Demonstrations of an interdisciplinary farming systems approach to planning adaptive agricultural research programmes

Material Information

Demonstrations of an interdisciplinary farming systems approach to planning adaptive agricultural research programmes
Series Title:
Report Cimmyt Eastern African economics programme
Place of Publication:
Publication Date:
Physical Description:
41 p. : ; .. cm.


non-fiction ( marcgt )
Spatial Coverage:

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact Digital Services ( with any additional information they can provide.
Resource Identifier:
67368359 ( OCLC )
000316843 ( CLC )

Full Text


46 c




REPORT No. 1. APRIL, 1977


Research Division, Ministry of Agriculture, Government of Kenya and
Department of Economics, Egerton College, Kenya in association with
International Maize and Wheat Improvement Centre.

Cimmyt Eastern African Economics Programme.
Cimmyt, P.O. Box 25171, Nairobi, Kenya, Telephone 48480

lL~tl/tb QLL/N S ~ IJ


The Research Division, Ministry of Agriculture, GOK and the Department
of Economics, Egerton College cooperated with CIMMYT in the demonstra-
tion of an interdisciplinary farming system based approach to planning
adaptive agricultural research in part of Siaya District, Nyanza
Province, Kenya. Chapter 1 of the report describes the objectives of
such demonstrations and the methods followed. Chapter 2 describes the
circumstances of people farming in the part of Siaya District with less
than 1500 mm rainfall per year.


Reliable family food supplies are a priority for Siaya farmers. Cha-
pter 3 describes their management tactics to provide the family with
a supply of food within the constraints of limited available land,
labour scarcity, particularly for weeding, and the low levels of cash-
they have available to spend on hired resources and purchased inputs.
A good farmer in Siaya, one respected by his fellow farmers, is a man
who produces enough maize regularly every year to feed his family on
their preferred food until the new harvest. He also produces enough
sorghum to allow his wife to brew beer throughout the year,.and a
surplus of crops to raise enough cash to meet his family's household
requirements and add to his herd of livestock from time to time. He
will have his own oxteam and plough.

The farming system operated by Siaya farmers is under considerable
pressure from increasing population. It is no longer possible to
maintain soil fertility by shifting cultivation, land is too scarce.
The arable area has grown at the expense of grazing and this increase
in the arable area is jeopardising the livestock enterprise and break-
ing up the interdependence of arable and livestock husbandry in the
system. A key decision (p.27) on the development of Siaya farming
will be the future of the livestock enterprise. On the cropping side
the relationship between sorghum, maize and cassava, all basic starch
foods, will be critical. Both the sorghums and cassava play an
insurance role in the farming system, underwriting the preferred
starch staple maize. The sorghums have a dual role; as insurance and
for brewing local beer. The real insurance sorghums are good stores,
they are hard and bitter and less prone to storage pests. They are
also long term and their establishment and maintenance require labour
early in the season to give them time to mature within the Siaya rain-
fall regime. These long term insurance sorghums take priority over
other crops at what is the optimal planting time in the season.

A high level of insurance cropping reflects the highly unreliable rain-
fall regime of this drier part of Siaya District. The start and finish
of the main rains is very variable. Flexibility in management is an
important asset in combating this type of uncertainty and the risks of
food crop failures it creates. For Siaya farmers the availability of
shorter term sorghum and maize varieties, particularlily varieties
with good storage characteristics, would considerably enhance their
flexibility of management in the face of uncertainties of rainfall.
Chapter 4 of the report evaluates the implications for a maize research
programme aimed at benefitting Siaya farmers.


Variety A shorter term maize variety (95 days in the drier parts of
the District) will give management flexibility to farmers with the
following advantages:
(1) The probability of being able to finish the crop in an increa-
sed number of years, even with a relatively late start to the rains
or an early finish.
(2) Increased probability of a successful short rains crop.
(3) The increased possibility of taking the two crops, long and
short rains, on the same land, important with increasing population
* (4) Earlier food supplies in seasons following poor harvests when
food stocks are low, avoiding the need to purchase on local markets
when prices are very high.
(5) Farmers growing a surplus will be able to sell on the local
market early in the new season when prices are high. Increased
use of the short-term varieties will gradually remove the seasona-
lity in food prices benefitting all in the area.

The use of purchased inputs. Forty six percent of the farmers sampled
were spending casl on farm inputs, their estimated average outlay was
N 260.00. Most money was spent on the hire of labour and oxen. Fifty
four percent of farmers interviewed reported no cash spending on farm
requisites. Nor does the Siaya farmers situation lend itself to a
credit scheme for input purchase.
(1) Most crop sales are surplus to food requirements and are made
in the local market. The cost of credit administration, particula-
rily in controlling repayments in this type of marketing situation,
is prohibitive for a programme aiming to reach the majority of
Siaya farmers.
(2) The risk ceiling of the average Siaya farmer, with a low level
of cash income, would discourage borrowings beyond an estimated
two to three hundred shillings.

The major cash inputs in any package are for fertilisers. Research
for Siaya should concentrate on efficient use of the animal manure
available throughout the area and already being used by 62% of the
farmers interviewed. Use of a shorter term variety is consistent with
relatively low levels of manuring. In the longer term artificial
will be needed, cattle numbers will not provide sufficient manure to
maintain fertility. Recommendations should initially be at low levels,
supplementary to the use of animal manure.

Constraints on changes in maize husbandry.
Time of planting and the frequency and timing of weeding are usual
features of improved crop husbandry which present problems for Siaya
farmers who operate under labour constraints during the period of crop
establishment and weeding. These constraints are likely to become
increasingly severe as the arable/livestock clash intensifies the need
for hand cultivation. Operating under these constraints farmers
spread the demand for labour by planting over a three month period.
Most of the cash presently spent on farms in Siaya is on oxen or labour
hire to alleviate these labour constraints. Asking farmers to plant


and weed at a particular time aggravates their situation. It increa-
*ses the peaks of labour required and implies they must increase cash
outlay on oxen or labour to boost family labour supply. Any such
increase in cash outlay is directly competitive with increases also
being urged for the purchase of intensifying inputs and'aggravates
further the farmers' already severe capital allocation problem. At
the same time, under the rainfall regime in Siaya, single plantings
at optimal times create increased risks, of major losses, and a fai-
lure in family food supplies. These labour constraints on Siaya
farmers' ability to plant and weed at a specific time suggest three
guidelines for research effort on maize.

(1) Varieties should be screened for insensitivity to time of
planting effect. Hand in hand with the shorter term variety such
selections would further increase the farmers' flexibility of
management under uncertain rainfall conditions.

(2) Varieties should be screened for rate of early'growth and
weed suppression.

(3) Experimenters should accept two weedings in the first two
months after germination as the maximum feasible level and work
within this level.

The final section of the report (p 37) details a five year maize
research programme designed to give profitable recommendations which
are acceptable to Siaya farmers because they are relevant to their
priorities and can be implemented within their resource constraints.



1.1. BACKGROUND ------------------------------------- 1
ORIGIN AND OBJECTIVES ------------------------------ 1
RESEARCH PLANNING --------------------------------- 2

2.1. REASONS FOR SELECTING SIAYA ----------------------- 3
2.2. ZONING SIAYA DISTRICT ------------------------------ 3
SOIL IN THE SIAYA ZONE ----------------------------- 4
FARMING IN THE SIAYA ZONE ------------------------ 9
OPERATION IN THE SIAYA ZONE ------------------------ 16

FARMING IN THE SIAYA ZONE -------------------------- 18
3.3. THE FARMING SYSTEM ---------------------------------- 22
FARMING SYSTEM IN THE SIAYA ZONE ------------------- 25
IN THE SIAYA ZONE ----------------------------------- 27
FOR CROP RESEARCH ----------------------------------- 28


4.1. INTRODUCTION ---------------------------------------- 29
4.2. VARIETY ---------------------------------------------- 30
4.5. THE USE OF PURCHASED INPUTS ----------------------- 33
4.4. METHODS OF PLANTING --------------------------------- 34
4.5. TIME OF PLANTING ------------------------------------- 35
4.6. WEEDING ---------------------------------------------- 36
1500 mm ANNUAL RAINFALL ----------------------------- 37

5.1. INTRODUCTORY ----------------------------------------- 39
5.3. MODIFICATIONS OF METHODOLOGY ----------------------- 40




By a Memorandum.of Agreement signed in November, 1975 Kenya Government
(GOK) agreed that the International Maize and Wheat Improvement Centre
(CIMMYT) should establish itself in.Kenya as a base for a Regional
Programme in Economics for Eastern Africa. The Ministry of Agricul-
ture, GOK, expressed interest and it is fitting that the first project
to have been completed under the programme is in association with the
Research Division of the Ministry and the Department of Economics,
Egerton College.
This report presents the results of this first project, undertaken in
Siaya District, Nyanza Province, Western Kenya. Fieldwork was comple-
ted there between early August and mid September, 1976.


