that reported by Department of 

Livestock Development (2004) and Thu and 

Preston (1999), but similar to reports by Liu et al 

(2002). The ADF content of corn stover was 

similar to that reported by Department of Livestock 

Development (2004). In addition, ADF content of 

water hyacinth and sugarcane top were higher than 

those previously reported (Thu and Preston, 1999; 

Kawashima et al., 2002) 

There are many factors that may affect 

fibrous (NDF and ADF) content such as stage of 

growth (Promkot and Wanapat, 2004), maturity 

and species or variety (Agbagla-Dohnani et al., 

2001), dried method and growth environment 

(Mupangwa et al., 1997) and soil types (Thu and 

Preston, 1999). These factors may partially explain 

differences in fibrous content between our study 

and others. 

Acid detergent lignin (ADL) content of 


3.67 to 12.64 %. Water hyacinth had the lowest 

ADL content while the brown salwood had the 

highest. Similar ADL content were observed in 

krapanghom, cassava hay and cavalcade hay. The 

ADL content of rice straw was lower than that

208 

reported by Thu and Preston (1999), but similar 

to other previous reports (Fonceca et al., 1998 and 

Department of Livestock Development, 2004). 

The difference of ADL content was probably due 

to difference in variety of rice straw (Agbagla- 

Dohnani et al., 2001) and soil type (Thu and 

Preston, 1999). 

Degradability characteristics 

The rapidly soluble fraction (a fraction), 

potentially degradable fraction (b fraction), rate 

of degradation of b fraction (c) and potential 

degradation (a+b) are presented in Table 2. Dry 

matter a fraction was highest (P


Table 2 In sacco degradation characteristic and effective degradability of crop residues and selected roughages. 

Parameters2 Treatment1 SEM 

WH KH CS CH SC CN RT CC 

DM degradation 

a, % 20.12d 27.58c 30.94b 20.65d 18.00 e 39.10 a 13.27f 29.54bc 1.24 

b, % 68.68a 61.53ab 53.04ab 67.56a 53.49ab 48.82b 62.49 ab 55.39ab 2.01 

c, h-1 0.011c 0.025b 0.022bc 0.020bc 0.018 bc 0.051 a 0.018bc 0.016bc 0.001 

a+b, % 88.81 89.12 83.98 88.21 71.49 88.92 75.77 84.94 2.05 

EDDM, % 44.35f 61.63b 58.49c 54.52d 40.83 f 74.48 a 41.38 f 52.60d 1.95 

OM degradation 

a, % 13.62e 24.60c 28.49b 20.85d 15.07 e 34.43 a 9.33 f 26.82b 1.45 

b, % 71.86ab 63.33ab 66.81ab 74.80a 56.16 b 56.11 b 75.78 a 57.06b 2.02 

c, h-1 0.011c 0.028b 0.017bc 0.0.016c 0.017 ab 0.044 a 0.014 bc 0.016bc 0.001 

a+b, % 85.49abc 87.93abc 95.30a 95.65a 71.24c 90.55 ab 84.0 abc 83.88abc 2.30 

EDOM, % 38.82f 61.21b 58.07c 53.75d 38.45 f 72.07a 40.47 f 51.20e 2.05 

CP degradation 

a, % 33.73c 27.44e 30.24d 12.48g 32.69cd 43.31b 28.74f 33.96a 1.83 

b, % 17.72e 72.56a 54.94b 74.22a 37.97d 45.91c 57.17cd 44.81c 3.21 

c, h-1 0.039b 0.025cd 0.017d 0.019cd 0.015d 0.066a 0.004e 0.057a 0.001 

a+b, % 51.45e 100.0a 85.18c 89.60bc 70.66d 89.22b 85.9bc 78.77b 2.80 

EDCP, % 45.44f 67.62c 55.19d 48.90e 49.05e 77.65a 38.87e 67.20b 2.10 

a, b, c, d, e Means within a row different superscripts differ (P

210 

cavalcade hay. The crude protein b fraction ranged 

from 17.72 to 72.56%. Crude protein b fraction 

was highest for krapanghom and lowest for water 

hyacinth. Similar crude protein b fraction was 

observed in Chinese spinach and cavalcade hay. 

The crude protein b fraction for water hyacinth 

was similar to those in previous studies (Khan et 

al., 2002). Crude protein b fraction for cassava 

hay was higher than those reported by Wanapat et 

al. (1997) and Promkot and Wanapat (2004). In 

addition, crude protein b fraction for rice straw 

was lower than that reported by Mgheni et al. 

(2001), but for corn stover was higher than that 

reported by Mgheni et al. (2001). Numerous 

factors may have influenced the differences in b 

fraction in this study. 

Degradation rate (c) of dry matter was 

fastest (P0.05). Potential degradation of dry 

matter for rice straw was lower than that reported 

by Mghani et al. (2001), but higher than that 

reported by Keir et al. (1997). Moreover, potential 

degradation of dry matter for cassava hay was 

higher than that reported by Promkot and Wanapat 

(2004). Potential degradation of organic matter 

was highest (P

esult into low degradability and induce low intake. 

The structure and solubility characteristics of 

protein in feedstuffs influence crude protein 

degradability in the rumen (Mahadevan et al., 

1980). Crude protein degradability of crop residue 

and selected roughages are much lower than in 

common concentrated feed (Woods et al., 2003), 

leguminous (Khandaker and Tareque, 1996), and 

aquatic plant (Khan et al. 2002). 

Numerous factors had effect of in sacco 

degradability, such as bag pore size (Vanzant et 

al., 1998), sample size (Nocek, 1985), washing 

procedures, grinding, diet of host animal, species 

of animal, sample preparation, incubation time and 

washing method (Olivera, 1998). Maturity of 

perennial forages affected all degradation fractions 

and degradation rate of dry matter and crude 

protein (Hoffman et al., 1993). Furthermore, 

chemical composition and processing of feedstuffs 

affected degradation characteristics (Huntington 

and Gives, 1997). Vitti et al. (1999) reported in 

sacco dry matter disappearance was highly 

correlated (P

212 

CONCLUSIONS 

Ruminal disappeared characteristic of 

crop residue and selected roughage differed among 

feedstuffs. Crop residue and selected roughage 

showed a great variation in chemical composition 

and degradability. The result in this study indicated 

that crop residues and selected roughages have a 

degradability ranked from the highest to the lowest 

were; Chinese spinach, krapanghom, cassava hay, 

corn stover, cavalcade hay water hyacinth, rice 

straw and sugarcane top, respectively. Importantly, 

crop residues and selected roughages are abundant 

and available for feeding ruminants in the dry 

season. 


The authors would like to express their 

gratitude to Mr. Pay Rodthong for taking care of 

the animal and the Department of Agricultural 

Technology, Faculty of Technology 

Mahasarakham University and Department of 

Animal Science, Faculty of Agriculture Khon 

Kaen University for supporting experiment 

facilities. This financial support for this research 

was partially provided by ‘The Khon Kaen 

University’s Graduate Research Fund, Academic 

Year 2003’. 


Agbagla-Dohnani, A., P. Noziere, G. Clement and 

M. Doreau. 2001. In sacco degradability 

chemical and morphological composition of 

15 varieties of European rice straw. Anim. 

Feed Sci. Technol. 94: 15-27. 

Alcaide, E.M., A.I. Martin Garcia and J.F. 

Aguilera. 2000. A comparative study of 

nutrient digestibility, kinetics of degradation 

and passage and rumen fermentation pattern 

in goats and sheep offered good quality diets. 

Live. Pro. Sci. 64: 215-223. 


AOAC. 1990. Official Methods of Analysis, 

Vol.1, 15 th Edition. Association of Official 

Analytical Chemists, Arlington, Virginia, 

USA. PP. 69-90. 

Baloyi, J.J., N.T. Ngongoni, J.H. Topps and P. 

Ndlovu. 1997. Chemical composition and 

degradability of Brachystegia spiciformis 

(Musasa) leaves and stems harvested over 4 

months from three sites in Zimbabwe. Anim. 


Broderick, G.A. 2003. Effect of varying dietary 

protein and energy levels on the production 

of lactating dairy cows. J. Dairy Sci. 86: 

1370-1381. 

Chinh, B.V., L.V. Ly, N.H. Tao, N.V. Hai and T.B. 

Ngoc. 2000. Study on processing, storing and 

using sugarcane leaves as ruminant feed. 

Workshop-seminar”Making better use of 

local feed resources” SAREC-UAF, January, 

2000. 

Department of Livestock Development, Ministry 

of Agricultural and Cooperative. 2004. Table 

of Chemical Composition of Feedstuffs. 

Chumnum Sahakorn HeangpadhesThai Ltd. 

37 p. 

Djajanegara, A. and P.T. Doyle. 1989. Urea 

supplementation compared with pre treatment. 

Effect on intake, digestion and live-weight 

change by sheep fed rice straw. Anim. Feed 

Sci and Technol. 25: 21-36. 

Fonseca, A.J.M., A.A. Daias-da-Silva and E.R. 

∅rskov. 1998. In sacco degradation 

characteristics as predictors of digestibility 

and voluntary intake of roughages by mature 

ewes. Anim. Feed Sci. and Technol. 72: 205- 

219. 

Hindrichson, I.K., P.O. Osuiji, A.A. Odenyo, J. 

Madsen and T. Hvelplund. 2001. Effect of 

supplementation with four multipurpose trees 

and Lablab purpureus on rumen microbial 

population, rumen fermentation, digesta 

kinetics and microbial protein supply of sheep 

and maize stover ad libilitum, pp. 22-38. In

Proc. TSAP Vol. 28. 

Hoffman, P.C., S.J. Sievert, R.D. Shever, D.A. 

Welch and D.K. Coms. 1993. In situ dry 

matter, protein and fiber degradation of 

perennial forages. J. Dairy Sci. 76: 2632- 

2643. 

Huntington, J.A. and D.I. Givens. 1997. Studies 

on in situ degradation of feeds in the rumen: 

1 Effect of species, bag mobility and 

incubation sequence on dry matter 

disappearance. Anim. Feed. Sci. Technol. 64: 

227-241. 

Ibrahim, M.N.M., S. Tamminga and G. 

Zemmelink. 1995. Degradation of tropical 

roughages and concentrate feeds in the rumen. 

Anim. Feed Sci. Technol. 54: 81-92. 

Kawashima, T., W. Sumamal, P. Pholsen, R. 

Chaithiang, W. Boonpakdee, M. Kurihara and 

M. Shibata. 2002. Feeding value of sugarcane 

stalk for cattle. Asian-Aust. J. Anim. Sci. 

15(1): 55-60. 

Keir, B., N.V. Lai, T.R. Preston and E.R. ∅rskov. 

1997. Nutritive value of leaves from tropical 

trees and shrubs: 1. In vitro gas production 

and in sacco rumen degradability. Livestock 

Research for Rural Development. 9(4): 1- 

8. 

Khan, M.J., H. Steingass and W. Drochner. 2002. 

Evaluation of some aquatic plants from 

Bangladesh through mineral composition, In 

vitro gas production and in situ degradation 

measurements. Asian-Aust. J. Anim. Sci. 

15(4): 537-542. 

Khandaker, Z.H. and A.M.M. Tareque. 1996. 

Studies on protein degradability of feedstuffs 

in Bangladesh. Asian-Aus. J. Anim. Sci. 

9 (6): 637-642. 

Liu, J.X., A. Susenbeth and K.H. Sudekum. 2002. 

In vitro gas production measurement to 

evaluate interactions between untreated and 

chemically treated rice straws, grass hay, and 

mulberry leaves. J. Anim. Sci. 80: 517-524. 

Lopez, S., F.D. De, B. Hovell, B. Manyuchi and 


R.I. Smart.1995. Comparison of sample 

preparation methods for determination of the 

rumen degradation characteristics of fresh and 

ensiled forages by nylon bag technique. Anim. 

Sci. 70: 439-450. 

Mahadevan, S., J.D. Eefle and F.D. Sauer. 1980. 

Degradation of soluble and insoluble proteins 

by rumen microorganisms. J. Anim. Sci. 50: 

723-728. 

Mgheni, D.M., E.E. Ndemanisho, T. Hvelplund 

and M.Weisbjerg. 2001. Evaluation of the 

feeding value of two tropical cereal straw, 

maize stover, rice straw and their botanical 

fractions by nylon and mobile bag technique. 

African Journal of Science and Technilogy 

(AJST). 2: 65-71. 

Mupangwa, J.F., N.T.Ngongoni, J.H.Topps and P. 

Ndlovu. 1997. Chemical composition and dry 

matter of forage legumes Cassia rotundiforlia 

cv. Wynn, Lablab purpureus cv. Highworth 

and Macroptilium atropurpureum cv. Siratro 

at 8 weeks of growth (pre-anthesis). Anim. 


NRC. 2001. Nutrient Requirements of Dairy 

Cattle. (7 th rev.ed.) National research council, 

National Academy Press, Washington, DC. 

Nocek, J.E. 1985. Evaluation of specific variables 

affecting in situ estimates of ruminal dry 

matter and protein digestion. J. Anim. Sci. 60: 

1347. 

Olivera, R.M.P.1998. Use of in vitro gas 

production technique to assess the 

contribution of both soluble and insoluble 

fractions on the nutritive value of forages. 

MS. thesis, University of Aberdeen, Scotland. 

Promkot, C. and M. Wanapat. 2004. Ruminal 

degradation and intestinal digestion of crude 

protein of tropical resources using nylon bag 

and three-step in vitro procedure in dairy 

cattle, pp. 434-447. In Proc. of the 

Agricultural Seminar, Animal Science/ 

Animal Husbandry. Held at Sofitel Raja 

Orchid Hotel 27-28 January 2004.

214 

∅rskov, E.R. and I. McDonald. 1979. The 

estimation of protein degradability in the 

rumen from incubation measurements 

weighted according to rate of passage. J. Agri. 

Sci. (Camb). 92: 499-504. 

SAS. 1996. SAS User’s Guide: Statistics, Version 

6.12 th ed. SAS Institute Inc. Cary, NC. 846 p. 

Sinclair, L.A., P.C. Garnsworthy, J.R. Newbold 

and P.J. Buttery. 1993. Effect of synchronizing 

the rate of dietary energy and nitrogen release 

on rumen fermentation and microbial protein 

synthesis in sheep. J. Agric. Sci. (Camb). 120. 

251-263. 

Shem, M.N., E.R. ∅rskov and A.E. Kimambo. 

1995. Prediction of voluntary dry matter 

intake, digestible dry matter intake and growth 

rate of cattle from the degradation 

characteristics of tropical foods. Anim. Sci. 

60-74. 

Thu, N.V. and T.R. Preston. 1999. Rumen 

environment and feed degradability in swamp 

buffaloes fed different supplements. 

Livestock Research for Rural Development. 

11 (3): 1-7. 

Van Soest, P.J., J.B. Robertson. and B.A. Lewis. 

1991. Methods for dietary fiber, neutral 

detergent fiber and non starch poly saccharine 

in relation to animal nutrition. J. Dairy Sci. 

74: 3583-3597 


Vanzant, E.S., R.C. Cochan and E.V. Titgemeyer. 

1998. Standardization of in situ techniques for 

ruminant feedstuffs evaluation. J. Anim. Sci. 

76: 2717-2729. 

Vitti, D.M., A.L. Abdalla, J.C. Silva Filho, N.L. 

del Mastro, R. Mauricio, E. Oven and F. 

Mould.1999. Misleading relationships 

between in situ rumen dry mater 

disappearance, chemical analyzed and in vitro 

gas production and digestibility, of sugarcane 

baggage treated with varying levels of electron 

irradiation and ammonia. Anim. Feed. Sci. 

Technol. 79: 145-153. 

Wanapat, M. 2003. Manipulation of cassava 

cultivation and utilization to improved protein 

to energy biomass for livestock feeding in 

tropics. Asian-Aust J. Anim. Sci. 16: 463- 

471. 

Wanapat, M., O. Pimpa, A. Petlum and U. 

Boontao.1997. Cassava hay: A new strategic 

feed for ruminant during the dry season. 

Livestock Research for Rural development. 

9: 1-5. 

Woods, V.B., F.P.O. Mara and A.P. Moloney. 2003. 

The nutritive value of concentrates feedstuffs 

for ruminant animals. Part I: In situ ruminal 

degradability of dry matter and organic matter. 

Anim. Feed Sci. Technol. 110: 111-130.


Comparative Efficiency of KU and ISO Plungers 

in Mixing Composite Bulk Raw Milk 

Jigme Wangdi1 *, Pravee Vijchulata1 , 

Pornsri Chairatanayuth1and Suwapong Swasdiphanich2 ABSTRACT 

The study was conducted to investigate the best mixing indicators among the milk composition 

i.e. milk protein, fat and SNF and to determine the efficiency of two plungers, i.e. Kasetsart University 

(KU) and International organization for standardization (ISO) for raw milk truck. Fat content was found 

to be the best indicator for mixing milk prior to sampling. Fat content stabilized after 10, 5 and 10 times 

of stirring raw milk truck’s chamber I, II, and III of 4,700, 3,900 and 7,400 kg capacities respectively; 

whereas the protein and SNF stabilized after 5 times of stirring irrespective of the chamber sizes in the 

study. Significant difference (p

216 

is constrained mainly due to factors, i.e., the wide 

variety of style and design of tanks, methods of 

agitation, volume of the products held and types 

of milk agitators (Grace et al., 1992). This problem 

was further aggravated by the creaming 

phenomenon, which in a tank was said to depend 

on the time that has elapsed from the completion 

of cooling of the milk until it was agitated for 

sampling (Likas and Calbert, 1954). Therefore, it 

is necessary to determine the adequate agitation 

time for the individual tank prior to sampling. 

However, there are no prescribed international 

standards on mixing length, besides the general 

consensus that the adequate agitation times are five 

minutes for tanks under 1,000 gallons and 10 min 

for tanks of 1,000 gallons and more (Goodridge 

et al., 2004). Such a long duration of mixing would 

not be necessary, considering the improvement in 

management of milk handling and transportation 

system. Several agitation methods, i.e. manual, 

mechanical and air compressor were adopted. In 

most developed countries milk trucks were 

installed with mechanical agitator and milk during 

transportation were periodically agitated, thereby 

reduced the time of agitation prior sampling. 

However, impacts of intermittent mixing on the 

quality of milk were not studied (Goodridge et al., 

2004). 


Figure 1 Different manual plungers; a) ISO standard and b) KU. 

In Thailand no regulation exists on the 

standard requirement for the bulk milk trucks. As 

such most raw milk trucks are not furnished with 

mechanical agitator. Therefore no intermittent 

mixing of milk takes place during transportation. 

Thus, obtaining representative sample at the dairy 

plant fully depended on the efficiency of manual 

plunging. Different types of locally designed 

plunger are adopted in mixing milk prior to 

sampling. This could be due to limited information 

available on the efficiency of ISO standard plunger. 

Therefore, this study was planned to investigate 

the best milk composition to be adopted as mixing 

indicator and also to compare the efficiency of KU 

designed versus the ISO standard plungers. 


The experiment was conducted on the 

chambers of bulk milk truck delivering raw milk 

to KU Dairy Center after approximately 3.5-4 hrs 

of transportation from Milk Collection Center 

(MCC). The capacity of the chambers was 4,700 

kg, 3,900 kg, and 7,400 kg for chamber I, II and 

III respectively. Two different plungers with 

following specification were used; a) the ISO 

standard made of stainless steel rod of 2 m in 

length, fitted with a 300 mm diameter disc

perforated with 12 holes each 30 mm in diameter 

on a circle of 230 mm in diameter from the center 

(ISO/FDIS, 1997) and b) the KU design made of 

stainless steel rod of 1.80 m in length, fitted with 

a stainless disc of 230 mm in diameter perforated 

with 4 holes each 62 mm in diameter on a circle 

of 205 mm in diameter from the centre. 

Raw milk samples 

Prior to transferring milk from the truck 

to storage tank, milk from each chamber was 

mixed at uniform interval of 0, 5, 10, 15, 20, 25 

and 30 times of stirring using both plungers. 

During each sampling time, 60 ml of raw milk 

were collected in plastic bottles of 80 ml capacity 

with the help of plastic jug from the top opening 

of each chamber. Samples were collected from all 

three chambers adopting the similar method and 

were analyzed for fat content and other milk 

compositions at the laboratory using calibrated 

ultrasonic milk analyzer, Ekomilk-M (Eon 

Trading, 2001). 


Data were analyzed under a split plot in 

Completely Randomized Design (CRD) using 

PROC GLM in SAS. The statistical model used 

was 

Yijkl = µ + ρi + δk(i) + αj + β(ρ)il +ραij 

+ εijkl 

where Yijkl was the dependent variable; 

µ was the overall mean; ri was the effect of 

chambers as main plots (i = 1, 2, 3), δk(i) was 

the main plot error; αj was the effect of stirrers 

(j = 1, 2); βl was the effect of stirrers (l = 1 and 2); 

β(ρ)il stirring times nested with containers; ραij 

was the interaction effect between the plunger and 

container, and εijkl was the subplot error at NID 

(0, σe 2 ). 

Statistical differences among different 

stirring times were analyzed using Duncan’s 

multiple range test (DMRT) according to Cody 

and Smith (1997). 



Milk composition 

The changing analysis values of milk 

composition sampling at various stirring times are 

illustrated in Table 1-3 for protein, solid-not fat 

(SNF) and fat respectively. The protein and SNF 

content in milk mixed with both plungers, 

stabilized after 5 times of stirring with no 

significant differences (p>0.05) with the rest of 

stirring times (Table 1 and 2), whereas the milk 

fat content statistically stabilized only after 10, 5 

and 10 times of stirring for chamber I, II and III 

respectively (Table 3). The shortest time to attain 

homogeneity in protein content could be explained 

by its nature of existence in dispersion form of a 

very tiny particle size ranging from 10 -4 to 10 -5 mm 

(Tetrapak, 1995), which could be easily disturbed 

and redistributed with mild agitation. Whereas, 

early stabilization of SNF could be explained by 

early stabilization of protein, which was one of 

the SNF components and the other water soluble 

components of SNF, i.e. lactose, minerals and 

certain vitamins which existed in true solution. The 

early stabilization of protein and SNF could also 

be in part explained by the nature of variation of 

milk components, where fat showed the widest 

variation, followed by protein, lactose and then 

ash (Table 4) (Jenness, 1988). The difficulty in 

stabilization of fat content was caused by 

stratification of fat globules during storage or 

transportation from gravity creaming and also from 

the action of cold agglutinin. Fat globule tended 

to float due to lower density as compared to skim 

milk; depleting the fat content at the lower portion 

of the container forming cream layer at the surface. 

