Grain Legumes and Green Manures for Soil Fertility in ... - cimmyt

Materials and Methods
The study used farmer participatory approaches,
complemented with laboratory-based analyses of
soils and plant materials. At least two participatory
rural appraisal (PRA) workshops were held at each
study site to discuss broader issues of soil fertility
management and local knowledge of leguminous
plants. Farmer involvement ranged from identification
of the legumes, and their niches, to seed collection.
Members of the local community leadership
that included councilors, headmen and resident national
extension officers organized and participated
in transect walks during the initial legume identification
exercise.
Transect walks and legume identification using
the Gwezu smell technique
Tr~ect walks were conducted in all study sites.
Based on the physical slope, farmers identified three
main field positions, namely, topland, midslope and
the relatively moist bottomland positions. During
the transect walk, particular attention was paid to
cropping patterns (including crop types), weed
status of the fields, and occurrence of naturally
growing herbaceous legumes. General discussions
ensued during the course of the walks and details of
cropping history and predominant weed species
were specifically discussed with farmers whose
fields were surveyed. Farmers were generally able
to distinguish legumes from non-leguminous plants
by considering mainly fruit morphology and likening
them to traditionally grown leguminous crops
such as groundnut (Arachis hypogaea), common bean
(Phaseolus vulgaris) and cowpea (Vigna unguiculata).
Because identification was done when most of the
species were not yet fruiting, applicability of this
approach was limited. To aid this process, the research
scientists then came up with an identification
approach based on the human sense of smell, hereinafter
called the Gwezu smell technique. The researchers
discovered that freshly harvested roots of
all the identified legumes invariably had ,a distinct
smell characteristic of an immature groundnut pod.
An immature groundnut pod is known as Gwezu in
Shona (Karanga dialect), a Zimbabwean vernacular
language. Plants were also uprooted, and presence
of root nodules was considered indicative of a legume,
taking care to differentiate true root nodules
from the galls caused by root-knot nematodes. The
field-identified legumes were then taken to the National
Herbarium laboratory of the Zimbabwe Ministry
of Lands~ Agriculture and Rural Resettlement,
for botanic identification.
Measuring species diversity and abundance
Transect walks and PRA group discussions resulted
in three possible scenarios for legume sampling to
determine the diversity and abundance of the legume
species: Scenario I - natural grazing areas that
have not been cultivated for more than five years;
Scenario II - fields that had not been cropped in the
current season (first season of fallowing); and Scenario
III - cultivated fields in which only the first
weeding had been done. After further consultation
with farmers, it was decided that measurement of
species abundance be focused on the latter two scenarios
since grazing by livestock would affect measurements
under natural grazing areas. Consequently,
the emphasis on Scenario I was only on determining
species diversity. Individual plant samples
were collected by farmers, field assistants and
researchers enclosed in polythene bags, and put in
cooler boxes for transportation to the National Herbarium
for identification. For Scenario II, only those
fields that were free from livestock disturbance during
the cropping season were sampled. The only
exception to the sampling protocol was at Mr Zindoma's
farm where a maize field abandoned soon
after crop emergence due to lack of fertilizer inputs
was additionally included.
Sampling and analyses of plants and soils
Sampling for Scenarios II and III was done by using
a network of 4 m x 4 m grids that were made out of
metal pegs and twine. The grid network was spread
over the desired field area and four replicate grids
randomly selected for sampling. For each replicate
sampling grid, all legume plants belonging to the
same species were uprooted, checked for nodulation
and put in a khaki sampling bag after cutting
off the roots from just above the soil line. Nonleguminous
plants were collectively sampled from 2
randomly located grids of 0.5 x 0.5 m 2 drawn from
within the 4 x 4 m 2 grid. In each agro-region the
process was repeated at each of the selected 10 farm
sites where fields meeting desired criteria were
found. All plant samples were oven-dried to constant
weight at 60°C and then measured for dry
mass. The dried samples were then ground in a
Wiley Mill to pass through a 1 mm sieve, Determination
of N, P and K concentrations was then done
using the methods given by Anderson and Ingram
(1993).
Soils were sampled from the respective field sites by
collecting 15 sub-samples from the 0 - 20 cm depth
per field site using a spade. The sub-samples were
mixed thoroughly in a clean polystyrene bucket, after
which a 1 kg composite sample was withdrawn
and put into a polythene bag for laboratory analysis.
The soils were analyzed for texture, pH, organic
C and plant available P according to methods by
Anderson and Ingram (1993).
Grain legumes and Green Manures for Soil Fertility in Southern Africa
69