A good deal has been heard of the need for improved technology which
is 'appropriate' or 'relevant' to the situation of the farmer. This
need is seen as particularly critical in the case of small farmers who
have their improved technology selected for them, usually by Govern-
ment. A huge edifice of training, extension effort, credit schemes,
often including transport and processing as well as infrastructural
investments, may be built around selected technology as the core of
a development programme. If the technology does not fit farmers'
priorities and circumstances it will be unacceptable to them and not
only the-research but also the development effort will have been

In 1972 CIMMYT began an economics programme to study whether eight
sets of improved technologies, four each in maize and wheat, were be-
ing absorbed by farmers in eight different countries. Fieldwork on
these studies was completed in mid-1975 and reports are now becoming
available. These studies reaffirm the need for technology to fit the
farmers circumstances. In the cases where adoption did not occur the-
re is evidence that the technology did not fit farmers situations.
CIVMMYT perceived that criticism of the research programme orientation
after the event was too late to be useful, research and development
funds having already been expended. CIMMYT biologists and economists
discussed the possibility of the interdisciplinary planning of resea-
rch programmes. The procedures derived have been and are being tested
in small studies in four countries. CIMMYT has felt sufficiently con-
fident of the procedures and convinced of the importance of more rele-
vant technological packages to focus the Eastern African Economics
programme on this work.

CIMMYT is urging a cooperative effort between biologists, as experts
in crop potential and crop production techniques and farm economists,
as experts in farmers' priorities and circumstances, in the planning
of adaptive research. Wherever research administrators manifest inte-
rest this programme demonstrates a set of procedures which allow such
a cooperative research planning effort. The project in Siaya, in
-association with the Research Division, Ministry of Agriculture GOK
and the Departnent of Economics, Egerton College, has been the progra-
mmes first such demonstration.


The set of procedures used has been derived from experience in CIMMYT
test projects and from consultation between biologists and economists
at CIMMYT. It is anticipated that procedures will need modifying to
the situation within the Kenya National Research Programmes. Such
modification will be possible on the basis of experience gained in
demonstration projects. It is emphasised that although CIMMYT's main
interest is in helping research administrators improve the effective-
ness of adaptive research in maize and wheat, the crops of CIMMYT's
mandate, the procedures can be utilised to structure research work on
all or any crops grown in the area studied. The approach embodied in
the procedures is essentially system oriented. For example in the
case of Siaya implications emerge.for research efforts on sorghum, ca-
ssava, and cotton as well as for maize.

Comments on possible improvements in the sequence with reference to
the organisation of agricultural research in Kenya are contained in
the final section of the report.


1. Definition of a target population with maize as an important
crop; the farmers of Siaya District.

2. A Pre-survey of the defined area by CIMMYT and Egerton eco-
nomists occupied six days and had four objectives.

(a) Zoning the Siaya area on the basis of locally identi-
fied variation in rainfall quantity and reliability.

(b) Discussion with local biological scientists at Western
Agricultural Research Station, Kakamega, on the problems of growing
maize under the climate, soil and pest conditions of Siaya.

(c) Description of the local farming system in the identi-
fied zone in Siaya to evaluate farmers' objectives, priorities and
resource constraints and the way their cropping pattern and management
practices reflect these.

(d) Establishing a framework for a survey of farmers with
the cooperation of Administration officials in five selected sub-
locations of Siaya and the junior field staff of the Ministry of Agri--

3. Designing a farmer questionnaire on the basis of the descri-
ption obtained of Siaya farming and of the problems of maize product-
ion highlighted by the staff at Kakamega.

4. A farmer survey, using eight Egerton students as enumerators
covered 150 farmers, thirty in each of five selected sub-locations,
in a fourteen day period.

5. Collection of general information including marketing,
prices, input distribution and meteorology by an Egerton economist
while supervising the farm survey work.

6. Tabulation by Egerton students, and analysis of data colle-
cted in the survey to verify pre-survey findings, taking some forty

7. Discussion of the findings of the pre-survey and survey with
senior crop scientists in Kenya isolate the implications for:
(a) the orientation of adaptive research for the area.
(b) breeding work for maize.
(c) agronomy work for maize.
8. Compiling a draft report.

This full sequence has been followed through in a five month period,
August to December 1976. It demonstrates the speed with which on
farm information can be brought to bear in designing a research prog-
ramme for a particular area, or in re-orienting the thrust of an exi-
sting programme. It can be a rapid and fairly cheap process. Should
a decision be taken at an annual research meeting to re-organise par-
ticular programmes, an inter-disciplinary planning approach, following
these procedures, would have detailed proposals ready for the follow-
ing years meeting. Most importantly such proposals could be based on
an organised study of the priorities and needs of the target farmer

It appears that the data from Karapul sub-location adjacent to the
town of Siaya has been.unduly influenced by its proximity to the town.
It seems likely that the rapid growth of Siaya, as a District town,
has distorted the local situation; Cultivated area per farm found in
the survey was only .95 ha, less than half the area cultivated per
farm in any of the other four sub-locations. The data presented on
the farmers' situation in Siaya is from the 120 farms interviewed in
the four sub-locations of Omia Malo and Nyawita in Bondo Division,
Kambare sub-location in Yala Division and Simenya sub-location in
Ukwala Division.
Siaya was selected as an area for the first demonstration of the app-
roach for three main reasons:
1. A large, dense, farmer population with maize as their major
and preferred starch staple food and as a locally important cash crop.
2. The ecological conditions of most of the District differ
sufficiently from the high and medium.potential areas to query the
suitability of the 500 and 600 series of hybrids for Siaya farmers.
3. Studies of Gerhart (1975) and Hesselmark (1974, 1976) had
shown low levels of adoption among the Siaya farmer population despite
some 10 years exposure-to a technological package based on available
hybrid varieties.
The main aim, reflected in 2 and 3 above, was to select an area in
which the approach would have an opportunity to prove itself. Clear-
ly, a demonstration in a high potential area, with well adapted varie-
ties and a high level of adoption among local farmers would give less
scope for the approach to demonstrate its potential.

Zoning within Siaya District was not a complex procedure. The hetero-
geneous factor is rainfall which falls away from the North and North
East towards the South of the District. The area around Kakamega in
the highlands to the North East of the District receive up to 2000 mm

annually and has 11 humid montns (Jaetzold 1974) in which crop growth
could be started. Moving South and West the rainfall drops steeply
and Yala in North East Siaya, is on the 1500 mm isohyet. At about
the 1500 mm isohyet the number of humid months fall to 10. More
importantly, of these 10, 4 months would be unable to reliable support
crops at water demanding stages of growth so two clear seasons are
distinguished. This bimodality continues to South and West to some-
where north of Bondo town itself. Here the August November rainfall
peak falls away and, on grounds of crop potential, the short season is
lost. Nevertheless in the Bondo area, as the survey results show,
most farmers attempt a short rains planting. The zone identified for
study is bounded by the line where the rainfall pattern breaks into
two clear seasons. It is assumed as the 1500 mm isohyet.

Although crop management problems intensify from North to South over
the zone, essentially similar problems are involved. Within the area
of the District receiving less than 1500 mm on average the potential
moves from good long and short rains crops in most years in the North,
to variable long rains crops and rarely successful short rains crops
in the South. The economic and social circumstances are so similar
through this zone that it was considered justifiable to examine it as
an entity despite the very marked change in rainfall reliability. The
zone covers South Gem location, Yala Division, the parts of Ukwala
Division south of the Yala-Ukwala-Bunyala road, Boro Division and most
of Bondo Division with the exception of the extreme southern and west-
ern areas bordering the lake.