Creaming was reported to onset after 40-50 minute 

of milk holding after cooling to 5 °C (Goodridge 

et al., 2004). 

Comparative plunger efficiency 

Comparative effectiveness of KU versus 

ISO standard designed plungers using protein,

218 

SNF and fat composition in raw milk as indicators 

are illustrated in Table 5-7. Using both types of 

plunger, protein and SNF stabilized after 5 times 

of agitation (Table 5 and 6). Milk fat content also 

attained homogeneity (p


Table 4 Variations in the composition of raw milk. 

Samples Trocher (1925) Overman et al. (1939) Herrington et al.(1972) 

676 individual milking 2426 3-day composites 868 bulk tank 

Fat (%) Mean 3.95 4.37 3.53 

SD 0.78 0.82 0.28 

CV 0.20 0.19 0.08 

Crude Mean 3.24 3.74 3.13 

Protein (%) SD 0.40 0.52 0.14 

CV 0.12 0.14 0.05 

Lactose (%) Mean 4.64 4.89 4.82 

SD 0.37 0.38 0.16 

CV 0.08 0.08 0.03 

Ash (%) Mean 0.70 0.72 0.72 

SD 0.05 0.05 0.01 

CV 0.07 0.07 0.02 

Total solid (%) Mean - 13.73 12.02 

SD - 1.23 0.63 

CV - 0.09 0.05 

Sources: Compiled by Jenness (1988). 

Table 5 Least Square means ±SE of protein (%) used as indicator for the efficiency of plungers. 

Stirring number Chamber I Chamber II Chamber III 

KU ISO KU ISO KU ISO 

0 2.69 ± 0.05b 2.81 ± 0.04b 2.43 ± 0.05b 2.57 ± 0.05b 2.72 ± 0.04b 2.45 ± 0.06b 5 3.07 ± 0.05 a 3.03 ± 0.04 a 3.08 ± 0.05 a 3.08 ± 0.05 a 3.05 ± 0.04 a 3.09 ± 0.06 a 

10 3.08 ± 0.05 a 3.05 ± 0.04 a 3.08 ± 0.05 a 3.08 ± 0.05 a 3.05 ± 0.04 a 3.09 ± 0.06 a 

15 3.08 ± 0.05 a 3.05 ± 0.04 a 3.07 ± 0.05 a 3.08 ± 0.05 a 3.06 ± 0.04 a 3.08 ± 0.06 a 

20 3.09 ± 0.05 a 3.05 ± 0.04 a 3.08 ± 0.05 a 3.09 ± 0.05 a 3.08 ± 0.04 a 3.08 ± 0.06 a 

25 3.08 ± 0.05 a 3.06 ± 0.04 a 3.08 ± 0.05 a 3.09 ± 0.05 a 3.08 ± 0.04 a 3.09 ± 0.06 a 

30 3.08 ± 0.05 a 3.08 ± 0.04 a 3.07 ± 0.05 a 3.08 ± 0.05 a 3.08 ± 0.04 a 3.09 ± 0.06 a 

N of individual stirring number for each chamber = 10. 

ab Means with different superscripts within same column are significantly different (p

220 

in terms of both manual sampling management 

and the relatively large variable of milk 

composition values between different intervals. 

The difference of analytical values are far beyond 

the acceptable range recommended by Grace et 

al. (1992) that “ adequate agitation is that degree 

of agitation which, at full tank capacity, results in 

a variation in fat content of the milk in the tank of 

not more than ± 0.1 % level as determined by an 

official AOAC milk fat test”. Adhering to the 

above recommendation by using the variable of 

fat composition as indicator, a trend requiring 

different mixing times for different chambers were 

evident in milk agitated with the two plungers 

(Table 8). More stirring times with the increase in 


the sizes of the chambers were observed. Milk in 

Chamber I and II agitated with KU plunger 

attained homogeneity after 15 times of stirring, 

whereas 20 times of stirring was required for 

chamber III (Table 9). Similar trend requiring more 

times to mix the milk homogeneously as the 

volume of milk in the tank increases were observed 

(Likas and Calbert, 1954). This might be due to 

small surface area of the disc which in turn could 

not produce sufficient disturbances to disperse the 

clustered milk fat that was formed during 

transportation time. The plunger should be long 

and the disc should be large enough to provide 

sufficient disturbances during mixing (ISO/FDIS, 

1997). On the other hand, the milk agitated with 

Table 7 Least square means ±SE of fat content (%) as indicator in for the efficiency of different plungers. 

Stirring Chamber I Chamber II Chamber III 

number N KU N ISO N KU N ISO N KU N ISO 

0 7 9.71 ± 0.31a 8 9.87 ± 0.20a 4 9.43 ± 0.41a 10 9.33 ± 0.20a 10 9.40 ± 0.28a 10 9.66 ± 0.24ax 5 10 4.72 ± 0.26 bc 10 5.51 ± 0.18 b 10 4.44 ± 0.26 b 10 4.91 ± 0.17 b 10 4.49 ± 0.26 b 10 4.98 ± 0.24 b 

10 10 4.31 ± 0.26 c 10 4.58 ± 0.18 c 10 4.30 ± 0.26 b 10 4.60 ± 0.17 b 10 4.40 ± 0.26 b 10 4.65 ± 0.24 bc 

15 10 3.97 ± 0.26 c 10 4.29 ± 0.18 cd 10 4.02 ± 0.26 b 10 4.24 ± .17 cd 10 4.18 ± 0.26 b 10 4.32 ± 0.24 c 

20 10 3.86 ± 0.26 c 10 4.15 ± 0.18 d 10 3.99 ± 0.26 b 10 4.13 ± 0.17 d 10 3.95 ± 0.26 b 10 4.07 ± 0.24 c 

25 10 3.83 ± 0.26 c 10 4.09 ± 0.18 d 10 3.97 ± 0.26 b 10 4.10 ± 0.17 d 10 3.85 ± 0.26 b 10 4.05 ± 0.24 c 

30 10 3.89 ± 0.26 c 10 4.06 ± 0.18 d 10 3.98 ± 0.26 b 10 4.04 ± 0.17 d 10 3.88 ± 0.26 b 10 4.05 ± 0.24 c 

abc Means with different superscripts within the same column are significantly different (p

ISO standard plunger attained homogeneity of 

milk after 20 times of stirring for all chambers. 

More time of stirring required to mix the milk in 

the small chamber could be attributed to the 

inconvenience encountered during mixing motion 

of the plunger with large disc diameter. The 

occasional contact of plunger to the side surface 

of the tanks while mixing resulted in non 

uniformity of mixing cycles. On the other hand, 

the increase in stirring time of KU plunger for large 

chamber was attributed to its short stirring rod. A 

more complete mixing motion diameter could be 

attained with a slight improvement of its length. 

This emphasized that the proportionality of 

plunger to the size of containers was imperative 

for effective mixing. 

CONCLUSIONS 

Among the three milk constituents 

determined for attaining homogeneity in mixing 

raw milk prior to sampling, milk fat was found to 

be the best indicator. Comparing KU versus ISO 

designed plungers for manually mixing milk in 

the milk truck chamber prior to sampling, the KU 

plunger performed relatively better for small size 

milk chamber. However, both plungers were equal 

in their effectiveness in mixing milk for larger 

chamber. With slight increase in the length of the 

KU plunger rod, at least 15 times of proper stirring 

motion were required for chamber of less than 


Table 9 Difference in mean between the fat test (%) of different mixing times with that of composite 

fat test for all chambers. 

Stirring number Chamber I Chamber II Chamber III 

KU ISO KU ISO KU ISO 

0 5.82 5.81 5.45 5.29 5.52 5.61 

5 0.83 1.45 0.46 0.87 0.61 0.93 

10 0.42 0.52 0.32 0.56 0.52 0.6 

15 0.08 0.23 0.04 0.2 0.3 0.27 

20 -0.03 0.09 0.01 0.09 0.07 0.02 

25 -0.06 0.03 -0.01 0.06 -0.03 0 

30 0 0 0 0 0 0 

7,500 kg capacity. As for the ISO standard plunger, 

at least 20 times of stirring are required for similar 

sample homogeneity. 


The authors would like to thank the staffs 

and management of KU Dairy Centre, Bangkok 

10900, Thailand for their kind supports and Swiss 

Development Corporation, Helvetas and Royal 

Government of Bhutan for the financial supports. 


Cody, R.P. and J.K. Smith. 1997. Applied 

Statistics and the SAS Programming 

Language. New Jersey. 403p. 

Eon Trading. 2001. Manufacturer of Ekomilk 

Ultrasonic Milk Analyzer (Manual). 

Bultech. Ltd., Bulgaria. 13p. 

Goodridge, L., A. R. Hill and R.W. Lencki. 2004. 

A Review of International Standards and the 

Scientific Literature on Farm Milk Bulk-Tank 

Sampling Protocols. J. Dairy Sci. 87(9): 3099. 

Grace, V., G. A. Houghty, H. Rundick, K. Whaley 

and J. Lindamood. 1992. Sampling dairy and 

related products, pp. 59-82. In R.T. Marchall 

(ed.). Standard Methods for the 

Examination of Dairy Products, 16 ed., 

American Public Health Association, 

Washington, DC.

222 

Harding, F. 1995. Compositional quality, pp. 75- 

96. In F. Harding (ed.). Milk Quality. Blackie 

Academic and Professional, Chapman and 

Hall, London. 

Hasanuzzaman, M., M. A. K. Azad, M. A. Barik 

and M. Z. Rahman. 2002. Milk fat production 

trend and effect of season on it at Sree-Nagor 

milk shed area under milk Vita throughout the 

year. Pakistan J. Nutri. 1 (5): 231-233. 

Hurley, W. L. 2004. Milk fat synthesis. Available 

source:http://classes.aces.uiuc.edu/AnSci308/ 

fatsynthesis. html Retrieved, January 26, 

2005. 

ISO/FDIS. 1997. Milk and milk products - 

Guidance on sampling. International 

Standard (Final Draft). ISO/FDIS, Geneva. 

36p. 

Jenness, R. 1988. Milk Composition, pp. 1-38. In 

N. P. Wong, R. Jenness, M. Keeney and E. H. 


Marth (eds.), Fundamentals of Dairy 

Chemistry 3 rd ed. van Nostrand Reinhold 

Company Inc., New York. 

Likas, B. J. and H. E. Calbert. 1954. A study of 

the influence of agitation time on the Babcock 

test of milk samples from farm bulk holding 

tanks. J. Milk and Food Techno. 17: 14-17. 

Ling, E. R. 1956. A Textbook of Dairy 

Chemistry, 3 rd ed., Chapman and Hall Ltd., 

London. 227p. 

Tetrapak. 1995. Dairy Processing Handbook. 

Tetra Pak Processing Systems AB, S-221 86 

Lund, Sweden. 436p. 

Walstra, P., T. J. Geurts, A. Noomen, A. Jellema 

and M.A.J.S. Van Boekel. 1999. Dairy 

Technology: Principles of Milk Properties 

and Processes, Marcel Dekker, Inc., New 

York. 727p.


Influences of Physicochemical Characteristics of Rice Flour 

and Cassava Starch on the Gelation of Calcium-Induced Egg 

Albumen-Flour Composite 

Parichat Hongsprabhas* and Kamolwan Israkarn 

ABSTRACT 

Mechanical properties of calcium-induced composite gel made of egg albumen (EA) containing 

15% protein (w/v) and 30% of rice flour (RF), cassava starch (CS) or a mixture of RF and CS at the ratio 

of 1:1, added with 0-70 mM calcium lactate and 1% of i-carrageenan were characterized. The twophase 

exponential decay model suggests that the calcium lactate concentration did not have significant 

effect on the stress decay mechanism of the flour-protein composite gel (p≥0.05). However, the type of 

flour filler regulated the stress decay behavior (p

224 

proteins described by Barbut and Foegeding 

(1993) and Doi (1993). The use of two-stage 

gelation of protein matrix introduced by 

Wongsasulak et al. (2004) could avoid phase 

inversion of the continuous protein phase to 

gelatinized starch phase occurred when the 

starch to protein ratio was higher than 3 : 10 

reported by the other investigators (Aguilera and 

Rojas, 1996; 1997; Aguilera and Baffico, 1997). 

The salt-induced gelation and film 

formation of the egg albumen (EA) and cassava 

starch (CS) has been previously demonstrated as 

a potential carrier and coating for the oil-soluble 

compound such as paprika oleoresin by 

Wongsasulak et al. (2006). The release 

characteristics of the core material from the 

starch-protein composite matrix are mainly 

dependent on the microstructure of the composite, 

which were influenced by the composition, 

types and concentration of salt, temperature 

and water availability during its applications. 

Apart from the salt types and their 

concentrations, the source of starch also played 

an important role in determining the mechanical 

properties of the protein-starch composite gel 

(Hongsprabhas and Dit-udom-po, 2006). The EArice 

flour (RF) composite gels were reported to be 

more rigid but had lower extensibility than the EA- 

CS composite gel. It was shown that the 

mechanical properties of the EA-RF, unlike those 

of the EA-CS, were dependent on the mechanical 

properties of the continuous protein phase although 

the starch to protein ratio was as high as 2:1. We 

hypothesized that apart from the volume fraction 

of the filler, the structural reinforcement of the 

composite was also relied on the mechanical 

strength of the filler. 

This study was aimed to further 

investigated the roles of starch granular 

characteristics on the stress transmitting 

mechanisms of the heated calcium-induced EAflour 

composite gels. A better understanding on 

the influences of the separated starch phase on 


the mechanical properties of the composite 

gel could help designing the mechanical properties 

of the composite structure containing high level 

of starch. 


Dried egg albumen (EA) powder (High 

gel type, SA Igreca, France) was obtained from 

Winner Group Enterprise Ltd., Thailand. The 

powder had a protein content of 86.25% (dried 

weight basis) determined by Kjeldahl analytical 

method using the N factor of 6.25 (AOAC, 2000). 

Food grade rice flour (RF, 10.28% moisture 

content, 7.59% protein, Jade Leaf Brand, Bangkok 

Interfood Co. Ltd., Bangkok, Thailand) and 

cassava starch (CS, 12.17% moisture content 

and no protein detected by Kjeldahl method, 

Jade Leaf Brand, Bangkok Interfood Co. Ltd., 

Bangkok, Thailand) were used in this study. 

The ι-carrageenan (Marcel Carrageenan, Quezon 

City, Philippines) was kindly supplied by Behn 

Meyer and Co. (T) Ltd., Bangkok, Thailand. 

Reagent grade of calcium lactate (Fluka, 

Fluka Chemie AG, Switzerland) was used in entire 

study. 

Thermal properties of food grade rice flour and 

cassava starch 

A Pyris 1 DSC (Perkin Elmer, USA) was 

used in this study to characterize thermal properties 

of food grade RF and CS. Flour suspensions of 15 

or 30% (w/w) were prepared in distilled water and 

pipetted into the stainless steel pans, which were 

later hermetically sealed. The samples were heated 

at a rate of 12.5 °C/min from 30 to 95 °C to 

determine the transition temperature and enthalpy 

of gelatinization. The heating rate used was similar 

to the rate used in the Rapid Viscoamylography 

previously reported (Hongsprabhas and Dit-udomp, 

2006).

Effect of flour filler on stress relaxation 

characteristics of the protein-starch composite 

gel 

The EA suspension (pH 7.6) containing 

15 % (w/v) of protein was prepared in distilled 

water, pre-heated at 55 °C for 5 min, cooled to 

room temperature (30 °C) and added with RF, 

RF+CS (1:1) or CS, i-carrageenan and calcium 

lactate solution to obtain the composite suspension 

containing final protein concentration of 15% (w/ 

v), flour filler 30% (w/v), i-carrageenan 1% (w/v) 

in the presence of calcium lactate 0, 10, 30, 50 or 

70 mM using method previously described 

(Hongsprabhas and Dit-udom-po, 2006). The 

suspensions were poured into polycarbonate tubes 

with the inner diameter of 20 mm. The tubes were 

placed in a water bath where the temperature was 

raised to 80 °C and held at that temperature for 30 

min to form gel. All gels were cooled to room 

temperature (30 °C) for 2 h prior to analyses. 

The cylindrical gel sections (20 mm 

diameter × 10 mm long) were compressed for 20% 

deformation between a lubricated stationary 

bottom plate and a moving upper plate with a 

crosshead speed of 100 mm/min using a Lloyde 

Texture Analyzer (Series 500, Fareham, UK) for 

300 s. The compressive stress at time t (σ t) were 

calculated as follow: 

F L L 

σt = t ( −∆ ) 

πr L 2 

where Ft is the compressive force at time t, L is 

the original sample length, ∆L is the corresponding 

deformation, and r is the original radius (Tang et 

al., 1994). The stress relaxation characteristics 

were determined as % relaxation which was 

calculated from the difference between initial 

stress (σ0) and residual stress at 300 s (σ300) compared with the σ0 as follow: 

σ0−σ300 %relaxation = * 100 

σ0 

The stress relaxation data of the EA-flour 

composite matrix were fitted to the two-phase 


exponential decay model by a nonlinear 

regression, using the Grafit software package 

(Leatherbarrows, 1992) as follows : 

( −K1* 

t) ( −K2* 

t) 

σt = a* exp + b*exp + c 

where σ t is the stress at time t. The a+b+c value 

referred as σ 0 that decayed over time t to plateau 

(c) or asymptotic residual stress with the decay 

rates K 1 and K 2. 

Effect of heating on granular characteristics of 

the composites 

The granular characteristics were 

determined as the Water Absorption Index (WAI) 

and microstructure. One mL of starch/flour 

suspension (0.67% w/v) in distilled water was 

shaken vigorously and allowed to stand for 10 min 

to absorb water at room temperature. The 

suspension was heated in a water-bath at 50, 60, 

70, 80 and 90 °C for 30 min in a quiescent 

condition. The sample was cooled down to room 

temperature for 1 min and centrifuged at 14,000 

rpm for 5 min (Spectrafuge 16M, USA). The 

supernatant liquid was discarded and the WAI was 

determined as the weight ratio of the water in the 

sediment and the dried weight of sample. 

The microstructure of starch/flour 

granules in the presence of excess water before 

and after heating at 80 °C for 30 min and cooling 

was examined under the Lieca DME Light 

Microscope (USA). Ten fields of unheated starch/ 

flour suspensions were observed for granule 

appearance and number. The sediment obtained 

after centrifugation was suspended in 0.5 mL 

distilled water and stained with 10% Lugol’s iodine 

solution (Autio and Salmenkallio-Marttila, 2001). 

The microstructure of each starch/flour matrix in 

the presence of EA at the ratio of flour to protein 

as 2:1 was also observed. The EA suspension 

(0.33% protein w/v) was prepared by heating the 

EA suspension at 80 °C for 30 min and cooled 

down to room temperature as described above. The 

pre-heated EA suspension was then added to

226 

starch/flour, shaken vigorously and the suspension 

was re-heated at 80 °C for 30 min prior to the 

examination under the light microscope. 


Thermal properties of RF and CS at 

different concentration were summarized in Table 

1. Gelatinization temperature of the RF was higher 

than that of the CS (p

Stress decay rate : K1 (s-1 ) 

Factor a (kPa) 

14 

13 

12 

11 

10 

9 

8 

7 

6 

5 

4 

3 

2 

1 

0 

0.020 

0.015 

0.010 

0.005 

0.000 

EA+RF 

EA+RF+CS 

EA+CS 

0 10 30 50 70 

Ca lactate (mM) 

EA+RF 

EA+RF+CS 

EA+CS 

Factor c (kPa) 

45 

40 

35 

30 

25 

20 

15 

10 

0 10 30 50 70 


5 

0 


EA+RF 

EA+RF+CS 

EA+CS 

Factor b (kPa) 

55 

50 

45 

40 

35 

30 

25 

20 

15 

10 

5 

0 

EA+RF 

EA+RF+CS 

EA+CS 

0 10 30 50 70 


Stress decay rate : K2 (s-1 ) 

0.4 

0.3 

0.2 

0.1 

0.0 

0 10 30 50 70 


EA+RF 

EA+RF+CS 

EA+CS 

0 10 30 50 70 


Figure 2 Two-phase exponential decay parameters describing stress relaxation characteristics of 

calcium-induced EA-flour composite gel after re-heating at 80°C for 30 min. Bars represent 

standard deviation.

228 

% Stress relaxation 

54 

49 

44 

39 

EA+RF 

EA+RF+CS 

EA+CS 

34 

0 10 20 30 40 50 60 70 80 


Figure 3 Effect of calcium lactate concentration 

on the stress relaxed after the EA-flour 

composite gel compressed for 300 s at 

20% deformation. Bars represent 

standard deviation. 

structure in the starch granules (Nussinovitch et 

al., 1990). It was note that the gelatinization 

temperature of the RF was higher than that of the 

CS as shown in Table 1. In addition, the smaller 

granules of RF also resulted in the higher number 

of the RF granular particles per unit volume of 

the composite. The denser arrangement of the RF 

within the composite network may also be 

responsible for more solid-like behavior of the EA- 

RF, compared with the EA-CS. 

Heating starch suspensions resulted in 

the changes in the microstructural geometry of the 

starch granules. The loss of original granular 

structure occurs after the water molecules 

penetrate into the granules. Thus, the resulting 

granules were mechanically weaker and more 

porous than the raw ones (Rao and Tattiyakul, 

1999). We further explored the granular 

characteristics of RF and CS in distilled water after 

heating for 30 min at different temperature. Figure 

4a illustrates that the heated RF absorbed water 

less than the CS when the temperature was raised 


from 50 to 90 °C in quiescent condition and the 

mixture of RF and CS absorbed more water in 

linear fashion when the ratio of CS increased. 

However, the WAI profile of each composite as a 

function of temperature appeared to be different 

as shown in Figure 4b. 

The WAI data were fitted by a nonlinear 

regression using the Grafit software package 

(Leatherbarrows, 1992) to characterize the 

influence of heating temperature on the WAI of 

RF and CS mathematically. The exponential 

equation (Eq. 1) could be used to describe the 

influence of temperature on the WAI of RF; while 

the power series model (Eq. 2) was best fit to the 

WAI of the CS (Table 2). The WAI response to 

heating temperature of the RF-CS composite 

seems to be a mix phenomenon of the two models. 

y = a*exp (b*x) Eq.1 

y = a*x b + c*x d Eq.2 

This study further showed that in addition 

to the pasting profile of RF and CS as previously 

reported (Hongsprabhas and Dit-udom-po, 2006), 

the mechanisms of starch/flour filler in reinforcing 

the mechanical strength of the EA-flour composite 

matrix also depended on volume fraction and 

strength of the hydrated granules. Figure 5 

illustrates the microstructure of RF and CS after 

being heated at 80 °C for 30 min in the absence 

and presence of pre-heated EA and in the excess 

of water. It appears that the RF could, in part, 

reinforce the EA-RF composite gel via the 

aggregation of deformed (swollen, collapsed and 

disintegrated) amylose-rich fraction. It is likely 

that the presence of pre-heated EA could induce 

the aggregation of the RF through protein-protein 

interactions at the flour-protein interfaces. It should 

be noted that the commercial RF contained 

substantial amount of protein while the protein in 

commercial CS was not detected. Nevertheless, 

the nature of the RF and EA proteins interfacial 

interactions in water-in-water dispersions needs 

further investigation.

CONCLUSION 

The structural reinforcement of starch/ 

flour in the protein-starch composite matrix could 

be evident when the dispersed phase, or fillers, 

remained separated phase in the supporting protein 

matrix. The EA-CS composite gel was weaker and 

more compliant to external force than the EA-RF 

gel due to the differences in granular 

characteristics; e.g., large granular size, low 

number of particles per unit volume and the 

weakness of the hydrated granular structure 

Water absorption index 

30 

25 

20 

15 

10 

5 

(a) (b) 

50 °C 

60 °C 

70 °C 

80 °C 

90 °C 

0 

0.0 0.2 0.4 0.6 0.8 1.0 

Mass fraction of CS in RF-CS flour composite 


after heating of the CS. The aggregation of the 

hydrated RF particles could further enhance the 

strength of protein-starch composite geometry 

during second heating step and the mechanical 

properties and applications of the composite 

structure accordingly. 


The authors would like to thank the 

Faculty of Agro-Industry, Kasetsart University for 

financial supports. 

0 

300 310 320 330 340 350 360 370 380 390 

Table 2 Non-linear regression parameters of the effect of heating temperature (x) on the Water 

Absorption Index (y) of rice flour (RF) and cassava starch (CS). 

Flour type Factor 

a b c d R 2 

RF 8.71 × 10 -7 0.045 - - 0.9622 

CS -6.71 × 10 12 -4.26 4.17 × 10 7 -2.181 0.9795 

Water Absorption Index 

30 

25 

20 

15 

10 

5 

RF : CS mass ratio 

1.0 : 0.0 

0.75 : 0.25 

0.50 : 0.50 

0.75 : 0.25 

0.0 : 1.0 

Temperature (K) 

Figure 4 Effect of cassava starch (CS) ratio in RF-CS flour composite on the water absorption index 

(WAI). Bars represent standard deviation.

230 


Figure 5 Light micrographs of food grade (a) rice flour and (b) cassava starch in distilled water before 

heating; (c) rice flour and (d) cassava starch after heating at 80 °C for 30 min in distilled 

water; and (e) rice flour and (f) cassava starch after heating at 80 °C for 30 min in pre-heated 

egg albumen suspension. Bar = 50 µm.


Aguilera, J.M. and P. Baffico. 1997. Structuremechanical 

properties of heat-induced protein/ 

cassava starch gels. J. Food Sci. 62(5) : 1048. 

Aguilera, J.M. and E. Rojas. 1996. Rheological, 

thermal and microstructural properties of 

whey protein-cassava starch gels. J. Food Sci. 

61(5): 962. 

Aguilera, J.M. and G.V. Rojas. 1997. 

Determination of kinetics of gelation of whey 

protein and cassava starch by oscillatory 

rheometry. Food Res. Intern. 30(5): 349. 

Association of Official Analytical Chemistry. 

2000. Official Methods of Analysis. 17 th ed. 

Arlington VA. 

Autio, K. and M. Salmenkallio-Marttila. 2001. 

Light microscopic investigations of cereal 

grains, doughs and breads. Lebensm.- 

Wiss.u.-Technol. 34: 18. 

Barbut, S. and E.A. Foegeding. 1993. Ca 2+ - 

induced gelation of pre-heated whey protein 

isolate gels. J. Food Sci. 58: 867. 

Doi, E. 1993. Gels and gelling of globular proteins. 

Trends Food Sci. Technol. 4: 1. 

Hongsprabhas, P. and S. Dit-udom-po. 2006. 

Influence of flour type on mechanical 

properties of calcium-induced egg albumenflour 

composite gel. www.bepress.com/ijfe/ 

vol2/iss1/art2. 

Leatherborrow, R.J. 1992. Grafit version 3.0. 

Erithacus Software Ltd., Staines. 

McClements, D.J., F.J. Monahan and J. E. 

Kinsella. 1993. Effect of emulsion droplets on 


the rheology of whey protein isolate gels. 

J. Texture Stud. 24: 411. 

Nussinovitch, A., M.M. Ak, M.D. Normand and 

M.M. Peleg. 1990. Characterization of gellan 

gels by uniaxial compression, stress relaxation 

and creep. J. Texture Stud. 21: 37. 

Rao, M.A. and J. Tattiyakul. 1999. Granule size 

and rheological behavior of heated tapioca 

starch dispersions. Carbohydrate Polymers. 

38: 123. 

Richardson, R.K., G. Robinson, S.B. Ross- 

Murphy and S. Todd. 1981. Mechanical 

spectroscopy of filled gelatin gels. Polymer 

Bull. 4: 541. 

Ross-Murphy, S.B. and S. Todd. 1983. Ultimate 

tensile measurements of filled gelatin gels. 

Polymer 24: 481. 

Wongsasulak, S., T. Yoovidhya, S. Bhumiratana 

and P. Hongsprabhas. 2004. Effect of aging 

temperature on diffusion and microstructure 

of egg albumen-cassava starch composite 

film. Presented at the IFT 2004 Annual 

Meeting and Food Expo, Las Vegas. 

Wongsasulak, S., T. Yoovidhya, S. Bhumiratana, 

P. Hongsprabhas, D.J. McClements and J. 

Weiss. 2006. Thermo-mechanical properties 

of egg albumen - cassava starch composite 

films containing sunflower-oil droplets as a 

function of moisture content. Food Research 

Intern. 39(3): 277. 

Tang, J., J. Lelievre, M.A. Tung and Y. Zeng. 1994. 

Polymer and ion concentration effects on 

gellan gel strength and strain. J. Food Sci. 

59: 216.


The Product Design of Puffed Snacks by Using Quality Function 

Deployment (QFD) and Reverse Engineering (RE) Techniques 

Wiwat Wangcharoen 1 *, Tipvanna Ngarmsak 2 and Brian H. Wilkinson 3 

ABSTRACT 

Quality function deployment (QFD) and reverse engineering (RE) techniques were used to 

design a new product that consumer wants. To prepare the information, the consumer test and quantitative 

descriptive analysis (QDA) were conducted to obtain liking scores and QDA attributes of commercial 

products. QFD was applied to relate consumer liking and QDA attributes of products, and to identify 

the priority and the direction of goodness for each QDA attribute. For a consumer group, it was found 

that flavours, both potato and seasoning flavours, were the most important attributes, followed by saltiness 

and sweetness. This target consumers wanted the product that was strong in seasoning aroma, but optimum 

in seasoning taste and saltiness, and slightly in sweetness. RE was then utilized to create the profile of 

the target product. The RE results showed that the target product should be stronger in seasoning flavours 

and saltiness, and slightly in sweetness. To compromise the results from both QFD and RE, the flavouring 

compound should be reformulated to include less salt to make puffed snacks strong in flavours but 

optimum in taste. 

Key words: food product design, puffed snacks, quality function deployment, reverse engineering 

INTRODUCTION 

The food industry relies heavily on new 

products to rejuvenate and maintain its business. 

The effective and successful new product 

development has to start with the customer (Saguy 

and Moskowitz, 1999). The quality function 

deployment (QFD) is a structured approach for 

integrating research of consumer needs and 

descriptions of the competitive environment with 

technical realities into a unique product 

specification. QFD methodology evolves around 

the “house of quality (HOQ)” a graphical 

representation of the interrelationships 

between customer wants and associated 

product characteristics (Rudolph,1995). The 

implementation of QFD in the food industry was 

started with a three day workshop arranged by the 

American Supplier Institute (ASI) in 1987 

(Charteris, 1993). However, there are few 

published applications of QFD for the improvement 

of food products. Most of the relevant information 

has only been published in the form of scientific 

working papers, or as reports, and so much of this 

1 Department of Food Technology, Faculty of Engineering and Agro-Industry, Maejo University, Chiangmai 50290, Thailand. 

2 Department of Food Technology, Faculty of Technology, Khon Kaen University, Khon Kaen 40002, Thailand. 

3 Institute of Food, Nutrition, and Human Health, Massey University, Palmerston North, New Zealand. 

* Corresponding auther, e-mail:wwwangcharoen@yahoo.com 


information is therefore not so readily available 

to people in general (Costa et al., 2001). 

Benner et al. (2003) criticized that the 

QFD approach ensures that the product is 

developed according to the wants of the target 

consumer group, but it is suitable for product 

improvements and not for truly innovative product. 

Moreover, it takes a large effort and a lot of time 

to conduct QFD for the first time, but once it has 

been executed it will speed up the time-to market 

and enable the company to improve the product at 

less cost. These authors suggested that if QFD was 

going to be used for food product development, it 

was important that simplifications were made to 

the product and its characteristics and interactions 

in order to keep HOQ matrices manageable. 

Another suggestion is the target values in matrices 

may be replaced by target intervals because of the 

fact that food ingredients are often still 

physiologically active materials that hence are 

subject to changes. 

An example of HOQ modification for 

food product development was proposed by Bech 

et al. (1994) to consider the relationships between 

sensory attributes, technical attributes, and 

consumer requirements. This new structure has 

been applied during market-based studies to 

improve the quality of smoked eel fillet (Bech et 

al., 1997), frozen peas (Bech et al., 1997), and 

chocolate couverture (Viaene and Januszewska, 

1999). Another sample was done for a Danish 

butter cookies company by Holmen and Kristensen 

(1996). They used the HOQ approach to identify 

the incompatibility between retailers and 

consumers, and transmit this information to 

internal departments and external suppliers for 

solving problems together. This modification has 

been called “the city of quality” 

To estimate the target value of each 

product characteristic, reverse engineering (RE) 

technique may be used. It is a way to identify the 

levels of one set of variables, given the levels of 

another set of variables by the task of inter-relating 


the data sets coming from products that have been 

systematically varied, or from products that are 

unrelated to each other but have been mapped in a 

coordinate system (Moskowitz, 2000). One of the 

benefits of using reverse engineering is to generate 

a specific sensory profile of product for 

maximizing consumer preferences or meeting the 

constraints of product development process. This 

techniques has been applied in some products such 

as ready- to-eat cereal (Moskowitz, 1997), juice 

(Moskowitz, 1998), and pasta sauce (Moskowitz, 

2000). 

In 2004, the value of savoury snack 

market in Thailand was expected to be worth 

10-12 billion baht and this market consisted of 

puffed snacks (40 %), potato chips (30 %), puffed 

rice (9 %), prawn crackers (8-9 %), fish snacks 

(8 %), and others such as nuts and popcorn 

(Tansattakij, 2004). Wangcharoen et al. (2002) 

showed that 47.3 % of Thai respondents in urban 

areas ate snacks on a daily basis and the main 

reasons for consuming snacks were that they were 

delicious, they stopped hunger, and they are fun 

to eat. Good snacks should be good in sensory 

attributes (taste, smell, and texture), inexpensive, 

convenient to consume, nutritious, and low in fat, 

and they should have long shelf-life, as well. This 

research was aimed to apply QFD and RE 

techniques in sensory attribute designs of 7 Thai 

snack categories for each target consumer group. 

But only results of puffed snacks for a target 

consumer group would be presented in this paper. 


1. Collecting data 

1.1 Preference data 

Forty three Thai consumers, who were 

18 – 45 years of age and ate snacks daily, were 

recruited for 1.5 hour sample evaluation which was 

splitted into 7 sections. In each section, they were 

asked to taste 3 brands, the leading brand and two 

main competing brands, from each of 7 snack

234 

categories including puffed snacks, potato chips, 

dried squid, popcorn, fish snacks, nuts, and prawn 

crackers. Before starting the first section, all 

panelists were advised how to evaluate samples. 

The 9-point hedonic scale was used for preference 

ratings (appearance, aroma, taste, texture, and 

overall liking), and there was a 10 minute break 

between each section. 

Data were then analyzed by SPSS 10.0. 

Cluster analysis was used to classify respondents 

by their overall liking scores for each product 

category. Overall liking of each respondent group 

for each product category was estimated from their 

attribute liking scores by multiple regression 

(stepwise) as the equation below. 

Overall liking = A (Appearance liking) + B 

(Aroma liking) + C (Taste 

liking) + D (Texture liking) 

1.2 Quantitative descriptive analysis (QDA) 

data 

Thirty semi-trained panelists were used 

to obtain QDA data for the same 21 snack products. 

They were selected and trained by the modified 

methods of ISO3972:1991 and ISO4120:1983, 

respectively. 

2. Quality function deployment 

A HOQ was built for each snack category 

to translate the attribute liking into QDA attributes. 

The importance scores of each attribute liking were 

considered from its coefficient in the multiple 

regression between the attribute liking scores and 

the overall liking scores. 

The impact between each pair of attribute 

liking and related QDA attribute, and the 

relationship between each pair of QDA attributes 

were considered by Pearson correlation. Cohen 

(1995) suggested that impacts in the middle of 

HOQ should be considered into 3 levels consisting 

of possibly related which was valued as 1, 

moderately related which was valued as 3, and 


strongly related which was valued as 9. In the 

study, therefore, if a correlation coefficient was ≥ 

0.9 or was an inverted U curve, then it was defined 

as being strongly related and scored as 9 for the 

impact of attribute liking and related QDA attribute 

(in the middle of house), and as + + or -- for the 

relationship between a pair of QDA attributes (on 

the roof). 0.6 ≤ correlation coefficient < 0.9 was 

defined as being moderately related and scored as 

3 and as + or –, respectively. In case of 0.4 ≤ 

correlation coefficient < 0.6 it was defined as 

being slightly related and scored as 1. The 

contribution (priority) for each QDA attribute was 

then computed by summing the multiplication of 

impact and importance scores. 

The direction of goodness for each QDA 

attribute was considered from the sign + and – of 

the correlation coefficient; + meant high intensity 

is good and vice versa in case of minus. For the 

inverted U curve, it meant the intensity of that 

sensory attribute should be optimized. 

3. Reverse engineering 

3.1 Creating a set of equations 

The preference scores and QDA data 

were transformed to z-scores. The 21 products 

were then mapped in a coordinate system by 

reducing their QDA data with principal component 

analysis (PCA). The products’ PC scores and their 

square terms were used as independent variables 

in the equations for estimating the degree of 

preference for each of the sensory attribute as well 

as the overall liking, and the equations for 

evaluating the intensity of the QDA attributes by 

multiple regression (stepwise). 

3.2 Finding the target 

The overall liking equation was 

maximized for each product category by using the 

solver command in Excel 2000. These maximized 

PC scores were then used in the QDA attribute 

equations to get the target product profiles.


In this paper, the product design of puffed 

snacks for a target consumer group would be 

presented only. 

1. Collecting data 

1.1 Preference data 

For puffed snacks, the respondents could 

be clustered into 2 different groups. The first group 

preferred puffed snacks 1 rather than puffed snacks 

2 and puffed snacks 3 and the second group was 

vice versa. Only the first group data is shown in 

Table1 and presented in this paper. The overall 

liking equation under Table 1 showed that the 

overall liking scores for this target group were 

highly affected by taste liking scores and 

moderately affected by texture and aroma liking 

scores. This result seemed to be agreed with 

Williams (1999) who mentioned that flavours 

and seasonings were very important for snack 

products. 

1.2 QDA data 

Semi-trained panelists generated 29 

QDA attributes for describing the sensory 

characteristics of 7 snack categories but the 

number of attributes used for each snack were 

different, such as there were 14 QDA attributes 

for puffed snacks. The QDA results showed that 

puffed snacks 1 were a bit milder in flavours and 

saltiness than puffed snacks 2 and both of them 


were significantly stronger in flavours and 

saltiness than puffed snacks 3. In addition, puffed 

snacks 1 were a bit hotter than puffed snacks 2 

and both of them were significantly hotter than 

puffed snacks 3, whilst puffed snacks 2 and puffed 

snacks 3 were significantly sweeter, harder, and 

more cohesive than puffed snacks 1 (Figure 1). 

2. Quality function deployment 

The HOQ for puffed snacks (Figure 2) 

shows that the important quantitative descriptive 

attributes for puffed snacks were flavour both 

potato and seasoning flavours (7.65), saltiness and 

sweetness (6.12), cohesiveness of mass (1.71), 

hotness (0.68), and hardness (0.19), respectively. 

Clearly, the consumers wanted puffed snacks that 

were more flavoured (positive correlation 

coefficient) and had a lower potato flavour 

(negative correlation coefficient) in the case of 

aroma liking. However, the care for flovour was 

needed as the results showed that there was a 

particular optimum flavour, and that if the flavour 

was increased or decreased from this level, then 

consumer acceptance would decrease, i.e., the 

graph produced an inverted U-shaped curve for 

taste liking. The product needed to be less sweet, 

less cohesive, rather hot, and not hard, and the care 

for saltiness was needed as this attribute also 

showed an inverted U-shape when liking plotted 

against intensity. This finding was agreed with 

Booth and Conner (1990) and Moskowitz and 

Bernstein (2000) who mentioned that factors or 

Table 1 Mean preference scores for 3 puffed snacks. 

Attribute Mean ± S.D. 

Puffed snacks 1 Puffed snacks 2 Puffed snacks 3 

Appearance 7.41 ± 1.40 a 6.64 ± 1.65 b 7.09 ± 1.23 ab 

Aroma 7.36 ± 1.05 a 7.05 ± 1.29 a 6.18 ± 1.59 b 

Taste 7.36 ± 1.14 a 6.41 ± 1.65 b 6.45 ± 1.53 b 

Texture 7.77 ± 0.69 a 7.59 ± 1.01 a 7.09 ± 1.38 b 

Overall 7.59 ± 1.05 a 6.68 ± 1.46 b 6.59 ± 1.53 b 

a, b Means with different letter in same raw are significantly different (p≤0.05) 

Regression equation (R2 = 0.994): 

Overall liking = 0.170 Aroma liking + 0.681 Taste liking + 0.193 Texture liking

236 

Cohesiveness of mass 

Crispness 

Sweetness 

Hardness 

Hotness 

Saltiness 

Length 

Flavours 

attributes which showed an inverted U-shaped 

curve against liking scores would be more 

important for the product acceptance of 

consumers. 

The correlation on the roof of the matrix 

of the HOQ showed that the colour of the products 

(yellow and orange) were influenced by the 

quantity of flavourings and the types of flavour 

added, and depending on the combinations could 

have an influence on the perceived hotness, 

saltiness and sweetness of the products. The 

analysis also indicated that hardness and 

cohesiveness were dependent on the thickness and 

surface smoothness of the products and that 

hardness and cohesiveness increased with the 

thickness or surface smoothness of the products. 

3. Reverse engineering 

3.1 Creating a set of equation 

The PCA reduced 29 QDA attributes to 

8 PCs with 90.35 % explained variance. The 

multiple regression for estimating the degree of 

preference and the intensity of QDA attributes 

were created. For example the overall liking score 

equation was: 


Thickness 

Potato flavour 

Smoothness of surface 

Yellow-body 

Figure 1 Quantitative descriptive analysis of 3 puffed snacks. 

Orange-flavours 

Quantity of flavours 

Puffed snacks 1 



Overall liking = 0.131 PC1 + 0.226 PC2 

– 0.202 PC3 + 0.123 PC4 – 0.104 PC5 – 

0.211 PC6 – 0.373 PC8 – 0.060 PC7 2 + 

0.063 PC8 2 [Adjusted R 2 = 0.808] 

3.2 Finding the target 

The overall liking equation was 

maximized for each snack category. The QDA 

profile of the target product was computed by the 

PC score set obtained from maximizing the overall 

liking equation. For the target puffed snacks, it 

should be an intenser yellow than the existing 

commercial puffed snacks and that the dusted 

flavouring should be pale orange in colour. The 

length should be nearly equal to those of the third 

most important brand of Puffed snacks defined as 

Puffed snacks 3. The thickness should be in the 

middle between Puffed snacks 1 and Puffed snacks 

2. The smoothness of the surface and the quantity 

of flavours should be closer to Puffed snack 1. 

The flavours should be stronger whilst the potato 

flavour should be moderate. The taste should be 

more salty but less sweet and hot as Puffed snack 

1. The texture should be moderately hard and 

cohesive (Figure 3). 