1. Rainfall. There are no East African Meteorological Grade 2
stations in Siaya District. Use has been made of records held at
Divisional level in the Ministry of Agriculture. Table 1 shows mon-
thly averages of pooled data for sites and years for Bondo (939 mm
annually), Boro(1225 mm annually) and Ukwala (1472 mm annually).

T e 1 Available observations on monthly average
rainfall by Division. (mm)

Division Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Bondo 42 59 89 175 139 44 55 75 68 69 88 36
Boro 47 61 156 180 147 102 89 81 100 122 100 30
Ukwala 53 80 111 262 202 113 80 80 139 132 179 33

The table demonstrates a marked similarity in the rainfall pattern of
the three Divisions but a marked disparity in levels. Of particular
interest are the differences in amount and distribution in the periods
February June and August November,the long and short rain seasons.
Rainfall levels fall moving South across the District, away from
higher rainfall areas in the North West and North East. Reliability
of rain during the cropping season, closely correlated with reliabi-
lity of production levels, is of vital importance to small farmers
dependent on the farm for staple food supplies. Table 2 presents very
crude measures of the reliability of the rain, on a monthly basis over

the crop seasons, in the three Divisions studied in Siaya. The mea-
pures were derived by ranking pooled inter-year and inter-site data
to give an indication of the frequency of various levels of monthly
rain. The Table shows levels, in mm, below which rainfall can be
expected to drop in the number of years indicated for critical months
in the growing seasons.
Table 2. Crude measures of monthly rainfall reliability (mm)

DIVISION Out of Long rain .season Short rain season
5 years P M A M J A S 0 N

BONDO 1 6 30 98 87 25 33 35 28 25
2 30 83 135 119 40 48 44 50 53
3 53 100 179 140 47 63 61 65 86
4 110 135 233 176 56 101 102 119 155

BORO 1 15 60 130 83 38 40 55 80 46
2 20 105 174 116 47 68 84 101 64
3 43 140 205 145 100 73 140 135 138
4 95 210 236 193 163 100 145 158 195

OUKWAA AND 1 30 20 155 156 59 46 75 86 100
2 48 76 230-186 96 68 120 109 112
3 68 146 288 217 110 93 168 147 178
4 120 214 374 254 145 1 135 213 202 264

Note: See Brown and Cocheme (1969) page 71.

2. Soils: There is little information .available on the soils
of Siaya District. Observation within the District suggests that
soils are more .generally fertile in the North and East of the District,
becoming a more variable mixture of sands with blacker, heavier soils
in the bottom lands in the South and West. There is more selective
use of lands for cropping in the South .and West, hardpan areas and the
lower lying black soil areas are left uncultivated. It is unlikely
that soils vary enough over the District to influence crop potential.
Similarly the small differences in the present cropping pattern across
the District are probably not attributable to soil variations.


1. Communications. The District is bounded along the East and
North side by the main, bitumenised road from Kisumu to Busia. There
is a .good network of all weather murram roads over the District join-
ing the many minor settlements and markets. The network of roads is
correlated with the density of the population and frequency of markets
and is less developed south of Bondo in the less densely populated
area of the Division.


2. Population. The District had a population of 383,188 at the
1969 census in an area of 2535 square kilometre. That is an average
density of 151 persons per square kilometre. Extremes of density were
from 27 persons per square kilometre in Kadenge sub location, Boro
Division, .to 498 persons per square kilometire in Mahera sub location,
Yala Division. However 60% of the sub locations in the District had
population densities between 130 and 260 persons per square kilometre.
An estimated 90% of the resident population have farming as their
major source of livelihood. Assuming population growth of the order of
2.5% per annum since 1969 and using the farmer survey data showing an
average of 7.4 persons per household, gives an estimate of some
57,000 farm families in the District with an average land area of 4.2
to 4.6 ha each and supporting over 420,000 people. Average population
density will be now be approaching 185 per square kilometre. The area
is wholly populated by the Luo people, the few exceptions being emplo-
yed persons in the towns and minor settlements of the District. Ori-
gins and social traditions are very homogenous through the District.
The identified zone contains an estimated 70% of the District popula-
tion; some 40,000 farm families.

3. Markets

(a) Structure. .The District has a network of local markets
which are very well patronised by farmers. The majority of transa-
ctions are of farmers selling produce and buying produce brought by
other farmers. All dealing is done through cash. Each minor settle-
ment has its market and market days, often two each week. 'Beyond
these transactions the main marketing channel for crops surplus to
local needs is the Siaya District Cooperative Union. Originally a
cotton buying organisation it now buys a range of products acting as
a purchasing agent for the Maize and Produce Board. The Cooperative
currently has Produce Buying Stores at nineteen of the local markets
throughout the District. With the exception of sugar, grown-mainly
in North and East Gem and sold mainly to local jaggery factories, and
cotton, crops sold to the cooperative are surpluses of the dominant
food crops in the area, maize and beans are particularly important.
There are 22 producer dairy groups in the District,sales are made loc-
ally, none of the milk is sold to K.C.C. meat is marketed in the minor
settlements and is slaughtered locally, animals sold by local farmers
are supplemented by the import of animals from other Districts.

(b) Value of marketed products. No statistics exist to show
the volume of trade in the local village markets. Most farmers
surveyed reported this as their main produce outlet. Table 3 presents
a summary of estimated sales values for produce moving through the
organised marketing channels for 1975. The value of sales through
channels which can be fairly reliably monitored amounts to S 200 per
farm family. Commercial, sales of sugar are restricted to farmers in
parts of Yala and Ukwala Divisions close to the centres of processing.
Similarly although 67% of farmers were cattle owners, only 29% of
owners, that is 19% of farmers reported selling cattle within 1976. At
the same time only 17% of all farmers reported milk sales, necessarily
these would also be stockowners. Taking out the value of sugar, meat
and milk from the value of crops marketed through channels which can
be observed it is clear that the majority of farmers only have contact
with their village market. Omitting the value of these three products
the average valued of cash sales through formal channels is only h 20
per. farm family. 20% of surveyed farmers reported no crop sales; 53%
of farmers reported maize and .43% sorghum sales, usually at the local
village market.

Estimates of the value of products marketed through'
Table 3. assessable outlets in Siaya District in 1975

Product Marketing Sales value
channel (s)

Cotton Siaya Cooperative Union -840,708
Milk Producer cooperative groups 330,670
Beef Individual sales 7,307,400
Maize (acting as agent 171,332
Beans Siay C.U(for Maize & Produce 112,740
Other crops (Board) 16,539
Sugar Cane Individuals to jaggery
factories 2,689,000
Sales value through assessable outlets SB 11,468,389

(c) Prices. Thus the main outlet for smallholders produce is
the local village market. The District Agricultural Office, Siaya
keeps price records for crops sold in local markets and these series,
collected at Divisional level, give a measure of recent trends as well
as comparison between areas and seasons. Seasonal variability is
marked and closely related to long rains and short rains harvests in
July/August and December respectively. Prices for selected crops and
years are presented on a monthly basis in Table 4 for Ukwala and Bondo

Selected Market prices (s/bag)
Bondo Divisions

in Ukwala and

Products .Maize Sorghum Cassava Beans

Ukwala (1974 58 56 35 97
Average to (1975 82 58 52 125
July (1976 82 77 54 113

3 year- (Pre IR harvest 82(2) 83(2) 55 147
averages (Post LR harvest 40(2) 35(2) 36 81

Bondo (1974 68 64 35 125
Average to (1975 80 64 34 168
July (1976 102 97 58 341
3 year (Pre T R harvest 91 91 52 253
averages (Post LR harvest 49 41 33 132

-Table 4.

The prices in the table have been ;elected to show; the trend over the
last three years, the very large variation between seasons and any
differences among prices in Bondo and Ukwala, only some 40 kilometres
apart. In general it can be said; prices have risen dramatically in
local markets over the last two years, most farmers need to buy as
consumers before the new harvest is ready,the length of the period
depending on the success of the previous years harvest, and finally
that most farmers are probably restricted to selling in their own
local market.