This target product profile that was

produced by the reverse engineering approach for 

puffed snacks produced an almost identical 

solution to the HOQ approach for the important 

product requirements (Figure 2), except the 

seasoning flavours and saltiness where the 

direction of goodness from HOQ suggested that 

they should be optimized, whilst the reverse 


Figure 2 The HOQ for puffed snacks. 

(Plus and minus values in the right and below sections of HOQ were Z-scores of attribute 

likings and QDA data of 3 puffed snacks to show differences of products and would be used 

for reverse engineering in the next step.) 

engineering approach proposed that these 

attributes should be increased. This incompatibility 

could be explained by the strong relationship 

between yellow colour and flavours, and between 

flavours and saltiness, of the existing products on 

the roof of HOQ showing that the flavouring 

agents currently used might be a yellow flavouring

238 

Cohesiveness of mass 

Sweetness 

Hardness 

Crispness 

Hotness 

Saltiness 

agent with added salt. The more flavouring 

the agents, the yellower and saltier the product. 

To compromise the results from both QFD and 

RE, the flavouring compound should be 

reformulated to include less salt to make puffed 

snacks strong in flavours but optimum in taste. 

Since QFD and RE had never been used together 

before, this research showed that they might 

disclose more useful information in product 

designs. 

Strong flavoured and hot puffed snacks, 

a result of the QFD and RE application for product 

designs in this study, might be a product which 

met consumer wants in a survey conducted by 

Wangcharoen et al. (2002) who showed that Thai 

consumers wanted functional puffed snacks with 

hot pungent flavoured Thai herbs. Uhl (2000) also 

mentioned that Thai people preferred foods with 

flavour blends which were hot, pungent, sweet, 

aromatic, and might be sour and salty depending 

on types of foods. 

CONCLUSION 

Length 

Flavours 

This work showed that the application 

of QFD and RE techniques could provide the 


Thickness 

Potato flavour 

Smoothness of surface 

Yellow-body 

Orange-flavours 

Quantity of flavours 

Figure 3 Quantitative descriptive analysis of 3 puffed snacks and a target product. 




Target product 

useful information to product developers. QFD 

helped the product developers to identify the 

importance of product attributes and the direction 

to improve them, whilst RE was beneficial in the 

intensity estimation of the target product’s 

attributes. The compromise of QFD and RE results 

might give the better result to product developers. 


The authors wish to thank National 

Center for Genetic Engineering and Biotechology 

(BIOTEC), Thailand, for supporting funds to this 

research. 


Bech, A.C., E. Engelund, H.J. Juhl, K. Kristensen 

and C.S. Poulsen. 1994. Qfood-Optimal 

Design of Food Products. MAPP working 

paper no. 19, MAPP. Aarhus. 18 p. 

Bech, A.C., K. Kristensen, H.J. Juhl and C.S. 

Poulsen. 1997. Development of farmed 

smoked eel in accordance with consumer 

demands, pp. 3-19. In J.B. Luten, T. Bresen 

and J. Oehlenschlager (eds.). Seafood from

Producer to Consumer, Integrated to 

Quality. Elsevier Science. Amsterdam. 

Bech, A.C., M. Hansen and L. Wienberg. 1997. 

Application of house of quality in translation 

of consumer needs into sensory attributes 

measurable by descriptive sensory analysis. 

Food Qual. Prefer. 8(5/6): 329-348. 

Benner, M., A.R. Linnemann, W.M.F. Jongen and 

P. Folstar. 2003. Quality function deployment 

(QFD)-can it be used to develop food 

products? Food Qual. Prefer. 14: 327-339. 

Booth, D.A. and M.T. Conner. 1990. Characterisation 

and measurement of influences on 

food acceptability by analysis of choice 

differences: theory and practice. Food Qual. 

Prefer. 2: 75-85. 

Charteris, W. 1993. Quality function deployment 

: a quality engineering technology for the 

food industry. J. Soc. Dairy Technol. 46(1): 

12-21. 

Cohen, L. 1995. Quality Function Deployment: 

How to Make QFD Work for You. Addison- 

Wesley Publishing Company. Massachusetts. 

348 p. 

Costa, A.I.A., M. Dekker and W.M.F. Jongen. 

2001. Quality function deployment in the food 

industry: a review. Trends Food Sci. Tech. 

11(9-10): 306-314. 

Holmen, E. and P.S. Kristensen. 1996. Downstream 

and Upstream Extension of the House of 

Quality. MAPP working paper no. 37. The 

Aarhus School of Business. Aarhus. 22 p. 

ISO. 1991. Sensory Analysis-Methodology- 

Method of Investigating Sensitivity of Taste. 

ISO3972. International Organization for 

Standardization. Geneva. 7 p. 

ISO. 1983. Sensory Analysis-Methodology- 

Triangular Test. ISO4120. International 

Organization for Standardization. Geneva. 

8 p. 

Moskowitz, H. 2000. Inter-relating data sets for 

product development: the reverse engineering 

approach. Food Qual. Prefer. (1/2): 105-119. 


Moskowitz, H. and B. Krieger. 1998. International 

product optimization: a case history. Food 

Qual. Prefer. 6: 443-454. 

Moskowitz, H.R. 1997. A commercial application 

of RSM for ready to eat cereal. Food Qual. 

Prefer. 3: 191-201. 

Moskowitz, H.R. and R. Bernstein. 2000. 

Variability in hedonics: indications of worldwide 

sensory and cognitive preference 

segmentation. J. Sensory Studies 15: 263- 

284. 

Rudolph, M.J. 1995. The food product development 

process. Brit. Food J. 97(3): 3-11. 

Saguy, I.S. and H.R. Moskowitz. 1999. Integrating 

the consumer into new product development. 

Food Technol. - CHICAGO 53 (8): 68-73. 

Thansattakij. 2004. Tasto launched “fried crab with 

curry powder” flavoured potato chips. 

Thansattakij 1907, 27-29 May 2004. 

Uhl, S.R. 2000. Handbook of Spices, 

Seasonings, & Flavorings. Technomic 

Publishing Company. Lancaster. 329 p. 

Viaene, J. and R. Januszewska. 1999. Quality 

function deployment in the chocolate industry. 

Food Qual. Prefer. (4/5): 377-385. 

Wangcharoen, W., T. Ngarmsak and B.H. 

Wilkinson. 2002. Consumption of Snack 

Products: Potential for Functional Snacks 

Containing Herbs in Thailand and New 

Zealand. Poster Presented at The 9th World 

Congress on Clinical Nutrition. London. 

United Kingdom. June 24-26, 2002: AB 220. 

Wangcharoen, W., T. Ngarmsak and B.H. 

Wilkinson. 2002. Suitability of using herbs as 

functional ingredients in Thai commercial 

snacks. Kasetsart J. (Nat.Sci.) 36: 426-434. 

Williams D. 1999. Flavors for snack-food 

application. Perfumers & Flavorist. 24: 29, 

31-32, 34.


Effect of Coating on Doughnut Cake Preference using R-index 

Tunyaporn Sirilert, Anuvat Jangchud * , Phaisan Wuttijumnong 

and Kamolwan Jangchud 

ABSTRACT 

Signal detection has been used to explain the underlying basis of difference testing from control 

and produces accurate different values on consumer acceptance. The number of samples presented in 

signal detection rating test was much lower than that of the instrumental method of limit test, which 

would require less time and cost for preparation by the experimenter. Signal detection on the preference 

of coated doughnut cake with different coatings from blend polymer between chitosan (CH) and 

methylcellulose (MC) at different ratios of CH: MC (1.00:0.00, 0.00:1.00, 0.75:1.25, 1.00:1.00 and 

1.25:0.75) were conducted by R-index method. Instrumental measurement on coated doughnut cake 

were also determined. Results showed that the instrumental method was highly correlated with R-index 

method. Coating solution of CH:MC at 1.00:0.00 and 1.25:0.75 showed the difference of instrumental 

measurement on brownish color, odor, hardness and oil content higher than that of the control. The 

optimum coatings at CH:MC of 0.00:1.00 (MC) and 0.75:1.25 (CM1) significantly reduced oil uptake 

without any effect on overall liking. 

Key words: R-index, consumer, preference, biopolymer coating, doughnut cake 

INTRODUCTION 

Deep fat frying is widely used in 

industrial or institutional preparation of food 

(Pinthus et al.,1992). Some fried food products, 

fat and oil may improve overall food palatability, 

mouthfeel, or flavor, but high fat diet are 

discouraged because of potential diet related 

diseases. Thus, oil uptake in fried foods has 

become a health concern. High consumption of 

lipids has been related to obesity and other health 

problems like coronary heart disease. Edible 

coating at the surface of the food comprises 

another possibility, but this technique has not been 

studied extensively. Williams and Mittal (1999) 

reported that methylcellulose (MC) film showed 

the best barrier property, because it reduced fat 

uptake more than hydroxypropylmethylcellulose 

(HPMC) for a pastry mix. It was reported that the 

cellulose derivatives were used to reduce oil 

absorption in fried products (Meyers, 1990). 

However, food coatings may become a good 

alternative to solve this problem. The effectiveness 

of a coating must be determined by its mechanical 

and barrier properties. 

Signal detection theory has been used to 

explain the underlying basis of difference testing, 

but implementation of signal detection procedures 

is rather time consuming. Therefore, short-cut 

signal detection procedures were developed 

allowing R-indice to be calculated from either 

rating or ranking data on a variety of food products 

Department of Product and Development, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author. e-mail addresses: fagiavj@ku.ac.th. 


(O’Mahony et al., 1985). 

Hedonic or ‘liking’ measurements have 

generally used category scales (5, 7, or 9 points) 

and uni- or bipolar magnitude estimation scales 

to give numerical estimations of liking (Pearce 

et al., 1986). More recently, signal detection 

procedures have been modified for hedonic 

evaluation, allowing the calculation of an R-index 

that gives the probability of preferring a test 

product over the control rather than the degree of 

difference from the control. Vie et al.(1991) used 

rating and ranking data collection procedures to 

calculate R-indice for potato chip preference and 

to determine the likelihood-to-buy. Swanson and 

Lewis (1992) used ranked data to calculate an 

R-index indicating honey preferences and a 

willingness-to-buy. 

Cliff et al.(2000) modified this index to 

develop a bipolar R-index. This allows for the 

‘signal’ samples to have both lower and higher 

intensities than the control sample. A further 

modification by the same authors provided a 

‘weighted-bipolar’ R-index. This addressed the 

bias associated with overestimation of the sample 

size. However, R-index is a probability rather than 

a relative preference score. It provides to the food 

researcher with an alternate interpretation for 

market research, which will ultimately save time 

and cost during development of food products. 

This research was intended to examine 

in term of probability on difference of coated 

doughnut cake from uncoated doughnut cake 

(control) and usefulness of the R-index to produce 

accurate different values with less preparation and 

correlation of rating data collection procedures 

with instrumental measurements for consumer 

preference, which would ultimately save time and 

costs. 


1. Preparation of coating solutions 

Two types of biopolymer were used in 


this study (chitosan, CH: DE>94.68; Thai Union 

Co. Ltd, Thailand and methylcellulose, 

MC: methoxyl group 27.5-32.0%; Fluka, 

Switzerland). Coating solution blending at 

different ratios of CH: MC was studied (1.00:0.00, 

0.00:1.00, 0.75:1.25, 1.00:1.00, 1.25:0.75 

named CH, MC, CM1, CM2, and CM3, 

respectively) and the concentration in the 

solutions was prepared at 1% (w/w). Coating 

solution was first dissolved thoroughly in 1% 

(w/w) formic acid solution (99% concentrate, 

BHD Co., England). After added with 

0.1% polyethyleneglycol (No.400 Fluka, 

Switzerland) and stirred by agitator mixer 

(Cole Co., Ltd., Illinois) for 20 minutes, the 

solution was then filtered with silk filter at 

120 mesh. Coating solutions were used in 

doughnut cake development. 

2. Doughnut cake preparation 

Six treatments of doughnut cake 

produced by coating with different biopolymer 

solution as ratios of chitosan: methylcellulose 

(1.00:0.00, 0.00:1.00, 0.75:1.25, 1.00:1.00 and 

1.25:0.75 w/v) were used in the tasting. All of 

doughnut cakes were prepared by compressing 

with mold to obtain 50 mm × 10 mm height 

dimension. One was uncoated sample (hereafter 

referred to as the control) and the other doughnut 

cakes were then dipped in coating solutions for 

10 second and allowed to dry at 45°C for 

15 minutes with tray drier (Model BWS, B.W.S. 

trading Ltd, Thailand). Frying of doughnut 

cake was carried out at 165°C for 2.0 minute 

(Pinthus, 1995). 

Prior to sensory evaluation, the six 

doughnut cakes were transferred into polyethylene 

bags. Each of doughnut cake was presented 

as one piece labeled with a three digit random 

order. They were evaluated using one or 

two scorecards, rating or ranking, which were 

randomly assigned to the bags.

242 

3. Physicochemical properties 

3.1 Oil content 

Oil content (OC) of fried products was 

determined (by dried samples) using combined 

technique of successive batch and continuous fat 

extraction (Soxtec Avanti-2050, USA), with 

petroleum ether (AOAC, 1995). 

3.2 Colorimetric measurements 

Experiments were carried out with a 

Minolta spectrophotometer (CM-3500d series, 

Japan) calibrated with a standard. The Hunter scale 

was used, lightness (L*) and chromaticity 

parameters a*(red-green), b*(yellow-blue) and 

overall difference of color (∆E*) was measured. 

Sampling five points of surface sample were 

measured and they were analyzed in triplicate, 

recording three measurements for each sample. 

3.3 Texture analysis 

Samples were measured by compression 

test using a texture analyzer Lloyd TA-500 (USA). 

Samples (specimen) were compressed with a 5 kg 

cell at speed of 0.5 mm/s with a cylinder 

compressive shape (10 cm diameter). Texture 

profile analysis was determined on hardness 

property. At least 10 samples were measured in 

each experiment. Samples were allowed to reach 

room temperature before performing the tests 

(Garcia et al., 2002). 

4. Sensory evaluation 

The consumer evaluation consisted of 40 

panelists, 20 males and 20 females, with age range 

of 15 to 55 years. Panelists were given a 

presentation on the basics of sensory evaluation, 

specifically the R-index method, in order to 

familiarize them with the type of test being used. 

Panelists were advised no bias in preference of 

doughnut cake. After the evaluation, comments 

were collected on their preferences. 

Panelists were splitted equally into two 

gender balanced groups for the different and 

acceptance tests. R-index was used for different 


test and all attributes of color (brownish), odor, 

texture (hardness) and oily sheen were obtained, 

panelists tested the six samples and scored samples 

‘least liked’ (left) to ‘most liked’ (right). For 

acceptability, 9 point hedonic scale was used. 

Panelists tested the products according to a 

randomized design and checked the appropriated 

box according to their degree of liking from 

‘extremely dislike’ to ‘extremely like’ (Vie et al., 

1991). 

5. Statistical analysis 

The R-index measures the degree of 

difference between a ‘control’ sample and a 

‘signal’ or treatment sample, in terms of the 

probability of distinguishing sample paired 

comparison. The experiment selected with a 

chance value more than 0.5 indicated samples to 

be distinguished. Values of the traditional R-index 

are ranged between 0.5 and 1.0 (50-100%), with 

higher values indicating better discrimination 

(Harker et al., 2002). 

The weighted-bipolar with R-indice 

was calculated following the method of Chiff 

et al. (2000). The weighting values were described 

in terms of propability in paired comparison and 

the significant signal and noise of samples with 

randomization were calculated using tables 

provided by O’Mahony et al. (1995). 

RESULTS AND DISSCUSSION 

1. Instrumental measurement 

Oil content, hardness, and color 

measurement of uncoated (control) and coated 

sample are shown in Table 1 

1.1 Oil content 

Blending of chitosan and methylcellulose 

at the given ratios (CH, MC, CM1, CM2, and 

CM3) were studied in doughnut cakes coating 

compared with the uncoated samples (control). 

The results showed that coating solution

significantly reduced oil content in doughnut cake 

(p≤0.05) described by a mechanism and natural 

barrier properties of coating solution, which 

depended on small porous size in microstructure 

of product (Debeaufort and Voilley, 1997). It also 

demonstrated that coating could reduce oil content 

in doughnut cake. Comparing between coating 

treatments, coating with only MC (0.00:1.00) 

showed the most effective on oil reduction, 

because of MC exhibited thermo-gelation, 

consequence as the best barrier properties of oil 

and moisture during frying (Grover, 1993). 

1.2 Color properties 

Coating affected color of fried doughnut 

cake significantly (p0.05) at 

0.00:1.00 as MC formed the best barrier of water 

loss and oil absorption (Garcia et al., 2002), which 

protected the evaporation of water from the food 

and absorption of oil by the food during frying 

(Krokida et al., 2000b). 

Table 1 The values of physicochemical in fried doughnut cake. 

Treatments Oil content Hardness L* a* b* ∆E* 

(Chitosan : (% wb.) (N) 

Methylcellulose) 

Uncoated 19.61± 2.03 a 186.96 a 64.31±0.23 a 13.44±0.28 d 40.08±0.30 b 1.48±0.82 d 

(Control) 

0.00 : 1.00 (MC) 13.31± 1.52 d 113.06 b 61.74±0.88 a 12.45±0.30 e 39.96±0.48 b 2.46±0.67 d 

1.00 : 0.00 (CH) 16.46± 2.55 b 115.14 b 51.18±0.74 c 18.34±0.33 a 38.64±0.58 c 13.05±0.90 a 

0.75 : 1.25 (CM1) 14.96± 2.04 c 114.22 b 58.44±1.42 b 14.64±0.94 c 41.33±0.61 a 5.26±1.75 c 

1.00 : 1.00 (CM2) 15.36± 2.15 c 115.04 b 57.75±0.59 b 16.61±0.25 b 41.59±0.22 a 7.14±0.65 b 

1.25 : 0.75 (CM3) 15.39± 2.53 c 116.29 b 53.79±1.67 c 16.04±0.75 b 40.58±0.51 b 7.97±1.82 b 

abcd Mean in the same column with different letters are significantly different (p≤0.05)

244 


Table 2 R-index measures of characteristics on doughnut cake. 

Treatments Weight-Bipolar Weight-Bipolar Weight-Bipolar Weight-Bipolar 

(Chitosan : (brownish color) (odor) (hardness) (oily sheen) 

Methylcellulose) R-index R-index R-index R-index R-index R-index R-index R-index 

more less more less more less more less 

0.00 : 0.00 (Control) 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 

0.00 : 1.00 (MC) nc 0.64 0.58 nc 0.62 nc nc 0.53 

1.00 : 0.00 (CH) 0.81* nc 0.69* nc 0.70* nc 0.72* nc 

0.75 : 1.25 (CM1) 0.70* nc 0.59 nc 0.63 nc nc 0.39 

1.00 : 1.00 (CM2) 0.74* nc 0.63 nc 0.64 nc 0.63 nc 

1.25 : 0.75 (CM3) 0.75* nc 0.69* nc 0.68* nc nc 0.43 

R-index followed by * are significant at p≤0.05 

nc , R-index not calculated (not significant at p>0.05) 

The relationship between instrumental 

method and R-index of hardness, oil content and 

brownish color were determined (Figure 1-3). 

Results showed that the instrumental method was 

correlated with R-index method, the distribution 

of the instrumental method was slightly shifted to 

the relatively right with the R-index distribution. 

This demonstrated that R-index method could be 

used as a tool to determine the qualities of product 

with less number of sample compared to the 

instrumental method (the number of sample for 

instrumental method was 108 and for signal 

detection rating test was 40). It was confirmed by 

Cliff et al. (2000) who mentioned that 

R-index method could be a measurement of 

magnitude and direction in the partition of the 

same-unsure responses. It required less time and 

cost for preparation by the experimenter. The 

optimum coating were MC (0.00:1.00) and CM1 

(0.75:1.25) whose oil content were lower than the 

control. As a consequence, it could be used to 

reduce oil uptake in doughnut cake. It was related 

with Williams and Mittal (1999) who reported that 

methylcellulose (MC) films showed the best 

barrier properties, to reduced fat uptake in pastry 

mix. 

3. Overall liking 

The sensory scores of doughnut cake are 

shown in Table 3. The high score of overall liking 

of MC (0.00:1.00) and CM1 (0.75:1.25) coatings 

were not significantly different from the control 

(p>0.05). This result agreed with R-index 

measurement on odor, hardness and oily sheen, 

where those samples were not significantly 

different from the control. 

CONCLUSION 

This study demonstrated that R-index 

method could be used to evaluate the degree of 

difference from the control in doughnut cake. The 

results showed that preference of consumer were 

affected by different coating solutions. However, 

optimum coating could be obtained and used to 

reduce oil uptake with non significant effect on 

overall liking. 

Table 3 Overall liking in doughnut cake with 9 

points hedonic scale. 

Treatments Mean ± SD 

Control (Uncoated) 6.02 ± 1.80ab 

0.00 : 1.00 (MC) 7.37 ± 1.74 a 

1.00 : 0.00 (CH) 3.90 ± 1.78 c 

0.75 : 1.25 (CM1) 6.42 ± 1.97 ab 

1.00 : 1.00 (CM2) 4.65 ± 1.99 bc 

1.25 : 0.75 (CM3) 4.45 ± 2.09 c 

abc Mean in the same column with different letters are 

significantly different (p≤0.05)

Hardness (N) 

120 

100 

80 

60 

40 

20 

0 

Con 

(0.00:0.00) 

MC 

(0.00:1.00) 


CH 

(1.00:0.00) 

CM1 

(0.75:1.25) 

CM2 

(1.00:1.00) 

CM3 

(1.25:0.75) 

Figure 1 Distribution of hardness detection in coated doughnut cake 

(■ Instrument method, --▲--R-index method) 

Oil content (%) 

Figure 2 Distribution of oil detection in coated doughnut cake 


Brownish color 

(delE*) 

25 

20 

15 

10 

5 

0 

14 

12 

10 

8 

6 

4 

2 

0 

Con 

(0.00:0.00) 

Con 

(0.00:0.00) 

MC 

(0.00:1.00) 

MC 

(0.00:1.00) 

CH 

(1.00:0.00) 

CH 

(1.00:0.00) 

CM1 

(0.75:1.25) 

CM1 

(0.75:1.25) 

CM2 

(1.00:1.00) 

CM2 

(1.00:1.00) 

Figure 3 Distribution of color detection in coated doughnut cake 


CM3 

(1.25:0.75) 

CM3 

(1.25:0.75) 

1.00 

0.80 

0.60 

0.40 

0.20 

0.00 

1.00 

0.80 

0.60 

0.40 

0.20 

0.00 

1.00 

0.80 

0.60 

0.40 

0.20 

0.00 

R-index score 

R-index scor 

R-index score

246 


This research was financially supported 

by the faculty of graduate school, Kasetsart 

University, Thailand. 


A.O.A.C. 1995. Official Methods of Analysis. 

Association of Official Analytical Chemists, 

Inc., Washington D.C. 1588. 

Cliff, M. A., M. O’Mahony., L. Fukumoto and 

M.C. King. 2000. Development of a ‘Bipolar’ 

R-index. J. of Sensory Studeies 15: 219-229. 

Flokes, M.J. and P.S. Hope.1985. Polymer blend 

and alloys. Chapman and Hall. London. 

NewYork, 430-440. 

Garcia,M.A., C. Ferrero, N. Bertola, M. Martino 

and N. Zaritzky. 2002. Edible coatings from 

cellulose derivatives to reduce oil uptake in 

fried products. Innovative Food Sci and 

Emerging Tech. 3: 391-397. 

Green, D.M. and J.A.Swets. 1996. Signal 

Detection Theory and Psychophysics. John 

Wiley, New York. 151 p. 

Harker, F.R., J. Maindonald, J. Murray, S.H. 

Gunson, I.C. Hallett and S.B. Walker. 2002. 

Sensory interpretation of instrumental 

measurements 1: texture of apple fruit. 

Postharvest Biology and Technology. 24: 

225-239. 

Krokida, M. K., V. Oreopoulou and Z.B. Maroulis. 

2000b.Water loss and oil uptake as a function 

of frying time. J. of Food Engineering 44: 

39-46. 

Meyers, M.A. 1990. Functionality of hydrocolloids 

in batter coating systems. In p 17- 

142. K. Kulp and R. Loewe (eds.). Batters 

and breadings in food processing, American 

Association for cereal Chemists. 

Muzzareli, R.A.A. 1985. Encyclopaedia of 

Polymer Science and Engineering. Wiley, 

New York. 3: 430-440. 


O’Mahony M., S.Y. Wong and N. Obbert. 1985. 

Sensory evaluation of navel oranges treated 

with low dose of gramma radiation. J. Food 

Sci. 50: 639-646. 

O’Mahony, M. 1988. Sensory difference and 

preference testing on the use of signal 

detection measures pp. 145-147. In H. 

Moskowitz (ed.). Applied Sensory Analysis 

of Foods. 

O’Mahony M. 1992. Understanding discrimination 

test: A user- friendly treatment of 

response bias, rating and ranking R-index tests 

and their relationship to signal detection. J. 

of Sensory Studies 7: 1-47. 

Pearce, J.H., B. Korth and C.B. Warren. 1986. 

Evaluation of three scaling methods for 

hedonics. J. of Sensory Studies 1: 27-46. 

Pinthus, E. J. and I. S. Saguy. 1992. Gel strength 

in restructure potato product affects oil uptake 

during deep fat frying J. Food Sci. 57: 1359- 

1360. 

Pinthus E.J., P. Weinberg. and I.S Saguy,. 1994. 

Initial interfacial tension and oil uptake by 

deep-fat fried foods. J. of Food Sci. 59 (804- 

807), 823. 

Pinthus E.J., P. Weinberg and I.S. Saguy. 1995. 

Oil uptake in deep fat frying as affected by 

porosity. J. Food Sci. 60: 4, 767-772. 

Robinson, K.M., B.P. Klein and S.Y. Lee. 2004. 

Utilizing the R-index measure for threshold 

testing in model caffeine solutions. Food 

quality and Preference 16: 283-289. 

Vie, A., D. Guilli and M. O’Mahony. 1991. 

Alternative hedonic measures. J. Food Sci. 

24: 225-239. 

Williams, R. and G.S. Mittal. 1999. Water and fat 

transfer properties of polysaccharide films on 

fried pastry mix. Lebensmittel-Wissenschaft 

und Technology 32: 440-445.


Application of Artificial Neural Networks for 

Reservoir Inflow Forecasting 

Varawoot Vudhivanich*, Santi Thongpumnak, 

Nimit Cherdchanpipat, Areeya Rittima and Nattaphun Kasempun 

ABSTRACT 

This study showed the application of the Artificial Neural Networks in forecasting the reservoir 

inflow. Two cases were studied, (1) single reservoir inflow forecasting and (2) multi-reservoir inflow 

forecasting. The problems were formulated as daily, weekly and monthly inflow forecast. There were 4 

types of model namely A, B, C and D according to the levels of data used as the input variables to the 

ANNs. Model A used all available data of that reservoir. Model B used the data having relatively high 

correlation with the reservoir inflow such as the first 3 lags of reservoir inflow, stream flow, rainfall and 

some meteorological data. Model C used only the first 3 lags of the reservoir inflow and stream flow 

data. Model D used the first 3 lags of reservoir inflow, stream flow and rainfall data. The 4 reservoirs 

namely Mun Bon, Lam Chae, Lam Phra Phloeng and Lam Takong reservoirs in Upper Mun basin, 

Nakhon Ratchasima province, were selected as the case study. Feed forwards back propagation algorithm 

was selected for the study. One to 3 hidden layers with different ANNs parameters were experimented. 

Two to 3 hidden layers were suitable for single reservoir problem while 1 to 2 hidden layers were 

suitable for multi-reservoir problem. Sigmoid transfer function was used in all the models. The initial 

weight, learning rate and momentum were in the ranges of 0.80-0.90. However they were not sensitive 

to prediction performance. For single reservoir forecasting, models A and B showed better performance 

(R 2 ) than models C and D. The monthly model showed the better result than the weekly and daily 

models. For multi-reservoir forecasting, the performance of the 4 models was not different. Model C 

was recommended since it required less data. The training and testing performance of daily, weekly and 

monthly models were not much different in case of multi-reservoir. 

Key words: artificial neural networks, forecasting, reservoir inflow, Upper Mun basin 

INTRODUCTION 

Reservoir inflow forecasting is an 

important task in reservoir operations. An effective 

reservoir inflow forecasting enables the reservoir 

operators to get the accurate information for 

decision making in planning and operating the 

reservoirs. With accurate and reliable forecast of 

inflow, flood and drought damages and inefficient 

utilization of water resources can be reduced. 

However, an accurate and reliable inflow forecast 

is usually difficult to obtain, particularly for a long 

lead time. 

The artificial neural networks (ANNs) 

Department of Irrigation Engineering, Faculty of Engineering at Kamphaengsaen, Kasetsart University, Kamphaengsaen Campus, 

Nakhon Pathom 73140, Thailand. 

* Corresponding auther, e-mail: fengvwv@ku.ac.th 


248 

are now becoming more and more popular in 

hydrological forecasting. ANNs is apart of 

artificial intelligence (AI) which has been widely 

applied in many fields. It is also called a machine 

learning algorithm or natural intelligent system. 

ANNs is a computational system that resembles 

the performance characteristics of the biological 

neural networks of the human brain (Vudhivanich, 

2001). It is in the class of black box model which 

inputs, outputs and functional performance are 

known, whereas the internal process is unknown 

or irrelevant. ANNs are information processing 

system that are composed of a number of neurons 

and interconnections between these neurons. The 

neurons are arranged in group called layer. 

Commonly, the basic ANNs architectures consist 

of three layers namely input layer, hidden layer 

and output layer. The most different characteristics 

of ANNs are their capability to recognize the 

patterns from the example outputs by the automatic 

weight adjustments. The selection of the best fit 

model is accomplished by a trial and error process 

(Tokar, 1996). 

Since 1990s, the successful applications 

for hydrological forecasting by means of ANNs 

techniques have been extensively carried out in 

water resource engineering. Tokar and Markus 

(2000) presented ANNs approach with backpropagation 

algorithm for rainfall-runoff 

modeling. In Thailand, the hourly water levels and 

discharges in Chao Pharaya River were forecasted 

for flood control study in tidal area by ANNs with 

back propagation algorithm (Tingsanchali and 

Manusthiparom, 2001). The ANNs with neurogenetic 

algorithm was developed to forecast water 

level for flood warning system in Hat Yai district 

(Supharatid, 2002). The artificial neural networks 

model was also developed to forecast the daily, 

weekly and monthly inflow of Lam Takong 

reservoir (Vudhivanich and Rittima, 2003) and the 

four reservoirs in the Upper Mun basin 

(Vudhivanich et al., 2004). Additionally, 

ANNs was also applied for runoff forecasting in 


Lam Phachi river basin (Phathravuthichai 

and Vudhivanich, 2003). Most of the mentioned 

researches used feed forwards back propagation 

algorithm with sigmoid transfer function in 

forecasting where the result is satisfactorily. 

In this paper, the ANNs model was 

developed for the reservoir inflow forecasting to 

benefit the reservoir operations. Four reservoirs 

namely Mun Bon, Lam Chae, Lam Phra Phloeng 

and Lam Takong reservoirs in Nakhon Ratchasima 

province were selected as the case study. 


Required data 

(1) The daily, weekly and monthly 

inflow of Mun Bon reservoir during 1995-2000, 

Lam Chae reservoir during 1999-2002, Lam Phra 

Phloeng reservoir during 1992-2000 and Lam 

Takong reservoir during 1987-2000. 

(2) The streamflow data of station 

M.49B near Mun Bon reservoir during 1995-2000, 

M.81 near Lam Chae reservoir during 1999-2002, 

M.145 near Lam Phra Phloeng reservoir during 

1992-2000 and M.89 near Lam Takong reservoir 

during 1992-2000. 


rainfall data as follows; 

- Mun Bon reservoir: station 25293 

(Chok Chai), 25112 (Khon Buri), 25152 (Ban San 

Chao Pho School) during 1995-2000. 

- Lam Chae reservoir: station 25093 

(Chok Chai), 25112 (Khon Buri) and 25152 (Ban 

San Chao Pho School) during 1999-2002. 

- Lam Phra Phloeng reservoir: station 

25511 (Lam Phra Phloeng), 25102 (Pak Thong 

Chai), 25093 (Chokchai) and 25152 (Ban San 

Chao Pho School) during 1987-2000, station 

25751 (Ban Wang Ta-Khian Thong) and 25781 

(Ban Tha Nam Sab) during 1992-2000. 

- Lam Takong reservoir: station 25541 

(Lam Takong), 25062 (Sung Noen), 25013 

(Muang), 25612 (Agriculture Office), 25644

(Lam Takong watershed research station), 25650 

(Chok Chai 4 farms) and 25272 (Pak Chong 

agrometeorological station) during 1992-2000. 


meteorological data of Nakhon Ratchasima station 

including the temperature, pressure and relative 

humidity during 1992-2002. 

Methods 

(1) Preliminary checking the 

abnormality and inconsistency of the data via time 

series plots and filling in the missing data by the 

distance weighted method. 

(2) Determing the autocorrelation and 

cross correlation matrices of the data in order to 

identify the tentative ANNs input structures for 

daily, weekly and monthly inflow forecast for both 

a single reservoir and multi-reservoir inflow 

forecasting. 

(3) Training and testing the ANNs 

model by adjusting the number of hidden layers, 

number of neurons in hidden layer, ANNs 

parameters such that the performance efficiency 

in term of R 2 would be acceptable. Selecting 80% 

of the records for training and using the remainder, 

20% of the records, for testing. 

Figure 1 Location map of Upper Mun basin. 


Description of the study area 

Upper Mun basin is situated in the 

northeast of Thailand covering the total area of 

37,970 km 2 in three provinces namely Nakhon 

Ratchasima, Buri Ram and Surin provinces as 

shown in Figure 1. Upper Mun basin is a subbasin 

of Mun River basin. It covers about 54.5% of the 

Mun River basin area. This basin is composed of 

four main reservoirs; Mun Bon(MB), Lam 

Chae(LC), Lam Phra Phloeng(LP) and Lam 

Takong(LK) reservoirs. All of them are located in 

Nakhon Ratchasima province. 

These reservoirs have been operated by 

Royal Irrigation Department (RID) mainly for 

irrigation and municipality purposes. Most of the 

water in the four reservoirs have been used for 

irrigation. The four reservoirs have the combined 

storage capacity of 836 mcm which can supply 

water to 353,650 rai of irrigable area in Mun Bon, 

Lam Chae, Lam Phra Phloeng and Lam Takong 

irrigation projects. The basic data of the four 

reservoirs are shown in Table 1. 

In addition, Mun Bon reservoir has 

allocated 0.0025 mcm per month of water for 

Charakae Hin sub-district municipality, Khon Buri 

district and 0.40 mcm per month for the

250 

downstream control. Lam Chae reservoir has 

allocated 5.13 mcm per month for the downstream 

control. Likewise, the amount of 1.93 and 2.04 

mcm per month are diverted from Lam Phra 

Phloeng and Lam Takong reservoirs, respectively, 

for municipal water supply of Nakhon Ratchasima 

province. However, the municipal water 

requirements are only small percentages of the 

irrigation water requirements. 


Preliminary analysis of data 

The cross correlation analysis between 

the reservoir inflow and the other input variables 

including stream flow, rainfall and meteorological 

data showed that in general there was some 

correlation between the inflow of each reservoir 

and the mean humidity at stations 43201 and 

431401. The inflow of Lam Phra Phloeng and Lam 

Takong reservoirs was highly related to the stream 

flow at station M.89 and M.145 with the 

correlation coefficient of 0.79 and 0.70 

respectively but the correlation with rainfall was 

not good, correlation coefficients of 0.11-0.24. 

However, the inflow of Mun Bon and Lam Chae 

reservoir were fairly related to the rainfall at station 

25112 and 25152 with the correlation coefficient 

of 0.23 and 0.32, respectively. 

The autocorrelation or serial correlation 

analysis showed that, for Mun Bon and Lam Chae 

reservoirs, the autocorrelation coefficients of daily 

inflows were highly significant upto the first 30 

days lag. The autocorrelation coefficients were in 

the range of 0.17-0.82 and 0.42-0.85 for Mun Bon 

and Lam Chae reservoirs, respectively. The first 7 

days lag autocorrelation coefficients were 

between 0.24-0.73 for Lam Phra Phloeng inflow, 

0.35-0.78 for Lam Takong inflow, 0.33-0.75 for 

the M.145 streamflow and 0.38-0.79 for the M.89 

streamflow. 

Eighty percent of entire records were 

selected as the training data set for ANNs and 


the other 20% were used as the testing data set. 

The description of data selected for training and 

testing is shown in Table 2. 

Reservoir inflow forecasting model formulation 

In this study, four types of ANNs inflow 

forecasting models namely model A, B, C and D 

were developed for both single reservoir and multireservoir 

inflow forecasting. Each type of the 

model was divided into daily, weekly and monthly 

model. Single reservoir model was designed to 

forecast the inflow at one step ahead (lead time 

equal to 1) or inflow (t+1) of each reservoir. Multireservoir 

model could forecast the inflow at one 

step ahead of the four reservoirs simultaneously. 

The autocorrelation and cross correlation of the 

data were used to identify the model inputs in 

preliminary formulation of the forecasting model. 

The four single reservoir models were 

developed to use different levels of inputs. Model 

A used all available data of that reservoir. Model 

B used the data having relatively high correlation 

with the reservoir inflow such as the first three 

lags of reservoir inflow, stream flow, rainfall and 

some meteorological data. Model C used only the 

first three lags of the reservoir inflow and stream 

flow data. Model D used the first three lags of 

reservoir inflow, stream flow and rainfall data. 

Phien et al.(2000) forecasted the daily river flow 

of one day lead time (Q t+1) of several stations 

including Srinakarind and Khao Laem reservoirs 

in Mae Klong river basin, Thailand, and the 

Chukha reservoir in Bhutan using the lag zero and 

lag one of flow (Q t, Q t-1) and those of rainfall (R t, 

R t-1) as the input variables. Anmala et al.(2000) 

used the artificial neural networks for forecasting 

the watershed runoff in Kansas, USA. Monthly 

precipitation and temperature formed then inputs, 

and the monthly average runoff was chosen as the 

outputs. However this study proposed some more 

input variables, the first three lags of reservoir 

inflow and other high correlated variables. 

Similarly, there were four types of the multi-

eservoir forecasting models A, B, C and D which 

used the same levels of input data to the single 

reservoir model. The multi-reservoir model 

utilized more information on cross correlation 

coefficients among the reservoir inflow which was 

one of the advantages. However, the multireservoir 

model required more computational time 

than the single reservoir model for training. The 

detail input variables of ANNs reservoir 

forecasting models A, B, C and D are shown in 

Table 3. 

Training and testing of ANNs forecasting model 

The multi-layer feed forward neural 

networks with back propagation algorithm was 

selected for this study. Sigmoid transfer function 

was used in all the models. The initial weight (IW), 

momentum (M) and learning rate (LR) were 

initially set between 0.80-0.90. The number of 

epochs for training varied between 10,000 to 

100,000 depending on the performance efficiency 

(R 2 ) of the training. There were no fixed rules for 

designing the structures of ANNs, number of 

hidden layers and number of neurons in hidden 

layers. Many times the best fit model was 


Table 1 The basic data of the four reservoirs in Upper Mun basin. 

Basic data MB LC LP LK 

1. Catchment area (km 2 ) 454 601 807 1,430 

2. Annual rainfall (mcm) 1,047 1,039 1,112 920 

3. Annual inflow (mcm) 98 218 158 223 

4. Storage capacity(mcm) 141 275 110 310 

5. Irrigable area (rai) 44,600 113,750 67,760 127,540 

6. Municipality & downstream control (mcm/month) 0.4025 5.13 1.93 2.04 

Table 2 The training and testing data set. 

Reservoir Available data Training data set Testing data set 

1. Mun Bon Apr 1995-Jan 2000 Apr 1995-Dec 1998 Jan 1999-Jan 2000 

2. Lam Chae Jan 1999-Mar 2002 Jan 1999-Feb 2001 Mar 2001-Mar 2002 

3. Lam Phra Phloeng Jul 1992-Mar 2000 Jul 1992-Dec 1998 Jan 1999-Mar 2000 

4. Lam Takong Jul 1992-Mar 2000 Jul 1992-Dec 1998 Jan 1999-Mar 2000 

accomplished by trial and error processes. The 

optimal ANNs design was considered from the best 

performance training and testing by using R 2 . 

• Single reservoir forecasting model 

The ANNs structures and the training 

parameters of the best fit single reservoir 

forecasting models are presented in Table 4. The 

time series plot of the actual inflow and the 

predicted or forecasted inflow were compared for 

the daily, weekly and monthly models in Figures 

2 to 4, respectively. The best fit models were 

models A and B. The number of hidden layers 

varied between one to three layers but mostly 2 to 

3 layers. The number of neurons in hidden layers 

varied considerably from one model to the others. 

The models were trained by 80% of the data. The 

training result showed that all the ANNs models 

could produce an acceptable result in reservoir 

inflow forecasting especially the monthly model. 

The monthly model of all the reservoirs showed 

the better performance than the weekly and daily 

models. The R 2 of monthly, weekly and daily 

models were 0.95, 0.88 and 0.83, respectively, for 

Mun Bon; 0.90, 0.73 and 0.74 for Lam Chae; 0.93, 

0.89 and 0.90 for Lam Phra Phloeng and 0.97, 0.96,

252 


Table 3 Detail input variables of different reservoir inflow forecasting models. 

Reservoir Single reservoir Multiforecasting 

model MB LC LP LK reservoir 

A B C D A B C D A B C D A B C D A B C D 

Target output 

Inflow MB(t+1) 

Inflow LC(t+1) 

InflowMB(t+1) InflowLC(t+1) InflowLP(t+1) InflowLK(t+1) Input variables 

Inflow MB(t) • • • • • • • • 

Inflow MB(t-1,t-2,t-3) • • 

Inflow LC(t) • • • • • • • • 

Inflow LC(t-1,t-2,t-3) • • 

Inflow LP(t) • • • • • • • • 

Inflow LP(t-1,t-2,t-3) • • 

Inflow LK(t) • • • • • • • • 

Inflow LK(t-1,t-2,t-3) • 

Flow89(t) • • • • • • • 

Flow89(t-1,t-2,t-3) • 

Flow145(t) • • • • • • • 

Ra25093(t) • • • • • • • • 

Ra25102(t) 

Ra25112(t) • • • • • • • • 

Ra25152(t) • • • • • • • • • • • 

Ra25272(t) • • • • • 

Ra25511(t) • • • 

Ra25521(t) • • • 

Ra25612(t) • • • • • 

Ra25644(t) • 

Ra25650(t) • • • • 

Ra25751(t) • • • • • 

Ra25781(t) • • • • • 

Ra25272(t-1,t-2,t-3) • 

Ra25511(t-1,t-2,t-3) • 

AvgRH431201(t) • • • • • • • • • 

AvgRH431301(t) 

AvgRH431401(t) • • • • • • • 

MinRH431301(t) • • • 

MaxRH431301(t) • • • 

MinTemp431201(t) • • • • • • • 

MinTemp431301(t) • • • 

MinTemp431401(t) • • • • • • • 

MaxTemp431201(t) • • • • • • • 

MaxTemp431301(t) • • • 

MaxTemp431401(t) • • • • • • • 

AvgTemp431201(t) • • • • 

AvgTemp431401(t) • • • • 

AvgPres431201(t) • • • • • 

AvgPres431301(t) 

AvgPres431401(t) • • • • 

Inflow LP(t+1) 

Inflow LK(t+1)


Table 4 Training and testing result of ANNs models for single reservoir forecasting. 