In Table 5 Maize and Produce Board 1976 prices are compared to local
market prices; the January July 1976 averages a repeat from Table
4 and the August October post harvest, 1975 prices.

S5 Maize and Produce Board 1976 buying prices and
Table local market prices (Ig/bag)

Products Maize Sorghum Cassava Beans

MPB buying price 59 39.80 11.25 97.65

.Jan-Jul'76 82 77 54 113
Aug-Oct'75 47 40 36 84

Bondo Jan-Jul'76 102 97 58 341
Aug-0ct'75 49 41 29 145

n.b. 1. For cassava MPB purchases specify 50Kg bag weights. The
Ministry refers to 90 Kg bag weights in some of the reports used as

2. Although the bean price for Mexican 142, the most marketed
variety in Siaya, is shown in the table, the MPB 1976 purchase price
is 201.30 per bag for other varieties.

It can be seen that the MPB prices are well pitched in relation to
local market prices for maize and sorghum in the post harvest period.
For the pre-harvest period it can be assumed that the market is demand
led by farmers who have used their home produced supplies.

4. Input distribution. There is a network for the distribution
of seeds and fertilizer throughout the District. Kenya Farmers Asso-
ciation (KFA) and Kenya Seed Company (KSC) cooperate in appointing
agents. There were 56 such agents appointed in over 40 market centres
across the District for 1975/76 season. It is difficult to assess the
efficiency with which the network operates during August when few in-
puts are required. Agents claimed that they could not get stocks on
credit and clearly the cash flow from the site of businesses they were
operating would not give cash to hold significant quantities of in-
puts, particularly fertilizer. On the other side, farmers had no
ready access to credit for the purchase of inputs at a time of the
season when farm cash flow was very poor. It is probable that farmers

would seek inputs on credit from stockists. Given a two way squeeze
on their own finances margins would have to be very high to attract
dealers into holding significant stocks.

Although a distribution network exists on paper it is doubtful whether
it operates effectively due to the limitations on cash flow of both
stockists and farmers. No central records were available with KFA to
show the history and present levels of input sales through agents in
the three relevant Divisions of Siaya District.


The resources Siaya farmers have at their disposal, both quantity and
quality, may form constraints on the types of change in farming tech-
nique they will be able to absorb. The situation of land, labour and
capital in farming in the zone and the way these resources are managed-
are described briefly.

1. Land

(a) Holding size was not measured in the course of the farmer survey.
Allowing 20 o waste and. non-agricultural land on the 2535 square kilo-
metres of area for the District gives 3.50 ha per farm family varying
between 2.50 ha for the more densely populated parts and 4.50 ha for
the less densely populated. Figures for the four sub-locations are
given in Table 6 with average cultivated areas in the 1976 long rains
season, based on farmers' and enumerators' estimates, included in the
Table. Many data are rough estimates, adequate for purposes of the

Table 6. Population, land availability and use survey areas

Sub location Mao Nyawita Simenya Kambare

Population 1969 census 4387 3749 1474 2221

-Population 1976 (at
2.5% per year) 5210 4460 1750 2630

Estimated farm population
at 90% 4690 4010 1580 2370

Family size (1976) 7.8 8.3 7.1 6.6

Estimated at 80%
usable land (Ha) 1480 3120 880 1520

Cultivated area per
family (Ha) 2.11 3.08 2.08 2.17

Total to Arable (ratio)

- ----- --- I -'----------^----~-^'-u------L -I_-_ ~ _~_





The data in the table demonstrate varying degrees of land pressure.
The most critical being in Omia Malo, Bondo Division (East), the least
critical in Nyawita, Bondo Division (West). In none of the areas how-
ever is the traditional fertility maintenance practice of shifting
cultivation any longer feasible, land cannot be rested more than one
year in two and for Omia Malo resting is possible only in one year in
three. Of the 122 farmers interviewed in the four sublocations 76
reported using manure on their fields. The great majority of these
were from the 82 cattle owners, though 9 farmers, that is 23% of non-
owners, reported manure use. Cattle owners averaged over 8 animals
each, though with wide variations between individuals. Assuming pro-
duction of one tone of manure per animal per year the animal population
of 688 head, among the 122 farmers, would allow an application of 10
tons of manure per hectare every four years on the 281 hectares they
cultivated, a modest, but useful level of manuring.

(b) ILand Adjudication has been completed in four of the five sub-
locations surveyed, Omia Malo being the exception. In adjudicated
areas the number of plots per household averaged 1.7, with 83% of
households having only one or two plots. By contrast, in Omia Malo
67% of farmers had three plots or more and stock owners had access to
communal grazing. While adjudication often has serious repercussions
on shifting cultivation as a fertility maintenance technique, in Siaya
population densities have already made the technique obsolete. More
immediately pressing is the need for farmers to manage their livestock
within their own landholding. Farmers interviewed in the survey re-
ported 24 cases of land renting to supplement their own holdings. In
17 cases (of which 12 were in Nyawita, the least densely populated sub
location) land was given or received free of charge, 5 cases were for
cash and two in kind. One farmers loaned land in return for the hire
of oxen on his own fields.

(c) The cropping pattern in the District is fairly homogeneous, alt-
hough there are localised specialities. Crops are grown on the long
rains; March to May, and the short rains; August to November. There
are major differences in the success rate of short rains cropping in
the higher rainfall areas in the North, and the lower rainfall area of
the South. Nevertheless attempts at short rains cropping are equally
intensive in the drier sub locations surveyed. Table 7 sets out the
cropping pattern on the surveyed farms in the 1976 long rains, and
information on the proportions of farms which planted the main crops
in the 1975 short rains and intended to plant in the 1976 short rains.
Intercropped areas are counted twice and cannot be added across the
table. There was extensive intercropping of maize and sorghum, but it
is noteworthy that significant areas of these basic starch staples were
planted as pure stands. The two staples were dominant throughout the
survey area. There were some peculiarities among the surveyed sub-
locations. Cassava was grown by 75% of farmers in three of the four
sub-locations, Nyawita the least densely populated sub location was
the exception. Beans were grown by 60% of farmers in three of the
sub-locations, in Omia Malo groundnuts competed with beans in popula-
rity, in Nyawita and Simenya cowpeas competed with beans.

, I

Tabe The Long Rains(LR) cropping pattern in the zone and supplementary information on
Table .7. Short Rains (SR) plantings.


LR 1976;
% farms planting
Average ha p. grower
Average ha for sample
% area intercropped

% farms planting in
SR 1975

% intending SR 1976






n. a.


3 24 12 2 13 3

3 0 48 2 8 0

Cassava figures show the area in the ground and numbers growing
Once mature cassava will stand pure until utilised.

n.a. Figures for area interplanted with cowpeas are not available
37% of cowpea growers intercropped.

at August 1976.


I ** '

Minor crops, recorded on a few farms were Green Gram and Sim Sim. A
high proportion of farmers grow small areas of bananas, sukuma wiki and
other vegetables.
Differences between the higher rainfall areas in Ukwala and Yala and
the drier Bondo areas can be noted:
i. Intentions of (1976) short rains maize and bean plantings are
equally high in each area, but actual (1975) plantings are
lower in the drier areas.
ii. The area of sorghum grown per household is higher in the drier
areas of Bondo.
iii. Cotton is grown more extensively in the drier areas of Bondo
with a significant number of short rains plantings.
2. Labour
(a) Availability: In the Siaya zone farm labour is supplied mainly by
the family supplemented by hired labour, usually casually employed at
periods when the workload is heavy. Table 8 gives details of the
labour force available for farm work in the areas surveyed. Permanent
off farm workers refers to family members who have emigrated from Siaya.