Reservoirs Structures of ANNs R2 Models Structures IW M LR Training Testing 

(1) MB 

Daily MB_D 10-13-13-1 0.9 0.9 0.9 0.83 0.44 

Weekly MB_B 11-6-6-6-1 0.9 0.9 0.9 0.88 0.45 

Monthly 

(2) LC 

MB_B 11-4-4-4-1 0.9 0.9 0.9 0.95 0.50 

Daily LC_B 10-8-8-8-1 0.9 0.9 0.9 0.74 0.70 

Weekly LC_B 10-8-8-1 0.9 0.9 0.9 0.73 0.74 

Monthly 

(3) LP 

LC_B 10-3-3-3-1 0.9 0.9 0.9 0.90 0.79 

Daily LP_A 15-17-17-17-1 0.8 0.8 0.8 0.90 0.72 

Weekly LP_A 15-10-10-10-1 0.8 0.8 0.8 0.89 0.74 

Monthly 

(4) LK 

LP_A 15-5-5-5-1 0.8 0.8 0.8 0.93 0.82 

Daily LK_A 10-63-1 0.9 0.9 0.9 0.86 0.55 

Weekly LK_A 10-14-14-1 0.9 0.9 0.9 0.96 0.59 

Monthly LK_B 8-5-5-5-1 0.9 0.9 0.9 0.97 0.80 

Inflow (mcm.) 

Inflow (mcm.) 

12 

9 

6 

3 

0 

30 

25 

20 

15 

10 

Apr-95 

5 

0 

Actual 

Predicted 

R 2 

Daily Model Training 

= 0.83 

D:10-13-13-13-1 

Jul-95 

Nov-95 

Mar-96 

Jul-96 

Predicted 

Nov-96 

3 

2 

1 

0 

Mar-97 

R2 R 

10 

8 

4 

3 

2 

Testing 

6 

1 

0 

= 0.70 

0 2 

Actual 

3 

4 

2 

0 2 

Actual 

3 4 

2 = 0.44 Testing 

1 1 

Jul-97 

Nov-97 

Mar-98 

Jul-98 

Nov-98 

(a) ANNs training and testing [MB_D] (b) ANNs training and testing [LC_B] 

R2 Daily Model Training 

Actual 

Predicted 

= 0.90 

A:15-17-17-1 

Predicted 

12 

9 

6 

3 

0 

Testing 

R 2 = 0.72 

0 3 6 

Actual 

9 12 

Jul-92 

Oct-92 

Feb-93 

Jun-93 

Oct-93 

Feb-94 

Jun-94 

Oct-94 

Feb-95 

Jun-95 

Oct-95 

Feb-96 

Jun-96 

Oct-96 

Feb-97 

Jun-97 

Oct-97 

Jan-98 

May-98 

Sep-98 

(c) ANNs training and testing [LP_A] (d) ANNs training and testing [LK_A] 

Inflow (mcm.) 

Inflow (mcm.) 

40 

30 

20 

10 

0 

0 

Predicted 

Jan-99 

Mar-99 

May-99 

Jun-99 

Aug-99 

Oct-99 

R2 Daily Model Training 

Actual 

Predcted i 

= 0.86 

A:10-63-1 

Dec-99 

Feb-00 

Predicted 

20 

15 

10 

5 

0 


Apr-00 

Jun-00 

R2 Actual 

Predicted 

= 0.74 

B:10-8-8-8-1 

Aug-00 

Oct-00 

Dec-00 

Feb-01 

Testing 

R 2 = 0.55 

0 5 10 15 20 

Actual 

Jul-87 

Feb-88 

Oct-88 

Jun-89 

Feb-90 

Oct-90 

Jun-91 

Feb-92 

Oct-92 

May-93 

Jan-94 

Sep-94 

May-95 

Jan-96 

Sep-96 

May-97 

Jan-98 

Sep-98 

Figure 2 Comparison of the actual and predicted inflow of selected ANNs daily models.

254 

Inflow (mcm.) 

Inflow (mcm.) 

Inflow (mcm.) 

40 

30 

20 

10 

0 

R2 10 

8 

Testing 

6 

4 

R 2 

= 0.45 

0 

0 2 4 6 8 10 

Actual 

2 Weekly Model Training 

Actual 

Predicted 

= 0.88 

B:11-6-6-6-1 

Predicted 

Apr-95 

Jul-95 

Nov-95 

Mar-96 

Jun-96 

Oct-96 

Feb-97 

80 

60 

40 

20 

0 

80 

60 

40 

20 

Predicted 

R2 50 

40 

30 

20 

10 

0 

= 0.74 Testing 

0 10 20 30 40 

Actual 

50 

May-97 

Sep-97 

Jan-98 

Apr-98 

Aug-98 

Dec-98 

Weekly Model 

Training 

R2 Actual 

Predicted 

= 0.89 

A:15-10-10-10-1 

Jul-92 

Nov-92 

Apr-93 

Aug-93 

Jan-94 

Jun-94 

Oct-94 

Mar-95 

Jul-95 

Dec-95 

May-96 

Sep-96 

Feb-97 

Jun-97 

Nov-97 

Apr-98 

Aug-98 

0 


Inflow (mcm.) 

Inflow (mcm.) 

50 

40 

30 

20 

10 

0 

120 

100 

80 

60 

40 

20 

0 

Jan-99 

Weekly Model Training 15 

10 

R 

5 

0 

2 Actual 

Predicted 

= 0.73 

B:10-8-8-1 

Predicted 

Mar-99 

May-99 

Aug-99 

Oct-99 

Predicted 

Dec-99 

Feb-00 

R2 50 

40 

30 

20 

10 

0 

= 0.59 Testing 

0 10 20 30 40 50 

Actual 

Testing 

R 2 = 0.74 

0 5 10 15 

Actual 

(a) ANNs training and testing [MB_B] (b) ANNs training and testing [LC_B] 

May-00 

Jul-00 

Sep-00 

Dec-00 

Feb-01 

R2 Weekly Model 

Training 

Actual 

Predicted 

= 0.96 

A:10-14-14-1 

Jul-87 

Apr-88 

Jan-89 

Oct-89 

Jul-90 

May-91 

Feb-92 

Nov-92 

Aug-93 

May-94 

Mar-95 

Dec-95 

Sep-96 

Jun-97 

Mar-98 

(c) ANNs training and testing [LP_A] (d) ANNs training and testing [LK_A] 

Figure 3 Comparison of the actual and predicted inflow of selected ANNs weekly models. 

Inflow (mcm.) 

140 

120 

100 

80 

60 

40 

20 

Monthly Model Training 

Actual 

predicted 

R 2 = 0.95 

B:11-4-4-4-1 

Predicted 

60 

40 

20 

0 

Testing 

R 2 = 0.50 

0 20 40 60 

Actual 

Apr-95 

Jul-95 

Oct-95 

Jan-96 

Apr-96 

Jul-96 

Oct-96 

Jan-97 

Apr-97 

Jul-97 

Oct-97 

Jan-98 

Apr-98 

Jul-98 

Oct-98 

0 

Predicted 

90 

60 

30 

Jul-92 

Jan-93 

0 

Testing 

R 2 = 0.82 

03 0 60 90 

Actual 

Jul-93 

Jan-94 

Jul-94 

Jan-95 

Jul-95 


R2 Actual 

Predicted 

= 0.93 

A:15-5-5-5-1 

Jan-96 

Jul-96 

Jan-97 

Jul-97 

Jan-98 

Jul-98 

Inflow (mcm.) 

140 

120 

100 

80 

60 

40 

20 

0 

Jan-99 

Predicted 

80 

60 

40 

20 

0 

Mar-99 

Te sting 

R 2 = 0.79 

0 20 40 60 80 

Actual 

May-99 

Jul-99 

Sep-99 

Nov-99 

Jan-00 


R2 Actual 

Predicted 

= 0.90 

B:10-3-3-3-1 

Mar-00 

May-00 

Jul-00 

(a) ANNs training and testing [MB_B] (b) ANNs training and testing [LC_B] 

300 

250 

200 

150 

100 

50 

0 

Jul-87 


R2 Actual 

Predicted 

= 0.97 

B:8-5-5-5-1 

Aug-88 

Sep-89 

Oct-90 

Nov-91 

Dec-92 

Predicted 

Jan-94 

120 

80 

40 

0 

Feb-95 

Sep-00 

Testing 

(c) ANNs training and testing [LP_A] (d) ANNs training and testing [LK_B] 

Figure 4 Comparison of the actual and predicted inflow of selected ANNs monthly models. 

Inflow (mcm.) 

Nov-00 

Jan-01 

R 2 = 0.80 

0 40 80 120 

Actual 

Mar-96 

Apr-97 

May-98

0.86 for Lam Takong. Lam Takong showed the 

highest performance in comparison to other 

reservoirs. 

In general, the performance of this study 

was comparable to the study of Phien et al.(2000), 

where the efficiency index or R 2 of daily flow 

forecast for both Srinakarind and Khao Laem 

reservoirs and Chukha reservoir were above 0.82 

for calibration and 0.70 for validation of the model. 

The study of Anmala et al.(2000) showed R 2 of 

one month flow forecast of El Dorado watershed 

in Kansas, USA, to be 0.74 for the training set 

(from 1948-1955), 0.66 for the validation set 

(from1956-1963), and 0.61 for the testing set( from 

InflowLK(t-3) 

InflowLP(t-3) 

InflowMB(t-3) 

InflowLC(t-3) 

InflowLK(t-2) 

InflowLP(t-2) 

InflowMB(t-2) 

InflowLC(t-2) 

InflowLK(t-1) 

InflowLP(t-1) 

InflowMB(t-1) 

InflowLC(t-1) 

InflowLK(t) 

InflowLP(t) 

InflowMB(t) 

InflowLC(t) 

Figure 5 ANNs model C for multi-reservoir inflow forecasting. 


1964-1993). Atiya and Shaheen(1999) used the 

backpropagtion algorithm neural networks for 

forecasting the river flow of Nile river in Egypt. 

The multistep ahead forecasting was employed by 

using 10-day and one month time steps. The 

normalized root mean square (NRMS) was used 

as the forecast performance indicator instead of 

R 2 . The NRMS was less than 0.60 in most cases. 

The one step forecast performed better than the 

multistep. 

The remainder 20% of the data were used 

for model testing. Although the testing phase 

showed R 2 lower than the training phase, most of 

the R 2 were above 0.50 and more than half of them 

InflowLK(t+1) 

InflowLP(t+1) 

InflowMB(t+1) 

InflowLC(t+1)

256 

were above 0.70. The testing result was quite 

satisfactorily. 

However, it appeared from Figures 2 to 

4 that there was higher forecasting error during 

the peak reservoir inflow than other periods. 

There were the common characteristics in all 

the daily, weekly and monthly models. If 

one would like to improve the accuracy of peak 

inflow forecast, one should develop the ANNs 

model for the peak inflow period in particular. It 

was also observed that the training parameters 

including the learning rate, initial weights and 

momentum were not sensitive to the prediction 

performance. 

• Multi-reservoir forecasting model 

The multi-reservoir inflow forecasting 

models A, B, C and D were experimented on daily, 

weekly and monthly data of Mun Bon, Lam Chae, 

Lam Phra Phoeng and Lam Takong reservoirs to 

find the best fitted ANNs models. The ANNs 

structure of model C is shown in Figure 5 as an 

example. The experiment covered one to three 

hidden layers with different neurons in each layer. 

The result of training and testing the ANNs showed 


that two hidden layers for daily and weekly models 

were indifferent from three hidden layers while 

one hidden layer for monthly model was 

indifferent from two and three hidden layers. The 

result showed that sigmoid transfer function was 

acceptable. The structures of selected models for 

multi-reservoir inflow forecasting are show in 

Table 5. The R 2 from training was in the ranges of 

0.49 to 0.98 but mostly above 0.80. This indicated 

good training performance as shown in Figure 6. 

The R 2 from testing was acceptable in general. It 

was in the ranges of 0.17 to 0.64 except Mun Bon 

reservoir where the R 2 was very low due to high 

variation of Mun Bon inflow during test period. 

The monthly model showed that, in general, the 

performance of multi-reservoir ANNs model as 

indicated by R 2 was not as good as the single 

reservoir model because the data available for 

training and testing were shorter. The data of the 

four reservoirs were available in different periods. 

For example, the inflow were available from 

1995-2000, 1999-2002, 1992-2000 and 1987-2000 

for Mun Bon, Lam Chae, Lam Phra Phloeng and 

Lam Takong, respectively. Thus only the data of 

Table 5 Training and testing result of ANNs multi-reservoir forecasting model. 

Models Parameters Structures R 2 training R 2 testing 

IW M LR of ANNs MB LC LP LK MB LC LP LK 

Daily 

A 0.8 0.8 0.8 28-22-22-4 0.72 0.81 0.84 0.82 0.02 0.57 0.32 0.29 

B 0.8 0.8 0.8 15-18-18-4 0.69 0.80 0.83 0.80 0.06 0.64 0.41 0.28 

C 0.8 0.8 0.8 16-18-18-4 0.74 0.83 0.81 0.82 0.05 0.63 0.34 0.24 

D 0.8 0.8 0.8 14-18-18-4 0.68 0.80 0.82 0.81 0.05 0.64 0.33 0.27 

Weekly 

A 0.8 0.8 0.8 28-13-13-4 0.82 0.87 0.68 0.63 0.05 0.59 0.21 0.48 

B 0.8 0.8 0.8 15-10-10-4 0.66 0.73 0.49 0.51 0.08 0.57 0.20 0.51 

C 0.8 0.8 0.8 16-10-10-4 0.89 0.98 0.89 0.92 0.05 0.39 0.21 0.58 

D 0.8 0.8 0.8 14-10-10-4 0.88 0.96 0.92 0.94 0.15 0.54 0.40 0.54 

Monthly 

A 0.8 0.8 0.8 28-21-4 0.96 0.77 0.89 0.84 0.04 0.51 0.19 0.56 

B 0.8 0.8 0.8 15-14-4 0.91 0.85 0.85 0.88 0.04 0.22 0.23 0.17 

C 0.8 0.8 0.8 16-14-4 0.50 0.69 0.78 0.89 0.05 0.43 0.24 0.63 

D 0.8 0.8 0.8 14-14-4 0.71 0.78 0.84 0.88 0.08 0.36 0.32 0.27

Inflow (mcm.) Inflow (mcm.) 

Inflow (mcm.) Inflow (mcm.) 

Inflow (mcm.) 

Inflow (mcm.) 

5 

4 

3 

2 

1 

0 

12 

10 

8 

6 

4 

2 

0 

16 

12 

8 

4 

0 

50 

40 

30 

20 

10 

0 

40 

30 

20 

10 

0 

120 

90 

60 

30 

0 

Jan-99 

Jan-99 


Actual 

Predicted 

C:16-18-18-4 

Apr-99 

Jul-99 

Oct-99 

Jan-00 

Weekly Model Training 

Actual 

Predicted 

C:16-10-10-4 

Apr-99 

Jul-99 

Oct-99 

Dec-99 

Apr-00 

Mar-00 


Actual 

Predicted 

C:16-14-4 

Apr-99 

Jul-99 

Oct-99 

Jan-00 

Apr-00 


Jun-00 

Jul-00 

R2 MB 

= 0.74 

12 

10 

8 

6 

4 

2 

180 

15 

R 

12 

2 LP 

= 0.81 

Oct-00 

9 

6 

3 

0 

Jan-01 

Jan-99 

Apr-99 

(a) Daily model 

R2 MB 

= 0.89 

60 

R 

50 

40 

2 LP 

= 0.89 

0 

Sep-00 

Nov-00 

40 

30 

20 

10 

30 

20 

10 

(b) Weekly model 

100 

R 80 

2 MB 

= 0.50 

0 

Feb-01 Jan-99 

60 

40 

20 

1600 

R 120 

2 LP 

= 0.78 

80 

40 

0 

Apr-99 

Apr-99 

Jul-99 

(c) Monthly model 

Figure 6 Comparison of the actual and predicted inflow of ANNs model C. 

Jul-00 

Oct-00 

Jan-01 

Jul-99 

Jul-99 

Oct-99 

Oct-99 

Oct-99 

Jan-00 

Dec-99 

Jan-00 

Apr-00 

Mar-00 

Apr-00 

Jul-00 

Jun-00 

Jul-00 

Sep-00 

R2 LC 

= 0.83 

R2 LK 

= 0.82 

Oct-00 

Nov-00 

Jan-01 

R2 LC 

= 0.98 

R2 LK 

= 0.92 

Oct-00 

Jan-01 

Feb-01 

R2 LC 

= 0.69 

R2 LK 

= 0.89

258 

1999-2000 could be used for training and testing 

the multi-reservoir ANNs model. 

It was difficult to distinguish which of 

the models A, B, C and D was the better one by R 2 

in Table 5. Some models gave good result for some 

reservoirs. Model C used less data, only the 

reservoir inflow and stream flow, was more 

attractive than the other models. Moreover, the 

performance of daily, weekly and monthly models 

were not much different. Theoretically, the ANNs 

multi-reservoir model should provide the better 

result, but it was not true for this case due to the 

limited data. The multi-reservoir problem required 

a lot longer training time than the case of single 

reservoir, particularly the case of daily model. This 

could be the disadvantage of the ANNs multireservoir. 

CONCLUSIONS 

For single reservoir forecasting, models 

A and B showed better performance (R 2 ) than 

models C and D. The monthly model showed the 

better result than the weekly and daily models. For 

multi-reservoir forecasting, the performance of the 

four models was not different. Model C was 

recommended since it required less data. The 

training and testing performance of daily, weekly 

and monthly models were not much different in 

case of multi-reservoir. However, the multireservoir 

problem required a lot longer training 

time than the single reservoir problem, particularly 

the case of daily model. This could be the 

disadvantage of the ANNs multi-reservoir inflow 

forecasting. In general, the single reservoir inflow 

forecasting showed the better result. 


The authors wish to thank Kasetsart 

University Research and Development Institute 

and Faculty of Engineering, Kasetsart University 

for funding this study. 



Anmala, J., B. Zhang and S. Rao. 2000. 

Comparison of ANNs and empirical 

approaches for predicting watershed runoff. 

J.Water Resour.Plng. and Mgmt., ASCE. 

126(3):156-166. 

Atiya, A. and S.I. Shaheen. 1999. A comparison 

between neural-network forecasting 

techniques-case study : river flow forecasting. 

IEEE Transactions on Neural Networks. 

10(2): 402-409. 

Phathravuthichai, S. and V. Vudhivanich. 2003. 

Flood forecasting in Lam Phachi river 

basin by MIKE11 Model and artificial 

neural networks. M.S. thesis, Kasetsart 

University, Bangkok. 

Phien, H.N., W. Phuetphan and N. Dukpa. 2000. 

B-spine networks for daily Flow forecasting, 

pp.1-10. In Proceedings of the International 

European–Asian Work shop on Ecosystem 

and Flood. Hanoi, Vietnam, June 2-29, 2000. 

Supharatid, S. 2002. Water forecasting by 

artificial neural network. Irrigation 

Development Institute, Thailand. 96 p. 

Tingsanchali, T. and C. Manusthiparom. 2001. A 

neural network model for flood forecasting 

tidal rivers. Water Engineering and 

Management Program, School of Civil 

Engineering, Asian Institute of Technology, 

Phathumtanee. 

Tokar, A.S. 1996. Rainfall-runoff modeling in 

an uncertain environment. Doctoral 

dissertation, University of Maryland, College 

Park, Md. USA. 

Tokar, A.S. and M. Markus. 2000. Precipitationrunoff 

modeling using artificial neural 

networks and conceptual models. 

J. Hydrologic Engrg., ASCE. 4(3): 232-239. 

Vudhivanich, V. 2001. Artificial neural networks. 

Wan Chuchart 4 January 2001. Irrigation 

Engineering Alumni Association under the 

Royal Partonage of the King, Bangkok. p89- 

102.

Vudhivanich, V. and A. Rittima. 2003. Application 

of artificial neural networks for reservoir 

inflow forecasting : Lam Takong reservoir 

case study. Final Report. Department of 

Irrigation Engineering, Faculty of 

Engineering, Kasetsart University, 

Kamphaengsaen Campus, Nakhon Pathom. 


Vudhivanich, V., Thongpumnak, S., Cherdchanpipat, 

N., Kasempun, N. and A. Rittima. 2004. 

Reservoir inflow forecasting by artificial 

neural networks, pp. 24-31. In Proceedings 

of 42 nd Kasetsart University Annual 

Conference. Kasetsart University, Bangkok.


Jordan Derivations on Rings 

ABSTRACT 

Orapin Wootijiruttikal and Utsanee Leerawat 

An additive mapping d : R→R is called a Jordan derivation on a ring R if d(a 2 ) = d(a)a + ad(a) 

for all a ∈ R. Two general forms of d n (aba) and d n (abc+cba), where a,b,c ∈ R and n ∈ , are 

established. It is also shown that if d is a Jordan derivation on a commutative ring R and P is a semiprime 

ideal or prime ideal of R where R/P is characteristic – free, then d(P) ⊆ P if and only if d n (P) ⊆ P for 

all positive integers n ≥ 2. 

Key words: derivation, Jordan derivation, ring 

INTRODUCTION 

An additive mapping d : R → R is called 

a derivation on R if d(ab) = d(a)b + ad(b) for all 

pairs a, b ∈ R. An additive mapping d : R → R is 

called a Jordan derivation if d(a 2 ) = d(a)a + ad(a) 

holds for all a ∈ R. Obviously, every derivation is 

a Jordan derivation. 

Herstein(1957) showed that if d is a 

Jordan derivation on a ring R such that 

characteristic is not 2, then d(aba) = d(a)ba + 

ad(b)a + abd(a) and d(abc+cba) = d(a )bc + 

ad(b)c + abd(c) + d(c)ba + cd(b)a + cbd(a) for 

all a,b,c ∈ R. 

In 1998, Creedon investigated a ring of 

characteristic – free, semiprime ideal on ring and 

proved that, if d is a derivation on a ring R and P 

is a semiprime ideal of R, such that R/P is 

characteristic – free and d k (P) ⊆ P, for any fixed 

positive integer k, then d(P) ⊆ P. 

In this paper we will consider d as a 

Jordan derivation on a ring R and establish general 

forms of d n (aba) and d n (abc+cba) for all positive 

integer n. 

We also show that if d is a Jordan 

derivation on a commutative ring R and P is a 

semiprime ideal or prime ideal of R where R/P is 

characteristic – free, then we have d(P) ⊆ P if and 

only if d n (P) ⊆ P for all positive integer n ≥ 2. 


For the proof results we need the 

following lemmas. 

Lemma 1. [Herstein, Lemma 3.1] Let R be a ring 

such that characteristic is not 2 and let d be a 

Jordan derivation on R. Then for all a, b, c ∈ R 

the following statements hold : 

(i) d(ab + ba) = d(a)b + ad(b) + d(b)a+ bd(a), 

(ii) d(aba) = d(a)ba +ad(b)a + abd(a), 

(iii) d(abc + cba) = d(a)bc + ad(b)c 

+ abd(c) + d(c)ba+ cd(b)a + cbd(a). 

Lemma 2. [Creedon, Lemma 4 ] Suppose P is a 

semiprime ideal of a ring R and L is a left ideal 

satisfying L n ⊆ P, for some positive integer n. Then 

L ⊆ P. 

Department of Mathematics, Faculty of Science, Kasetsart University, Bangkok 10900, Thailand. 

* Corresponding author, e-mail: fsciutl@ku.ac.th 



Theorem 1. Let R be a ring such that characteristic 

is not 2. If d is a Jordan derivation on R, then for 

all a, b ∈ R and for each positive integer n, 

dn (aba) 

⎛ n ⎞ i j k 

= ∑ ⎜ ⎟d 

( a) d ( b) d ( a) 

⎝i 

j k⎠ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

PROOF. The proof will be finished by induction 

on n. 

If n = 1, the result is just by Lemma 1 

(ii) in materials and methods, so we assume n > 1 

and assume that d n (aba) 

= 

∑ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

⎛ 

⎜ 

⎝i 

n ⎞ i j k 

⎟d 

( a) d ( b) d ( a ) . 

j k⎠ 

Since d n+1 (aba) = d(d n (aba)), d n+1 (aba) = 

⎡ 

⎤ 

⎢ 

⎛ n ⎞ 

⎥ 

i j k 

d⎢∑⎜⎟d ( a) d ( b) d ( a) 

⎥ 

⎢ 

i j k 

i++ j k= n ⎝ ⎠ 

⎥ 

⎣⎢ 

ijk ,, ∈{,, 

012, 

K, 

n} 

⎦⎥ 

= ∑ 

⎛ n + 1 ⎞ i j k 

⎜ ⎟d 

( a) d ( b) d ( a). 

⎝i 

j k⎠ 

i++ j k= n+ 

1 

ijk ,, ∈ {,, 012, K, 

nn , + 1} 

the proof is complete. # 

Theorem 2. Let R be a ring such that characteristic 

is not 2. If d is a Jordan derivation on R, then for 

all a,b,c ∈ R and for each positive integer n, 

d n (abc+cba) 

= ∑ 

i++ j k= n 

i,j,k ∈{0,1,2, 

K,n} 

⎛ 

⎜ 

⎝i 

n ⎞ i j k 

⎟[ 

d ( a) d ( b) d ( c) 

j k⎠ 

i j k 

+ d () c d () b d ()] a 

PROOF. We linearize the result of Theorem 1 by 

replacing a by a+c to obtain 

dn ((a+c)b(a+c)) 

⎛ n ⎞ i j k 

= ∑ ⎜ ⎟d 

( a+c) d ( b) d ( a+c) 

⎝i 

j k⎠ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 


= 

∑ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

⎛ 

⎜ 

⎝i 

+ ∑ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

⎛ 

⎜ 

⎝i 

⎛ 

+ ∑ ⎜ 

++ = ⎝i 

i j k n 

ijk ,, ∈{,, 

012K , , n} 

+ 

∑ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

= d n (abc) 

+ 

∑ 

i++ j k= n 

ijk ,, ∈{,, 

012K , , n} 

⎛ 

⎜ 

⎝i 

⎛ 

⎜ 

⎝i 

⎛ 

+ ∑ ⎜ 

++ = ⎝i 

i j k n 

ijk ,, ∈{,, 

012K , , n} 

+ d n (cbc). 

n ⎞ i j k 

⎟d 

( a) d ( b) d ( a) 

j k⎠ 

n ⎞ i j k 

⎟d 

( a) d ( b) d ( c) 

j k⎠ 

n ⎞ i j k 

⎟d 

() c d () b d () a 

j k⎠ 

n ⎞ i j k 

⎟d 

() c d () b d () c 

j k⎠ 

n ⎞ i j k 

⎟d 

( a) d ( b) d ( c) 

j k⎠ 

n ⎞ i j k 

⎟d 

() c d () b d () a 

j k⎠ 

On the other hand, d n ((a+c)b(a+c)) 

= d n (aba + cbc + cba + abc) 

= d n (aba) + d n (cbc) + d n (cba+abc) 

Comparing the two equations, we get 

the result. # 

For the next results, we need the 

following lemmas. 

Lemma 3. Let P be an ideal of a ring R such that 

characteristic is not 2. If d is a Jordan derivation 

on R, then for all a ∈ P and for all positive integers 

r we have d r (a s ) ∈ P for all s = r+1,r+2,... 

PROOF. We will proof by induction on r. 

Obviously, for r = 1 d(a 2 ) ∈ P. If s ≥ 2, d(a s+1 ) = 

d(aa s-1 a). Using Lemma1(ii) in Materials and 

methods, with b = a s-1 we obtain d(aa s-1 a) = d(a) 

a s-1 a + ad(a s-1 )a + aa s-1 d(a). Since a ∈ P, 

d(aa s-1 a) ∈ P. Hence d(a s ) ∈ P for all s ≥ 2. Next,

262 

we assume that r ≥ 2 and d r (a s ) ∈ P for all r=2,...,n 

and s = r+1,r+2,... Since d n+1 (a n+2 ) = d n+1 (aa n a), 

by theorem 1 we have d n+1 (aa n a) 

= 

∑ 

i++ j k= n+ 

1 

ijk ,, ∈ {,, 012, K, 

nn , + 1} 

⎛ n + 1 ⎞ i j n k 

⎜ ⎟d 

( a) d ( a ) d ( a). 

⎝i 

j k⎠ 

Therefore d n+1 (aa n a) ∈ P because a ∈ P. Hence 

d n+1 (a n+2 ) ∈ P. Let n+2 = t, u ∈ and suppose 

that d n+1 (a s ) ∈ P for all s = t+1,t+2,...,t+u. Since 

d n+1 (a t+u+1 ) 

= ∑ 

i++ j k= n+ 

1 

ijk ,, ∈ {,, 012, K, 

nn , + 1} 

⎛ n + 1 ⎞ i j n+ u+ 1 k 

⎜ ⎟d 

( a) d ( a ) d ( a) 

⎝i 

j k⎠ 

and a ∈ P, then d n+1 (a t+u+1 ) ∈ P. Hence d n+1 (a s ) ∈ 

P for s > n+1. Thus we conclude that d r (a s ) ∈ P 

for all s = r+1,r+2,... # 

Lemma 4. Let P be an ideal of R such that 

characteristic is not 2 and let x∈P. Suppose that d 

is a Jordan derivation on R. Then d n (x n ) – n!(dx) n 

∈ P for all positive integers n. 

PROOF. By definition of prime ideal, the result 

holds for n = 1. If n ≥ 2, suppose d n (x n ) – n! (dx) n 

∈ P. Since d n+1 (x n+1 ) = d n+1 (xx n-1 x), by Theorem 1 

we have d n+1 (xx n-1 x) 

= ∑ 

i++ j k= n+ 

1 

ijk ,, ∈ {,, 012, K, 

nn , + 1} 

⎛ n + 1 ⎞ i j n−1k ⎜ ⎟d 

( x) d ( x ) d ( x). 

⎝i 

j k⎠ 

Since x ∈ P, 

⎛ n + 1 ⎞ i j n−1k ∑ ⎜ ⎟d 

( x) d ( x ) d ( x) 

i 

i++ j k= n+ 

⎝ j k 

1 ⎠ 

ijk ,, ∈ {,, 012, K, 

nn , + 1} 

⎛ n + 1 ⎞ n−1 n−1 

= ⎜ ⎟ dxd ( ) ( x ) dx ( ) + P. 

⎝1 

n -1 1⎠ 

Thus d n+1 (x n+1 ) ∈ (n+1)n [d(x) 

d n-1 (x n-1 )d(x)] + P, and so d n+1 (x n+1 ) ∈ (n+1)! 

(dx) n-1 + P. Hence d n+1 (x n+1 ) - (n+1)!(dx) n+1 ∈ P. 

Inductively we have d n (x n )-n!(dx) n ∈ P. # 


Theorem 5. If d is a Jordan derivation on a 

commutative ring R and P is a semiprime ideal of 

R for which R/P has characteristic-free, then d(P) 

⊆ P if and only if d n (P) ⊆ P for all positive integer 

n ≥ 2. 

PROOF. Suppose that d n (P) ⊆ P for all positive 

integer n≥2. Lemma 4 says that d n (x n ) – n!(dx) n ∈ 

P for all x∈P. Since d n (x n )∈P, we see that n! [d(x)] n 

∈ P. Hence, [d(x)] n ∈P. Consider the ideal 

+P of R, we have (+P) n ⊆ P. 

Therefore, by Lemma 2 in Materials and Methods, 

we see that (+P) ⊆ P, and hence d(P) ⊆ P. 

The reverse implication is obvious. # 

The following theorem can be proven in 

a similar way. 

Theorem 6. If d is a Jordan derivation on a 

commutative ring R and P is a prime ideal of R 

such that P≠R and R/P has characteristic-free, 

then d(P) ⊆ P if and only if d n (P) ⊆ P for all 

positive integer n≥2. 

CONCLUSION 

Results of the studying show that: 

1. If d n is a Jordan derivation on a ring R 

such that characteristic is not 2, then 

1.1 d n (aba) 

⎛ 

= ∑ ⎜ 

++ = ⎝i 

i j k n 

ijk ,, ∈{,, 

012,..., 

n} 

1.2 (abc+cba) 

= ∑ 

i++ j k= n 

ijk ,, ∈{,, 

012,..., 

n} 

⎛ 

⎜ 

⎝i 

n ⎞ i j k 

⎟d 

( a) d ( b) d ( a) 

j k⎠ 

⎟[ ] 

n ⎞ i j k 

d ( a) d ( b) d ( c) 

j k⎠ 

+ d i (c)d j (b)d k (a)] 

for all a,b,c ∈ R and for any positive 

integer n.

2. If d is a Jordan derivation on a 

commutative ring R and P is a semiprime ideal of 

R (or prime ideal of R such that P≠R) for which 

R/P has characteristic-free, then d(P) ⊆ P if and 

only if d n (P) ⊆ P for all positive integer n≥2. 


This work was supported by a grant from 

the Graduate School, Kasetsart University. 


LITURATURE CITTED 

Creedon, T. 1998. Derivations and prime ideals. 

Mathematical Proceedings of the Royal 

Irish Academy. 2: 223-255. 

Herstein, I.N. 1957. Jordan derivations of prime 

rings. Proceedings of the American 

Mathematical Society. 8: 1104-1110.


Geoinformatic Public Domain System Model “ SWAT “ in Thailand 

ABSTRACT 

Hansa Vathananukij 

The Chaophraya river basin has been supremely presumed to regime characterize central plain 

area. Prototype domains were researched with the SWAT model on medium and huge areas of both 

unregulated and regulated reaches upon steep slope, hilly and plain features on the Chaophraya tributaries 

in northern Thailand. 1,500 and 10,200 sq. kilometers domains situated in mountainous feature were 

unregulated studiousness, the other largest research domain of 14,500 sq.kilometers was regulated by 

Pasak Cholasit reservoir. Essential hydrological data inquiry was two stations at Ping river, nine stations 

at Nan river and eleven stations at Pasak river. 1:50,000 geoinformatic system was much conveniently 

spatial analyzed for federation and assessment. Base flow severance was stream flow hydrograph 

separation and analysis program was assembled by sliding interval methodology. Data analysis interface 

was modified by data analysis programming and data access methodology was approached by basic 

standard query language. This appropriate model calibrations and verifications, admissible ensued on 

above ninety percentage of correlation efficiency and affirmable best arbitrated on large scale-mild 

slope potentiality together with un-implied in both continual rainfall investigation and sufficient number 

of stations. 

Key words: SWAT, soil and water assessment tools, geoinformatic system, stream flow hydrograph 

separation and analysis program, standard query language 

INTRODUCTION 

While most popular water resource 

models employed in Thailand, were costly 

licensed, Thai researchers had composed several 

models but still repelled. Nowadays, an innovative 

SWAT(soil and water assessment tools) has been 

a rather renowned public domain model in water 

resource which is updated to geoinformatic system 

model (Bahram,1992). Whereas the Chaophraya 

regime had been regional characterized by 

monsoon, spatial and intensive heavy rainfall 

increased mean rainfall intensity attitude as rainfall 

variations with plenitude that could cause floods, 

landslide and debris flow (Vathananukij, 1998). 

This inundation simulation which mainly caused 

of water-related extreme event (Micheal,1998; 

Muttiah and Allen, 1999), was attempted to 

structure and verify through standard public 

domain SWAT/GIS model to the Chaophraya 

tributaries. 

METHODOLOGY 

Public domain SWAT model 

SWAT the “Soil and Water Assessment 

Tools”, a model in water resource with continuous 

time, basin scale and geoinformatic system 

Water Resource Engineering Department, Faculty of Engineering, Kasetsart University, Bangkok 10900, Thailand. 

Corresponding author, e-mail: fenghsv@ku.ac.th 


interface, has been capably counseled maxima to 

five hundred drainage basins calibration. Adequate 

hydro-meteorology continuance investigation 

might insinuate reliable correlation coefficient 

while SWAT model provided contingent 

spontaneous precipitation (Srinivasan et al., 1995; 

Rosenthal et al.,1995; Srinivasan et al., 1996). 

Thus model was physically based requiring 

specific inputs for weather, soil, topography, 

vegetation, land management practice and 

watershed, which would be essential partitioned 

into hydrologic response units (HRU). These HRU 

(sub-watershed or sub-basins) severely based on 

land uses, soil types, channels, ponds or reservoirs. 

Hydrology simulation was separated into land 

phase and stream phase, while land phase 

controlled main channel transportation on water, 

sediment and agricultural chemicals but stream 

phase would be transported through channel grids 

into basin outlet (Sloto and Crouse, 1996; Steven, 

1998). Flood hydrograph and base flow analysis 

employed hydrologic respond unit together with 

stream flow hydrograph separation and analysis 

program (HySeP) (Vathananukij and 

Thanasiriyakul, 2002). Hydrological sub basin was 

basically on the following formulae. 

Surface runoff duration time could present with 

N = A 0.2 …(1) 

where N = Number of days with no runoff 

A = Drainage area in square-miles. 

Water balance equation was 

SW t = SW + ∑(R i - Q i - ET i - P i - QR i) …(2) 

where SW = Soil water content (15 Bar) 

t = Time (days) 

R = Daily precipitation 

Q = Daily runoff 

ET = Daily evapotranspiration 

P = Daily percolation 

and QR = Daily return flow. 

SCS curve number method was (infiltration 

determination) as follows: 

Q = 

( ) 

R−02 . s 

R+ 08 . s 

2 

, R> 02 . s …(3) 


Q=0.0, R < 0.2s …(4) 

s = 254 100 ( −1) CN 

…(5) 

where Q = Daily runoff 

R = Daily rainfall 

s = Retention parameter 

and CN = Curve number. 

Assigned based on soil type, land 

cover and initial moisture 

conditions. 

The SWAT model was principaly 

designed to assess continuous incidents and long 

time periods of both natural and au-natural 

activities which could advantageously employed 

to formulate and calibrate from small to very large 

basins. This research was proved that local 

geoinformatic system execution could perform and 

illustrate precisely synchronism with this public 

domain model (Vathananukij, 2003a; 2004 a). 

Data assimilation 

Intricacy and simplicity model as SWAT/ 

GIS has been tenuous simulated upon tropical 

precipitation in Thailand (Vathananukij, 2003b). 

Figure 1 shows three different prototype domains 

at Ping(Chiang Mai), Nan and Pasak river basins 

whose topography with mainly mountainous and 

hilly features. Two domains occupied 1,500 and 

10,156 square kilometers while other comparable 

prototype, Pasak river domain, was mountainous 

on upper part and slope down to dam site outlets, 

occuping 14,520 square kilometers. 

Selective hydro-informatic data from 

precipitation gauge stations under responsibility 

of the Royal Irrigation Department (RID) were 

stations number: 28013, 28032, 28042, 28053, 

28073, 28102, 28111, 19052, 19092, 19113, 

19342, 19351, 19360, 19411, 25132, 25172, 

25272, 25470, 25612, 36013, 36023, 36032, 

36043, 36052, 36062, 36082, 36092, 36104, 

36122, 36192 while Thailand Meteorology 

Department (TMD) responsibility were stations

266 

Figure 1 Prototype domains. 

numbers: 327008, 327027, 376203, 331008, 

331009 and Agro-meteorological stations were 

numbers: 28142, 28152 and 19342. According to 

sufficient precipitation from 1980 to 2003 at 

Chiang Mai, Nan and Pasak provinces, rainfall 

consistency and analysis are illustrated in Figure 

2 (Chow, 1988; Vathananukij and Paewpisakul, 

2003a). 

Geoinformatic system (George, 1997) 

which illustrated topography, soil, land cover and 

river system is used as interface to equalize digital 

data and converted to model format(Craig,1981), 

thus became the most essential availability for 

model initiation, calibration and verification 

occurrence (Vathananukij, 2003a; 2004 b). All 

digital information commencement and concern 

were recommended as presented in Table1. Figure 

3 shows those substantial digital information on 

these three prototype domains (Vathananukij, 

2005). 

Calibration 

Streamflow investigation by RID 


gauging stations (P64 and PN8 along periods 1990 

to 2000 A.D. for Chiang Mai prototype; N.17, 

N.42, N.49, N.50, N.51, N.63, N.1, N.13A and 

N.35 from 1980 to 2001 A.D. for Nan prototype; 

S.4B, S.9, S.10, S.12, S.13, S.14, SM.1 and SM.2 

from 1982 to 2003 A.D. for Pasak prototype) have 

been calibrated and verified through SWAT model 

(Vathananukij, 2002; 2003b; 2004b). Most 

calibration consequence greatly affirmed model 

consistency. Substantial calibrated domains were 

shown in Figure 4. 

DATA ANALYSIS PROGRAMMING 

Programming was principled on stream 

flow with two principal sources; surface runoff 

and ground water, where base flow was separated 

from daily stream flow using sliding interval 

method adapted from USGS program named 

“HySeP” (Arnold et al., 1995). SWAT model was 

separately calibrated against both observed surface 

flow and base flow between two gauge stations 

which covered entire period of interest. Base flow 

hydrograph separation is shown in Figure 5. 

Digital data base with spreadsheet software have 

capability to use macro script for both data 

calibration and tabular presentation ( Ling, 1996; 

Mauro et al., 1998; Vathananukij, 2003). Data 

access basic software and standard query language 

were mainly developed for substance utilization 

in order to convenient digitally calibrate and 

analyze SWAT model. Results are illustrated in 

Figure 6. 

RESULTS 

Overwhelm simulation majorly caused 

water-related normal to extreme event, was 

attempted to structure and verified the 

geoinformatic public domain SWAT model. 

Comparison outcome on correlative efficiency 

illustrated the best presage on mild slope feature 

and good relation on steep slope feature where

egulated prototype signified above ninety seven 

percent and unregulated prototype index over 

seventy percent. Table 3 shows results in feature 

Figure 2 Hydro-informatics location and consistency analysis. 


slope, a number of rain gauge intensity and 

correlation coefficient between simulated and 

investigated surface flow. 

Table 1 Geoinformatic system information for SWAT model. 

Data type Resource department Scale 

Topography - Royal Survey Thailand Department 1:50,000 

Land Cover - Land Development Department 1:50,000 

Soil - Land Development Department 1:50,000 

International Geosphere–Biosphere Programe (IGBP-DIS) 

Weather - Thai Meteorology Department (TMD) N/AN/A 

Royal Irrigation Department (RID) 

Stream Flow - Royal Irrigation Department (RID) N/A

268 

DISCUSSION 

Consequences indicated that SWAT 

model (the first standard geoinformatic public 

domain model which exercised through Thailand 

hydroinformatic system) which could perform 

basis calibration up to five hundred sub-basins, 

has been standardized and affirmable best 

arbitrated (62~97 % efficiency)on both normal to 

large scale area (1,000 to 14,000 squared 

kilometers domains) and mild slope potentiality 

although un-implied and insufficient in continual 


Figure 3 Geoinformatic system analysis linkage and performance. 

precipitation observation. 

In addition to data analysis program 

development which was much efficient 

investigation to operate, calibrate and verify 

through SWAT model. 

Open source SWAT model assessment 

did essentially employ on geoinformatic system 

and digital elevation model capability. Furtherance 

could be more concerned in integration and 

management together with decision supporting 

system automation program. 

Figure 7 illustratively indicated both

Figure 4 Threshold stream network and calibrated domains. 

Figure 5 Base flow separation. 

Figure 6 Data analysis program for model calibration. 

Kasetsart J. (Nat. Sci.) 40(1) 269

270 


Table 3 Correlation on calibrated basin. 