Table 8. The labour force on farms in the Zone


Family normally resident
Average number in family 7.4
Average number over 15 years 3.9

Percentage of households with family members
as: permanent off farm workers (%) 43
: temporary off farm workers (%) 39

From sampled farms, numbers of members:
in permanent jobs, non resident 81
Away at boarding school 68
taking some casual employment in 1976 14

Percentage of households:
employing permanent hired labour 1
employing casual hired labour 40

Percentage of households receiving cash from
non resident relatives
regularly 11
occasionally 36

Several points are noteworthy. Adult members of 435 of families have
emigrated from the zone. Numbers emigrating are much higher in the
densely populated areas surveyed. As might be expected

Nyawita sub-location, with the lowest population density and largest
cultivated area per family has the lowest number of emigrants and the
lowest proportion of households with emigrants. Whereas in the three
more densely populated sub-locations 55% of households receive some
level of support from outside the area, in Nyawita only 25% of house-
holds receive support, none regularly. On the otherside of the coin
Nyawita farmers send a larger proportion of their children to school
and hire much more casual labour suggesting a higher level of farm
generated income.

(b) Use: Traditionally the Luo people had very marked ,specialisations
in their use of family labour. Having been pastoralists the responsi-
bilities of the men were towards the livestock. This had carried over
into the responsibility for seedbed preparation done, whenever possi-
ble, by oxen. The women were responsible for the weeding and harve-
sting. This traditional specialisation has largely broken down,
though vestiges were observed in recording crop management on over
200 individual maize and sorghum fields. Particular households stuck
to traditional practice, women performed the work without the aid of
men, particularly on the weeding operation. A further important
aspect is the extent to which farmers are prepared to hire labour for
the operations traditionally done by women as a substitute for employ-
ing male family labour on these operations. Clear evidence on this
aspect is difficult to obtain, only 3 male heads of households inter-
viewed hired labour to supplement their women and children in opera-
tions in which they, or other men in the family, did not participate.
This is a low proportion of total hirings. It is impossible to rule
out family contingencies as a reason for specialisation but, on bala-
nce, there is evidence of specialisation persisting in some 12 (11%)
of male headed households, particularly in the weeding operation. It
seems the tradition will disappear in the next decades.

Farmers were asked to name what they considered the busiest and the
second busiest period of the year. As they were asked to answer for
two periods the incidence of responses does not directly rdlect the
relative severity of the two work peaks. The incidence of responses
is reported in-Table 9. Most farmers mentioned two or even three
month periods, particularly in the long rains season. Table 9 also
shows how the incidence of reported labour hirings coincides with the
busiest periods.

Farmers consistently named February to May as the major peak period
with August to October as a secondary peak. The exception to this was
Simenya where farmers distinguished between December to February and
April to May with a small proportion of farmers putting emphasis on
land preparation for the second rains crops in September and October.
There are numerous hirings for weeding in Nyawita, the sub-location
with low'population density in which most farmers own and operate ox-
ploughs and cultivate larger areas. This contrasts with the other
three relatively densely populated sublocations, with less farmers
owning oxploughs, where the hiring of ploughs, and in the case of
Kambare labour for land preparation, predominates.

3. Capital

(a) Availability: Capital is defined as the surplus remaining for
re-investment out of incomes after consumption requirements have been
met. It is inevitably closely bound up with income levels. No attempt
was made in the course of the survey to collect information on income

Table 9. Farmers opinions on the busiest season and the
I incidence of labour and ox plough hiring on farms

Sub location J F M A M J J A S 0 N D

Omia Malo
Opinion on busy period 7 23 26 28 20 3 4 28 25 9 0 0

Ox hire-ploughing 4 4 2 1 4
(land preparation 2 2 -
Laour(weeding 1 6 5 1 -

Opinions on busy period 18 24 30 32 26 0 12 25 30 18 0 8

Ox hire-ploughing 2 -
(land preparation 1 -. -2
abureeding 1 12 17 6 2 1 1 -

Opinions on busy period 16 14 9 24 21 9 1 4 4 6 3 7

Ox hire-ploughing 3 2 1 1 3 2 -
Labour(land preparation -
(weeding 1 1 -

Opinions on busy period 14 18 23 22 13 3 3 17 18 13 9 2

ox hire-ploughing 3 4 1 -
abour(land preparation 3 4 4 1 1
(weeding 2 3 1 1 -

n.b. Land preparation includes planting

levels. Some emphasis was placed on sources of
obtained is set out in Table 10.

income and information

Sources of cash income are very varied. Most farmers (80) reported
a dependence on crop sales, mainly of food crops in the local markets.
Most farmers had two sources of cash income.

The productive investments into which surpluses are channelled in the
Siaya zone are cattle and ox ploughs as assets, the hire of labour or
machinery and the purchase of improved seeds and fertilizer. Evidence
from the collected data strongly suggests a run down in assets is
taking place, cattle numbers and the number of farmers owning ox-

Table 10. Numbers of farmers reporting various sources
of family income in the Siaya Zone

Source Omia
Source ma Nyawita Simenya Kambare Total

Crop sales 24 27 20 27 98

Livestock sales. 7 .12 6 2 27
Milk sales 8 7 1 3 19

Prom relatives regular 3 0 4 7 14
living away occasional 12 8 15 10 45

Temporary off farm work 6 2 3 3 14

Hire out oxploughs .9 10 0 0 19

ploughs, are falling. This is most probably caused by pressures ari-
sing on grazing land from increasing arable requirements and the conti-
ngency demands for cash for major household expenditures met by the
sale of animals. Cattle numbers on the surveyed farms have fallen
from an estimated 862 five years ago to 688 head in August 1976 a 20%
reduction. The number of owners has increased over the same period
from 76 to 82 or 67 of the farmers surveyed. The average herd size
has dropped from 11.3 to 8.4 head per owner. Some 98 farmers, 80%o of
the sample, reported owning cattle at one time. 27 farmers, a third
of livestock owners, reported sales of livestock in the first nine
months of 1976. 96% of these sales were to realise funds for consu-
mption. Over the last three years 74% of owners reported sales of
stock to meet cash needs for consumption and 76% of these sales were to-
make payments of school fees or to purchase food. 37% of farmers repo-
rted owning ox ploughs, 10% more reported having previously owned
ploughs. Table 11 sets out the information collected on the use of
cash as farm working capital. No information was collected on the
costs of improved seed and fertilizer purchased by farmers. The main
use for working capital was in the hire of labour and ox-ploughs. As
has been detailed in Table 9 most of the hirings were for seedbed pre-
paration and weeding.

Without information on the quantities of fertilizer and seed purchased
it is not possible to assess exactly the average level of outlay of
working capital of farmers in the Siaya zone. On rough assumptions
that each purchasing farmer is spending S 25.00 on seed and S 100.00
on fertilizer gives an aggregate Si 2225.00 for purchased inputs, and
hs 15,250 total outlay by surveyed farmers, that is an average as 125.00
per farm. Out of the full sample 67 farmers were using working capi-
tal with an estimated average outlay of S 228., of these 7 were buying
only hybrid seed with an assumed outlay of Ss 25.00, and 4 were buying

Table 11.

Use of working capital by surveyed farmers in
1976 long rains season

Item Omia Myawita Simenya Kambare Total

Hire of labour and oxploughs
Farmers hiring: no 14 19 9 14 56
% of survey farmers 47 59 30 47 46

Number of brings 28 36 11 31 106
Total cost K.s. 3894 5430 1095 2604 13023

Average cost per hiring 139 151 100 84 123
Average outlay per hirer 278 286 122 186 232
Average outlay p. survey farmer 130 170 37 87 107

Purchase of seed and fertilizer
Farmers buying seed: no. 5 13 12 7 37
% 17 37 40 23 30
---- --------------------- --------------------------------- -----

Farmers buying fertilizer: no. 2 5 6 1 14
7 16 20 3 11

seed plus fertilizer; an assumed outlay of S' 125.00.. Of the residual
56 farmers (46% of those surveyed) using working capital, 26 were
hiring labour or machinery and purchasing seed (16) or both seed and
fertilizer (10). These 46%h ad an average outlay estimated at
S 260.00 per farmer.

4. Management

Management is the coordinating resource, provided by the farmer him-
self. He works within the natural and economic circumstances of the
area and within the constraints of the other resources to satisfy his
family's needs and priorities. Evaluation of the management strate-
gies of Siaya farmers depends on identification of their objectives
and priorities. This leads into section 3, a key part of the report
which analyses the present farming system as a manifestation of farmers
strategies for achieving their production objectives.


The key circumstances which are likely to influence decision making
by Siaya farmers on the development of their farms are:

(1) The high variability of rainfall and relatively short growing

(2) Population growth; breaking down the effectiveness of shifting
cultivation and creating competition for land between use for arable
and for grazing.

(3) Adjudication, reinforcing the effects of population growth but
also creating individual responsibilities for grazing management.

(4) Wide seasonal fluctuations in selling and buying prices for food-
stuffs, particularly starch staples, in the local markets.

(5) Falling soil fertility.

(6) Capital scarcity and a running down of numbers of oxen and ploughs
as capital assets.

(7) Labour shortages in the February to May period of the long rains
season and the possibility of these becoming more acute; a result of
the apparent rundown in capital assets.



Most people have the same objectives in production activities; family
welfare comes first, then cash income; initially to meet 'necessary'
purchases and subsequently to give command over an ever widening hori-
zon of semi-luxury and luxury goods. These three 'objectives' are
clearly graded. To the man earning S 50,000 a year basic family wel-
fare is no problem. His priorities are increasing his command over
wider range of luxury goods can he afford that second car! The man
producing to the value of s 2,000 a year has quite different priori-
ties he operates close to subsistence level and the welfare of his
family dominates his production objectives. Dominance in objectives
depends where you are on the income scale. Attitude to risk also
depends where you are on the income scale. A man earning S 50,000 a
year may happily speculate with s 2,000 of his income, he may have to
go for a smaller model of second car if his speculation fails but the
consequences are not far reaching. Clearly a man earning s 2,000 a
year can't speculate with it, the very survival of his family is at
stake. Vis-a-vis the S 50,000 a year man's 4% speculative spending,
this man's risk ceiling is only Ss 80.00. Farmers producing close to
the substistence level will only move away from a tried and tested
system, as long as it is meeting their priorities, in small steps.

Siaya farmers are low on the income scale, family welfare dominates
their production objectives. Family welfare in the form of an assured,
reliable, supply of food and enough cash to buy necessities. Within
this dominant priority of a reliable food supply the family have prefe-
rences; these are now discussed.

Traditionally sorghum and finger millet were the starch staples for
Siaya families, supported by supplementary cassava and sweet potatoes.

Finger millet and to a lesser extent sorghum have been replaced by
maize as the preferred starch staple. Of 122 farmers interviewed in
four sub locations 101 expressed a preference for eating maize ugali,
11 a preference for sorghum ugali and 10 preference.
Maize ugali is considered tasty by itself though some form of relish
is 'usually eaten with it. Sorghum ugali is considered by most fami-
lies to be heavy and tasteless, 91% of farmers reported mixing sorghum
with cassava flour to lighten it. It also needs a strong relish to
'liven it up'. Sorghum ugali has a reputation for 'sitting on the
stomach' and causing constipation if taken by itself. In the lake-
shore areas, with ready access to low cost fish, it is considered as
a good 'vegetable' to eat with sorghum. In addition to fish and meat,
eaten: when cash is available, a wide range of vegetables are used
with ugali. Cowpea leaves and sukuma wiki are the most common culti-
vated vegetables and are widely sold in the local markets. Green
grams and groundnut are also common. Beans are not used as a side dish
but are mixed with maize as an ugali very acceptable by itself. A wide
range of local plants are used as vegetables; some, including mito,
sewewe, akao, ododo and onulo may be cultivated, others, including
osuga, omboga, odielo and Apoth (ochte),are gathered. When making
relish for use with ugali vegetables are often combined, popular combi-
nations are cowpea leaves/mito/apoth/onulo, osuga/ododo and Akao/
omboga, said to be particularly good with ghee. Mito is pepperminty
in flavour and has a reputation in curing bad stomaches, with Akao
and omboga added it is used to.stimulate milk flow in nursing mothers.
Sweet potatoes, relatively widely grown, is considered as a supplemen-
tary starch staple, not a meal in itself. Sorghum is of major import-
ance as a base for local brewing. Home brewing is a mark of capable

In this context a good farm manager, a man respected by his fellow
farmers in Siaya, is one who produces enough maize, every year, to feed
his family on their preferred food through to the new harvest, who
produces enough sorghum to allow his wife to brew regularly all thro-
ugh the year, and who produces a surplus to bring in enough cash to
meet the families needs for household requirements, fish, meat,clothes
and children school fees, and to add to his small livestock herd from
time to time. Management strategies, in the face of the prevailing
local natural and economic circumstances, are directed towards achiev-
ing a reliable food supply within the resource constraints operating
on the farmer as manager.


This section details the evidence collected on the hazards that Siaya
farmers feel they face, contingencies created by these hazards and the
actions taken within the household to combat them. The food prefere-
nces described in Section'3.1 are often subordinated to the basic
priority of food sufficiency due to hazards beyond the farmer's cont-
rol. Farmers were asked to state their major problem in keeping their
family's supplied with food in an effort to identify what they percei-
ved as the most important hazards influencing the achievement of their
primary production objective. Table 12 reports their answers.

The main hazard reported by farmers in three of the four sub-locations
was unreliable rainfall. In Simenya it was shortage of land. An
attempt was made to evaluate the effects of all such hazards on family

Table 12.

Identification of the main production hazards
recognized by Siaya farmers

Omia T tal
Source Mal Nyawita Simenya Kambare No
Mue al o No %

No reply 2 1 3 1 7 6
No problem 13 1 2 2 18 15
Hailstorm 5 0 0 0 5 4
Drought, unreliable rains 15 19 5 16 55 45
Shortage of capital 1 12 6 2 21 17
Shortage of labour 0 3 1 6 10 8
Shortage of land 0 1 12 0 13 11
Poor soil, erosion 1 0 4 2 7 6
Pest attack 4 0 1 6 -11 9

food supply and the type of action taken by the farm family to allevi-
ate the contingency arising. Table 13 reports the incidence of some
hazards and some resulting contingencies for the years 1974 76.

Table 13.

The incidence of some hazards and some contingencies
arising in Siaya farming 1974-76 (numbers of farms

Year 1974 1975 1976

Maize attacked by army worm 2 4 49

Maize damaged by hail 19 55 29

Replanted some maize fields 19 33 56

Replanted some sorghum fields 14 19 42

Ate maize throughout the year .26 30 17

Finished off all stored sorghum 77 73 56

". t : -

A more detailed examination was made of the food supply situation
arising from the 1975 long and short rain harvests. Table 14 gives
the results.

FT e arm families starch staple consumption for food and beer
from homegrown and purchased sources August 1975 July 1976

Period By end During During During
Oct 1975 Nov-Jan Feb-Apr May-Jul

MAIZE No. finishing 11 18 55 38

No. starting 103 6 .5 2
Farm SORGHUM No. finishing 4 14 35 60

CASSAV No. starting 49 8 16 5
No. finishing 3 8 8 59

No. starting 5 14 39 29
MAIZE No. finishing 0 3 5 79

Bought SORGHUM No. starting 4 6 10 11
No. finishing 1 1 1 28

No. starting 10 4 5 3
CASSAVA No. finishing 0 1 3 19

The table should be seen in the context of farmers reports on their
own 1975 grain harvests. There was considerable variation between
sub-locations; in Simenya and Nyawita the majority of farmers reported
good 1975 harvests, in Omia Malo and Kambare the majority reported
poor 1975 harvests. Overall the split was about fifty-fifty good and
poor harvests. First maize pickings were reported in May, however
98% of reported maize harvests occurred after July 1st and 77% after
the end of July, though 25% of farms reported eating green maize by the
end of June. Only 31% of farm families had enough maize to last until
the May/July period, 69% finished their home grown maize supplies by
the end of April. Households purchasing maize increased significantly
in the period February to July, but all reported finishing purchasing
by the end of July. A higher proportion of households had home grown
supplies of sorghum and cassava lasting through until July and purcha-
ses of these starch staples were not so widespread as of maize.