Calibration Stream Basin area Basin Rain gauge Correlation coefficient (R 2 ) 

area gauge Km 2 slope intensity total flow 

Area1 P.64 487.18 0.266 1 0.5795 

Area2 PN.8 1,567.67 0.299 1 0.5361 

Area3 N.17 1,091.15 0.306 0 0.3512 

Area4 N.42 2,047.11 0.296 2 0.6219 

Area5 N.49 153.16 0.317 0 0.4784 

Area6 N.50 194.61 0.324 0 0.4775 

Area7 N.51 758.53 0.240 1 0.5284 

Area8 N.63 776.05 0.202 1 0.4099 

Area9 N.1 4,495.06 0.241 8 0.6008 

Area10 N.13A 8,566.91 0.237 8 0.7045 

Area11 N.35 10,156.01 0.230 11 0.7008 

Area12 S.10 300.79 0.100 0 0.6037 

Area13 SM.1 1,132.70 0.248 2 0.7585 

Area14 S.4B 3,321.00 0.181 3 0.8285 

Area15 S.12 476.11 0.230 0 0.5450 

Area16 SM.2 7,329.49 0.158 6 0.8089 

Area17 S.13 395.25 0.100 0 0.7182 

Area18 S.14 1,252.77 0.087 0 0.8572 

Area19 S9.(Un) 14,323.80 0.068 13 0.7397 

Area20 S9.(Re) 14,323.80 0.068 13 0.9708 

comparison and correlation at outlet verified 

features (Chiang Mai basin, upper Nan basin, 

regulated & unregulated Pasak basin). 

CONCLUSON 

Public domain system model SWAT, has 

been executed on most water resource freeware 

in this decade, which become more and more 

standardize to technological contemporary. The 

advantages of SWAT model basely calibrate up to 

five hundred basins, could perform best results on 

these selected domains. Differences were 

permissible for both surface runoff and base flow 

fractions since investigated values are on large 

scale verification. The SWAT model has been 

elective appraised on both unregulated and 

regulated prototypes, which their issuances had 

confirmative best allied on large scale-mild slope 

potentiality and un-implied in both continual 

rainfall investigation and sufficient number of 

stations. Bondage correlation efficiency in large 

tropical basin of both unregulated steep slope and 

regulated mild slope basins were illustrated in 

Table 2, Figure 7 and Figure 8. Data analysis 

program development has been much efficiently 

investigated to operation, calibration and 

verification the model while digital elevation 

model was combined to become essential assistant 

on water resource decision supporting system. 


Research on geoinformatic public 

domain models has been consistently subsidized 

under Research and Development Unit on 

Geoinformatic Public Domain System / Research 

and Training Center on Resource Management and 

Geoinformatics, Faculty of Engineering, Kasetsart 

University. Together with GAME-Chaophraya

Figure 7 Surface flow comparisons and correlations. 


Figure 8 Correlation of flows, number of basin rain gauge and basin slope. 

research project, National Research Council of 

Thailand. Viscous bountiful diversity from The 

University of Tokyo, Tokyo, Japan directly 

concerned to project success. 


Arnold, J.G., W.Rosenthal, R.Srinivasan, 

K.W.King and R.H.Griggs. 1995. Swat–Soil 

and Water Assessment Tool : Draft Users 

Manual, USDA–ARS, Temple, Texas.110p. 

Bahram,S.1992. Implementation of a 

Distributed Hydrologic Model within 

Grass. USDA – ARS, Texas. 300 p. 

Chow, V.T. 1988. Applied Hydrology. Mc-Graw 

Hill, New York. 273 p. 

Craig, A.H. 1981. Soil-Vegetation Relations in 

The North Continental Highland Region of 

Thailand. A preliminary investigation of soil 

- vegetation correlation. Soil Survey Division, 

Department of Land Development, 

Bangkok.112 p.

272 

George, B. K. 1997. The GIS Book. 

Understanding the Value and Implementation 

of Geographic Information Systems, Word 

Press, New York.414 p. 

Ling, B. 1996. ARCSWAT. Users Manual, 

Department of Geography, State University, 

New York of Buffalo. 200p. 

Muttiah, M. and P.M.Allen.1999. Continental 

Scale Simulation of the Hydrologic Balance, 

J. American Water Res.Ass., 35(5): New 

York. 1200 p. 

Mauro, D.L., R.Srinivasan and G.A.Jeff. 1998. An 

Arc-View GIS Extension. Tools for 

Watershed Control of Point and Non-Point 

Sources, USDA–ARS, Temple, Texas. 80p. 

Michael, F.G.1998. GIS And Environmental 

Modeling. Fort Collins, GIS World Books, 

New York. 486 p. 

Rosenthal, W.D., R.Srinivasan and J.G.Arnold. 

1995. Alternative river management using 

a linked GIS-hydrology model. Trans. 

ASAE.38(3):7., Texas.110 p. 

Srinivasan,R., B.W.Byars and J.G.Arnold. 1995. 

Swat–Grass. users manual, USDA-ARS 

Temple, Texas. 150 p. 

Srinivasan,R., J.G.Arnold, W.Rosenthal and 

R.S.Muttiah. 1996. Hydrologic Modeling of 

Texas Gulf Basin Using GIS. USDA-ARS 

Temple, Texas. 535 p. 

Sloto, R.A. and M.Y.Crouse. 1996. HYSEP:A 

Computer Program for Stream-flow 

Hydrograph Separation and Analysis, U.S. 

Geological Ecological Survey Water- 

Resources Investigations Report 96- 4040, 

Texas. 46 p. 

Steven, T. 1998. Brush/Water Yield Feasibility 

Studies, Black Land Research and Extension 

Center, 720 E. Black Land Rd., Temple, Texas, 

80 p. 

Vathananukij,H.1998. River Engineering, 

Physics Center Co.Ltd., Bangkok. 120 p. 

Vathananukij,H. and K.Thanasiriyakul. 2002. 

Thailand Tributaries Relativity Assessment 

through Geoinformatic System and SWAT 

Modeling, pp.62-68 In Proceedings 

International Workshop on GAME-T and 


Hydro-meteorological Studies in Thailand 

and Southeast Asia, Chiang Rai. 

Vathananukij,H. 2003a. Geographic Information 

System/Geoinformatic System, Physics 

Center Co.Ltd., Bangkok. 151 p. 

Vathananukij,H. 2003 b. Public Domain 

Model (SWAT/GIS) Development on 

Nan River Basin, pp.15-23 In Proceedings 

International Symposium on the Climate 

System of Asian Monsoon and its 

Interaction with Society, Khon Khaen. 

Vathananukij,H. and T.Paewpisakul. 2003. Design 

Flood Application through Geoinformatic 

System and Autocad Land Development 

Desktop(LDT) Model, pp.1087-1093. 

In Proceedings International Fourth 

Regional Symposium on Infrastructure 

Development in Civil Engineering, 

Bangkok. 

Vathananukij,H. 2004a. Portentous Flood 

Exposition on Geoinformatic System, 

pp.25-28. In Proceedings International 

Symposium on Water Resource and Its 

Variability in Asia in the 21 st Century, 

Epochal(International Congress Center), 

Tsukuba, Ibaraki. 

Vathananukij,H. 2004b. Sustention Models for 

Water Resource Management, volume II 

pp.668–676. In Proceeding 2 nd APHW 

Conference Jointly with 1 st AOGS Annual 

Meeting, Suntec International Convention 

and Exhibition Center, Singapore. 

Vathananukij,H. 2005. Near Real Time MODIS/ 

TERRA on Hydro-informatics for Agriculture 

in Thailand, pp.10-22 In Proceedings 

International Seminar on Near Real Time 

Agriculture Activity Monitoring Using 

Multi-Temporal MODIS Earth 

Observation Satellite Data: Faculty of 

Engineering, Kasetsart University, Bangkok. 

Yan,Z. and F.Chris. 1998. Watershed 

Management Tool Using SWAT and Arc- 

Info. Center for Agricultural Resource and 

Environmental Systems (CARES), 

University of Missouri – Columbia, 95 p.

NOTICE TO CONTRIBUTORS 

The Kasetsart Journal (Natural Science) is published by Kasetsart University Research and 

Development Institute (KURDI) under the authorization of Kasetsart University to serve the interest of 

scientists and researchers. Research articles must be a report of the original work in any aspects of natural 

sciences. The Journal emphasizes and promotes the highest standards of ethics in the conduct of reporting 

research. Appropriateness is judged by the scientific approach to study the problems related to natural 

sciences. The governing operational factor is the general soundness and integrity of science being reported 

as judged by peer reviewers and/or the scientific editors. The Journal is published quarterly and is issued 

in March, June, September and December. 

GENERAL EDITORIAL POLICIES 

Authorship criteria and author responsibilities 

Kasetsart Journal is proud of its high quality research journals and is dedicated to maintaining 

its high level of professionalism. However, because of reports of fraud and plagiarism in the scientific 

literature, we remind authors of their obligations when submitting manuscripts for publication. 

Authorship criteria. Authorship is restricted to those who (1) have contributed substantially to one or 

more of the following aspects of the work (conception, planning, execution, writing, interpretation, and 

statistical analysis) and (2) are willing to assume public responsibility for the validity of the work. 

Exclusivity of work. The corresponding author must verify, on behalf of all authors (if more than one), 

that neither this manuscript nor one with substantially similar content has been published, accepted for 

publication, or is being considered for publication elsewhere. 

Copyright 

Copyright to published manuscripts becomes the sole property of the KURDI. In instances where 

the work cannot be copyrighted (works authored solely by government employees as part of their 

employment duties), this requirement is waived. 

Reproduction of all or a portion of a Kasetsart Journal article by anyone, excluding authors, is 

prohibited, unless permission is received from KURDI. Authors have the right to reproduce extracts from 

the paper with proper acknowledgment and retain the right to any patentable subject material that might be 

contained in the article. 

Disclaimer 

Opinions expressed in articles published in this journal are those of the author(s) and do not 

necessarily represent opinions of the KURDI. KURDI does not guarantee the appropriateness for any 

purpose, of any method, product, process, or device described or identified in an article. Trade names, when 

used, are only for identification and do not constitute endorsement by KURDI. 

Criteria for manuscript acceptance 

Manuscript acceptability is based primarily on appropriateness and importance of the topic; 

clarity of objectives; originality; appropriateness of the experimental design, methods and statistical 

analysis; depth of the investigation; substance of the results; thoroughness with which the results are

discussed; and appropriateness of the conclusions. 

Reprints 

Following acceptance of a paper and prior to publication, the author will be given appropriate 

number of reprints free of charge. 

MANUSCRIPT REQUIREMENTS 

Unless otherwise stipulated, the style and format of manuscripts submitted to Kasetsart Journal 

(Natural Science) should follow the “Scientific Style and Format: The CBE Manual for Authors, Editors 

and Publishers” 1994, 6th ed. (Cambridge University Press, New York). For convenience, refer to articles 

in the latest issue of the journal or for more details, contact the Kasetsart Journal editorial office. 

Use the English language (American spelling and usage) and the SI system (Système International 

d’Unitès, often referred to as “International Units”) for measurements and units. 

Manuscripts preparation 

Manuscripts on original research should include the following elements. 

Title page, as p. 1. Include: 

• Full title (be concise) 

• Name(s) of author(s) and author affiliation(s) with complete address(es) 

• Contact information for the corresponding author, including full name, complete mailing address, 

telephone, fax, and electronic mailing address. 

• Previous address of author(s) if research was conducted at a place different from current affiliation. 

Abstract, as p. 2. Include: 

• An abstract not exceeding 200 words; all acronyms and abbreviations defined; no references cited. 

State what, where and how it was done, major results. 

• Five key words. 

Introduction. In two pages or less, review pertinent work, cite key references, explain importance of 

the research, and state objectives of your work. 

Materials and Methods. Provide sufficient detail so work can be repeated. Describe new methods in 

detail; accepted methods briefly with references. Use subheadings as needed for clarity. 

Use of trade names. Trade names are to be avoided in defining products whenever possible. If 

naming a product trade name cannot be avoided, the trade names of other like products also should be 

mentioned, and first use should be accompanied by the superscript symbol TM followed in parentheses by 

the owner’s name. If a product trade name is used, it is imperative that the product be described in sufficient 

details so the nature of the product will be understood by professionally trained readers. Do not use trade 

names in titles. 

Use of abbreviations and acronyms. At first text use, define in parentheses. Do not use 

abbreviations and acronyms in titles.

Statistical analysis. If variation within a treatment (coefficient of variation, the standard 

deviation divided by the mean) is small (less than 10%) and difference among treatment means is large 

(greater than 3 standard deviations), it is not necessary to conduct a statistical analysis. If the data do not 

meet these criteria, appropriate statistical analysis must be conducted and reported. 

Results. Present results concisely using figures and tables as needed. Do not present the same 

information in figures and tables. 

Discussion. Discuss principles and relationship, point out exception. Show agreement with published 

research work. The significance of the work or conductions should be presented in the end of discussion. 

Conclusion. State conclusions (not a summary) briefly. 

Literature Cited. List only those references cited in the text. Required format of references is described 

below. 

Acknowledgments. List sources of financial or material support and the names of individuals whose 

contributions were significant but not deserving of authorship. Acknowledgment of an employer’s 

permission to publish will not be printed. 

Appendix. This section is rarely needed in a research paper but can be added if deemed necessary (e.g., 

complicated calculations, detailed nomenclature). 

Tables. Number each table with Arabic numerals. Place a descriptive caption at the top of each table. 

Print one table per page. Columns and their headings are usually (but not always) used to display the 

dependent variable(s) being presented in the table. Foot-notes should be identified by lower case letters 

appearing as superscripts in the body of the table and preceding the footnote below the table. 

Figures (graphs, charts, line drawings, photographs). Supply one illustration per page with the figure 

number indicated under the illustration. Figure captions should be double-spaced and listed consecutively 

on page(s) separate from figures: use Arabic numerals. Include one original set of illustrations with an 

original manuscript, marked, “ORIGINAL.” Illustrations for other copies of the manuscript may be 

photocopies, provided they are clear. Exceptions are photomicrographs, gel electrophoretic patterns, etc.; 

furnish four glossy photographs for each of these. 

Authors are responsible for obtaining permission to reproduce previously copyrighted illustrations. 

Proof or certification of permission to reproduce is required. 

Lettering, data lines, and symbols must be sufficiently large so as to be clearly visible when the figure 

is reduced to a size commonly used in the journal. When a color presentation is deemed necessary, please 

note this in the cover letter of the submission.

LITERATURE CITED FORMAT 

Manuscripts should follow the name-year reference format of the Council of Science Editors 

(formerly Council of Biology Editors). Cite only necessary publications. Primary rather than secondary 

references should be cited, when possible. It is acceptable to cite work that is “in press” (i.e., accepted but 

not yet published) with the pertinent year and volume number of the reference. Work that is “submitted” 

but not yet accepted should not be cited. 

In text. Cite publications in text with author name and year. Three or more authors use “et al.” 

In parenthetical citations, separate author and year with a comma. Use suffixes a, b and c to separate 

publications in same year by the same author. Semi-colons separate citations of different authors. Cite two 

or more publications of different authors in chronological sequence, from earliest to latest. For example: 

• The starch granules are normally elongated in the milk stage (Brown, 1956). 

• Smith et al. (1994) reported growth on vinasse. 

• ...and work (Dawson and Briggs, 1984,1987) has shown that . . . 

• ...and work (Dawson, 1984; Briggs, 1999) has shown that . . . 

In Literature Cited section. List only those literature cited in the text. References should be 

listed alphabetically by the first author’s last name. Single author precedes same author with co-authors. 

Type references flush left as separate paragraphs. Do not indent manually. Write the names of book and 

journal in bold letters. Let the text wrap with first line hanging indented. Use the following format. 

• Journal articles: Author(s). Year. Article title. Journal title, volume number: inclusive pages. 

Example: Citation in text: (Smith et al., 1999) 

Smith J.B., L.B. Jones and K.R. Rackly. 1999. Maillard browning in apples. J. Food Sci. 64 : 512-518. 

• Books: Author(s) or editor(s). Year. Title. Publisher name. Place of publication. Number of pages. 

Example: Citation in text: (Spally and Morgan, 1989) 

Spally M.R. and S.S. Morgan. 1989. Methods of Food Analysis. 2nd ed. Elsevier. New York. 682 p. 

• Chapter: Author(s) of the chapter. Year. Title of the chapter, pages of the chapter. In author(s) or editor(s). 

Title of the book. Publisher. Place of publication. 

Example: Citation in text: (Rich and Ellis, 1998) 

Rich R.Q. and M.T. Ellis 1998. Lipid oxidation in fish muscle, pp. 832-855. In J.J. Moody and W. Lasky, 

(eds.). Lipid Oxidation in Food. 6th ed. Pergamon. New York. 

For journal abbreviations and other examples of reference formats please refer to articles in a 

2000 issue of this journal or contact the editorial office at KURDI. 

EDITORIAL REVIEW AND PROCESSING 

Peer Review. All submitted manuscripts are screened by the Scientific Editor for importance, 

substance, appropriateness for the journal, general scientific quality, and amount of new information 

provided. Those failing to meet current standards are rejected without further review. Those meetings these 

initial standards are sent to expert referees for peer review. Referees identities are not disclosed to the author. 

Author identities are also not disclosed to the referees. Referee comments are reviewed by an Associate 

Editor and he/she, often after allowing the author to make changes in response to the referee’s comments,

advises the Scientific Editor to either accept or reject the manuscript. The Scientific Editor informs the 

author of the final decision. The review process ordinarily is completed within 3 months. If the process is 

delayed beyond that point, authors will be notified. 

Rejected manuscripts. Rejected manuscripts may, in some instances, be appropriate for 

publication in other journals. Rejected manuscripts including original illustrations and photographs will be 

returned to authors. 

Accepted manuscripts. The author(s) will be asked to review a copyedited pageproof. The 

author(s) is responsible for all statements appearing in the galley proofs. The author will be informed of the 

estimated date of publication. 

Inquiries regarding status of the manuscript. Direct inquiries to: Orawan Wongwanich, 

Manager, Kasetsart University Research and Development Institute (KURDI), Kasetsart University, 50 

Pahonyothin Road, Chatuchak, Bangkok 10900, Tel. 66-2-579-0032, 579-5548, 561-1474, Fax. 66-2- 

561-1474, E-mail: rdi@nontrLku.ac.th. 

SUBMISSION MANUSCRIPT 

Submit the following items. 

Cover letter. Identify the corresponding author and provide his/her full name, address, numbers 

for telephone and fax, and e-mail address. 

Manuscript. Double-space all components of the manuscript except tables. In 12 point Times 

or Times New Roman.Type on one side of A4 paper. Use one inch margins. Number all pages. 

Send an original manuscript (with original figures, marked “ORIGINAL”) and 3 photocopies. 

Hard copies are required. Staple each manuscript copy, including the original manuscript in the upper left 

corner. Include no paper clips or binders. 

Disk. Include an IBM-formatted, 3-1/2″ disk, containing the manuscript in Microsoft Word 

(version 97 preferred). 

MAIL MANUSCRIPT TO: 

Orawan Wongwanich, Manager, 

Kasetsart University Research and Development Institute (KURDI) 

Kasetsart University, 

50 Pahonyothin Road, Chatuchak, 

Bangkok 10900, Thailand 

Tel. 66-2-579-0032, 579-5548, 561-1474 

Fax. 66-2-561-1474 

E-mail: rdi@nontri.ku.ac.th

PRE-SUBMISSION CHECKLIST 

(see “Instructions for Authors” for additional information.) 

Cover letter & form 

❑ Full contact information for the corresponding author (full name, address, phone, 

fax, e-mail). 

Manuscript 

❑ Original manuscript (with original figures, marked ORIGINAL) 

❑ Three photocopies 

❑ Page numbering 

❑ Each manuscript stapled in upper left corner (no paper clips or binders) 

❑ Abstract of less than 200 words 

❑ 5 key words 

❑ Figure captions listed consecutively on a page separate from figures 

DISK 

❑ 3-1/2 in IBM-formatted disk with full manuscript Microsoft Word (version 97 preferred) file. 

(submission of the title page and abstract by electronic means is an acceptable alternative)

Kasetsart Journal 

Application for Membership 

Name : __________________________________________________________________ 

Title : __________________________________________________________________ 

Home Address : ____________________________________________________________ 

Tel :_________________________________ Fax : __________________________ 

Office Address : ____________________________________________________________ 

Tel :_________________________________ Fax : __________________________ 

E-mail Address : ____________________________________________________________ 

Type of application : ❑ New Membership 

❑ Membership Extension 

Membership type : ❑ Regular (non – government) Membership 

Submission information : 

❑ Student Membership 

Student membership (Natural Science) 

❑ Application and 60 bahts dues payment 

❑ Letter of Recommendation from major advisor or the copy of student ID 

Student membership (Social Science) 


❑ Letter of Recommendation from major advisor or the copy of student ID 

Regular membership (Natural Science) 


Regular membership (Social Science) 

❑ Application and 80 bahts dues payment

Send application materials to 

Payment information : 

Manager of KU Journal 

KURDI Kasetsart University 

50 Paholyothin, Bangkhen Bangkok 10900 

❑ Cash 

❑ Send money order made payable to KURDI , Kasetsart Post Office 

Preferred address for mail 

❑ Home 

❑ Office 

Do Not Write In This Space 

For Office Use Only 

Membership Number ___________________ 

Receipt Number _______________________ Receipt Date ______________________ 

______________________________ Finance Officer 

( ____________________________ ) 

Date ______________________________

Letter of Recommendation 

This is to certify that __________________________________________________ 

is currently the _____________________ (first -, second -, third -, fourth – year, graduate) 

student majoring ____________________ , Faculty of ____________________________ , 

University. 

( ____________________________________ ) 

Academic Title _________________________ 

Major Advisor 

Date _________________________________

Text and Journal Publication Co., Ltd. 

158/3 Soi Viphawadi Rangsit 5, Viphawadi Rangsit Road, Ladyao, Chatuchak, Bangkok 10900 

tel. 0-2617-8611-5 fax. 0-2617-8616

kasetsart journal natural science

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?