With maize as the most commonly purchased food farmers were asked where
they obtained money from to buy maize during 1976. Their answers are
set out in Table 15.

abe 1. Sources of money for
ae 1. the 1975 harvest

farmers buying maize since

Omia Total
Sub location Malo Nyawita Simenya Kambare
N____o %

Animal sales 5 9 5 4 23 19

Crop and vegetable 7 2 5 7 21 17

Relatives 2 2 9 7 20 16

Casual labouring 0 6 3 5 14 11

Poultry and milk sales 1 1 3 3 8 7

There are a wide range of tactics available within the household for
spinning out food supplies. Maize may be mixed with sorghum or cassava
flour. Maize may be purchased, sometimes from the sale of another
stored crop. .In extreme situations, when the cash cannot be raised or
the maize price is considered too high, then sorghum and cassava flour
will be mixed in making ugali. In emergencies sorghum is diverted
from brewing and.used for food, albeit the varieties are not preferred
for ugali. Oloro and Ochuti, two red, bitter and hard varieties of
sorghum with storing properties upto 2 years long under good management-
will be kept as a fall back food and brewing supply, often being sold
from store or eaten once it can be seen that the new harvest will be
adequate. Control of the foodstuffs is left with the senior woman of
the household. She is responsible for the state of the food stores and
for liaison with her husband, in times of shortage, as to how supplies
will be spun out. She will for example initiate the plan for the ope-
ning of a new cassava field in anticipation of scarcities 6 months or
a year ahead. Storage-of foodstuffs is an important part of household
management. Foods are stored inside the house, in the smoke of the
fire and insect damage minimised by the use of ash scattered in the
stored crop. Keeping qualities are enhanced if grain is removed from
the store and re-dried from time to time and new ash mixed in. Some
farmers now use purchased insecticides for storing grain. Out of the
122 sampled 65% still used ash, 8% used purchased insecticide alone,
7% used insecticide with ash and 20% reported not using either.


Actions in the household reacting to food contingencies are 'curative'
facets of the farm families strategies to ensure a reliable supply.
Such actions are taken once contingencies show up. 'Preventive' fa-
cets are reflected in the farmers' decision making in managing his
farm. Here the farming system is briefly analysed under categories
of decision made by the farmer within local natural and economic cir-
cumstances, his resource constraints and his experience of the inci-
dence of the various hazards facing him. A pre-occupation with relia-
bility of food supply reflects its dominance in farmers objectives.

1. Cropping pattern

The cropping pattern is a straight reflection of the consumption pat-
tern just described. Maize as the major crop among the survey farmers,
155.3 ha including intercropped areas in the 1975 long rains, closely
followed by sorghum, 118.3 ha also including intercropped areas, with
the two crops sometimes mixed on the same plot. Some 75% of farmers
in three sub locations grew a much smaller area of cassava, Nyawita
sub location, with a high average area cultivated per farm being an
exception with only 25% of farmers interviewed growing cassava. Sweet
potatoes grown by 39% of farmers on a very small average area, and
Finger millet grown by 25% of farmers, were back up starch staples.
Again Nyawita with a high proportion of ox owners and a large culti-
vated area per farm was exceptional with 56% of farmers interviewed
still growing Finger millet. When farmers were asked which crops they
had stopped growing or reduced the area of finger millet and the longer
term red sorghums featured most prominently. Other crops featuring
most prominently in the pattern were the legumes and pulses used as
complements in the menu to the starches. These varied to some extent
from area to area; Beans was grown by 60% of farmers in three sub-loca-
tions; Omia Malo was the exception with only 17%. Cowpea, green gram
and groundnuts were others. Dominance of the basic starch staples in
the economic activity of the Siaya zone is also demonstrated by the
number of farmers selling these crops, mainly in the local markets.
21% of farmers interviewed (wholly in Omia Malo and Nyawita) reported
selling cotton, and 16% reported selling beans often quoted as a cash
crop. However figures for maize (53%) and sorghum (43%) were much
higher, with 40% of maize sellers and 53% of sorghum sellers making
some of their sales during the period of high local market prices in
April June.

A range of varieties of the main crops are grown by Siaya farmers.
Notes follow on the main varieties identified in maize, sorghum and

(a) Maize Three varieties were noted:
i. Rachar: which means white, was the preferred variety for
making ugali. It is a relatively good storer and takes about
5 months to mature. It is also called Ndere.

ii. Nyamula: Is a yellow maize, not considered as palatable as
the alternatives. It is shorter term than Rachar and consi-
dered drought resistant. It is also called Kipindi.

iii. Hybrid: (Either H512, H511 or H632). Is considered very pala-
table even preferred to Rachar by those farmers using it.
It is longer term than Nyamula and does not store well.

(a) Sorghum. There exists a profusion of sorghum varieties and there
is some confusion in identification with the same variety having diffe-
rent names in different parts of the zone.

i. Ohunjo: A light brown variety favoured for both ugali and for
brewing. It gives a lighter coloured appearance and less bitter
taste than the red varieties when used in ugali. It is a short
term variety taking about 3 months to mature but is vulnerable to
birds when planted early and matures in advance of other varie-
ties. It stores very poorly and will be the first sorghum used
by the family. It is often sold for cash because of its value
for brewing and poor storage quality.

ii. Oloro similar to Ochuti though ochuti is goose-necked.
Both are popular for their bird resistance and good storage quali-
ties but are not preferred for either food or beer. They are
mainly 'insurance' varieties. Both are long maturing, red varie-
ties and have a hard seed with a bitter taste. Once the new
harvest is ensured these varieties may be sold. They can be "--
stored for over two years with good management.

iii. Kumba: A white (glumed) variety, favoured for both ugali
and beer but very prone to bird damage as it is (relatively) sweet.
It has a long maturity period and a relatively poor yield. It is
a fading variety and little seed is available now.

iv. Serena: Short talked, quick maturing in two and half months,,
a red/brown variety. It is well liked for both ugali and beer
but will store for less than six months. Because of its short
maturity it can be grown on the short rains and sold on a high
price market for brewing early in the new year. Seed is diffi-
cult to get.

Other varieties recorded were: Andiwo, Jagopari, Othuwa, Hongrangai,

(c) Cassava. As with sorghum there is a problem with different names
for the same variety on different farms. The three most common varie-
ties were:
i. Duruma: Can be eaten fresh from the ground after boiling-or
made into chips. It can be consumed less than a year after plan-
ting and remains palatable for two years after planting.
ii. Khamisi: Is not eaten fresh from the ground, only used for
chipping. In times of severe hardship the leaves will be used for
a vegetable. Similar in growth pattern to Duruma.
iii. Rateng: Is reputed to be a high yielder and can remain pala-
table for well over two years after planting.

Other names noted were: Kiganda/Nyaganda, and Rabuor.

2. The Crop calendar and seasonal decisions

The range of crops and varieties discussed are management tools for the
Siaya farmer and give him flexibility in deciding, within his resource
constraints,which crops and varieties to grow, how much of each and
which will take priority for establishment. The farmer has two major
types of decision during the cropping year: what cropping pattern to
adopt, and how to meet contingencies which arise as the season progre-

The farmer will have a 'usual' cropping pattern as a basis for his
planting decisions. This will depend on his families food preferences
and his attitude to risk, particularly vis-a-vis the balance between
maize and sorghum. At the beginning of the, season this 'usual' pat-
tern will be modified by the current food stock situation from the last
harvest and by knowledge of any special resource constraints (his wife
may be pregnant) likely to operate in his forthcoming planting season.
A wide range of scenarios can be example: with good stocks on hand of
maize and storable sorghums the farmer may decide on initial priority
to a large area of the preferred Rachar maize variety, then a small
area of long term sorghums, Oloro or Ochuti, to supplement his store
and finally, in the long rains planting, a large area of preferred
short term sorghums, Ohunjo or Serena to give good brewing prospects
and possibly as a source of cash income. With very low food stocks his
priorities may be exactly the opposite; initially an area of Ohunjo or
Serena giving the best chance of early new food supplies, then-a large
area of the storable long term sorghums' trying to assure food grains
through until the next season, finally an area of the short term, though
less preferable, Nyamula maize. In such a case quantity takes clear
priority over preferences. This sort of decision process faced again
at the short rain plantings though, because of the short season, the
alternatives are more limited, the long term varieties are excluded.

Within this framework of decisions on cropping pattern unforeseeable
contingencies arise, two of the common ones for Siaya are described:
-(a) Rains, having started enough to allow planting, stop and early
planted maize is lost. The farmer may decide to replant the same
variety, replant a shorter term though less preferable variety or re-
plant with Ohunjo, a short term sorghum to give as much guarantee as
possible of a grain crop. Similarily decision situations would arise
with an attack of army worm.
(b) The grain crops are maturing badly, rain is scarce in May and June
and grain filling is suffering. Depending on the urgency with which
he sees a need for contingency food supplies the farmer may plant sweet
potatoes, with prospects of food in four or five months, or cassava,
with much greater flexibility as to when the production will be used.
Farmers reported using cassava roots five months after planting, others
reported expectations of using cassava roots three and a half years
after planting. Adroit use of the crops and varieties with which he
is familiar is an important tool for Siaya farmers.One of the most vi-
tal aspects of farming in the sub-optimal rainfall areas is to have
enough flexibility with available management tools to react to hazards
of weather, pests and disease beyond the farmers control. In Siaya
flexibility is reduced towards the South and West of the District; the
short rains is a less reliable standby and even in the long rains the
longer term varieties must be planted early to give them the possibility
of maturing within the available rainfall.

3. Crop Management practices

A description of the major crop management practices is given in the
course of section 4 of the report. Not all management practices can be
directly related to the farmers primary objective of an assured family
food supply, however many can be indirectly related to it. Some of the
intercropping practiced is directly related to assurance of food supply.
In particular the mixing of maize and sorghum with the proportions of
the two crops reflecting the farmers' own balance between a preference
for maize, particularly perhaps the early green cobs, and a desire for
certainty in grain production from the efforts expended in preparing the
field. It is noticeable that sorghum is the.more dominant in such mix-
tures in the drier areas of the South and West.

S A major management practice indirectly related to the need for an
assured food supply is time of planting. Farmers fully appreciate the
yield benefits of early planting, particularly for the long maturing
grain varieties. Constraints on land preparation prevent all crops
from being early planted and the considerations of food stocks as well
as the seasonal contingencies outlined in the last section order the
priorities. It is quite clear that the very late planting of cotton
is a direct consequence of the priority given to food grain establish-
S- .-ment;' --The-, traditional method-of- planting by broad c ast ing- though how hro-w
breaking down,at least in maize, lent itself more easily to the urgency-
of getting crops established as early as possible. Thus important
management practices, which agricultural experimental programmes usually
seek to influence, can be seen to be the indirect result of the priori-
...ties of the Siaya farmer and the constraints within which he has to meet
these priorities.


The farming system in the Siaya zone is in a rapid state of evolution,
under very heavy pressure from increasing population density. Of the
sampled areas Nyawita sub-location is probably the closest to 'the
S.. :rad~tional.' Siaya farming system. Simenya-..seems -perhaps~ to have bee.nlm
pushed farthest away from it but is relatively fortunate in having far
greater flexibility in management because of the better rainfall regime.

Increasing population has exerted pressure by raising the area required
for growing crops with two major consequences:

1. There is no longer the possibility of maintaining soil ferti-
* lity by shifting cultivation.
2. The arable area has grown at the expense of grazing, reducing
the livestock carrying capacity of the remaining grazing land.

These two consequences have had inter-related effects on the evolution
of the Siaya zone farming system, the effects have been intensified and,
usefully, made clearer to farmers in those areas where adjudication has
taken place. Falling fertility has brought falling yields and the
increased use of cattle manure. 62% of farmers reported using manure,
and 12% of those farmers using were not livestock owners. The inability
to shift the cultivated area has brought on Striga as a major weed pest;
67% of farmers acknowledged it as a problem on sorghum and 52% felt it
was also a problem on maize. It has intensified the labour needed for
weeding; 55% of farmers felt that uprooting was the only way to get rid

of it. On the other side of the coin the reducing grazing area has
begun to reduce the.cattle numbers on Siaya farms. It is possible
that the reported fall in numbers over the last five years in a cycli-
cal phenomenon. The fall off in numbers owning ox ploughs on the other
hand indicates it is a general trend. Given the fact that farmers in
the area aspire to own oxen and plough and the contrast between the
two areas; the reporting of 17 farmers operating ploughs in Nyavwta in
1976 and only 1 in Simenya supports it as a general trend. As cattle
number fall, particularly if there are low calving percentages, long
calving intervals and high calf mortality, as would be expected under
overstocked conditions, it becomes more difficult to maintain a team
of working oxen. Farm surveys in Ethiopia (Kelber and Mela 1970)
suggest that 5 6 head of stock need to be kept to support one ox in
working condition. Even using a figure of 4 head per ox implies a herd
size of 8 animals to support a working pair and 16 to support two pairs.

Thirty one farmers, that is 30% of-the sample reported working ox ploughs
during the 1976 season, predominantly in the Southern sub locations
were the shorter season would increase the urgency of timeliness in land
preparation. Of those working ox ploughs 81% had a herd size greater
than eight head of'cattle, a further 13% of farmers currently working
ox ploughs reported they owned more than eight head tive years ago.
Falling numbers of cattle in general and draught animals in particular
has two disturbing effects on the future development of the farming

1. The use of manure as an alternative to shifting cultivation for
maintaining soil fertility is jeopardised.

2. Increasing use of hand cultivation techniques implies even less
timely planting due to labour constraints or a reduction in the area
cultivated per family. It has particularly difficult implications for
the use of longer term varieties in the South and West where the relia-
ble rainfall season is much shorter, and timeliness in cultivation and
planting more critical.

It is apparent from the extensive feeding of the leaves and stalks of
maize and sorghum, reported by 93% of cattle owners, that the feeding
situation is difficult. Leaves are stripped from maize still standing
in the field once plant growth has stopped and while the cob is milky.

Extrapolations of these apparent trends in the Siaya farming system are
inevitably gloomy. The combination of falling yields and increasing
labour constraints will reduce the usefulness of the traditional mana-
gement tools and the flexibility of action of the farmer. One must
foresee an increased incidence of food scarcity in the area, lower cash
surpluses as the farmers priority for an assured food supply further
dominates his production decisions, a sacrifice of preferences for maize
and a move into the lower risk starch staples; sorghum and cassava.


To allow the planning of adaptive agricultural research which effecti-
vely serves Siaya farmers, decisions are needed on the direction in
which the farming system should be developed. In such decisions it must
be borne in mind that if the direction chosen for development is to be
acceptable to Siaya farmers it must be consistent with their priorities
and the steps along the path must be within their resource capabilities
to implement. We have seen which way the system is moving, the major
question is whether to try to reverse this trend; preserve the live-
stock enterprise integrating it more closely with the cropping, or acce-
pt the trend.and try to ensure an efficient use of arable land without
a livestock component.

Elements of a true mixed farming system have emerged in Siaya. The long
established tradition of draught animals and more recently the use of
animal manure and the feeding of maize and sorghum leaves and stalks to
the animals. In the adjudicated areas stockowners, grazing animals on
their own land, have begun to be more aware of the pressures and more
conscious of their own responsibilities for solving the problem. Given
the very strong tradition and attraction to livestock among the people,
they are likely to be receptive to developments in grazing management.

The potential benefits of keeping livestock are there; draught power
with lower human energy requirements and better timeliness of operations,
manure, milk, meat and potentially, grass as a valuable rotation crop.
However, given the population density, large herds can not be supported.
The evidence shows they are already fading out. With the relatively
poor rainfall regime only modest stocking densities will be feasible.
Traditional herd structures are too extensive to support enough milk
cows and draught animals. It seems probable that a choice will be needed
between draught and milk animals and that either draught or milk animals
will have to be brought into the area from outside, or if the choice is
for milk, artificial insemination introduced. The ultimate form of
livestock enterprise in Siaya farming is well into the future. Infor-
mation from this present study, concentrating on maize research, is
inadequate to draw conclusions. The questions posed need answers if
livestock and pasture research is to be oriented to the needs of Siaya
farmers. The questions posed are:

1. Given the population density and the stocking densities feasible
under.the rainfall regime can the benefits from livestock; draught use,
manure, milk, meat and grass as a rotation crop, justify the use of
scarce land by a livestock enterprise in the Siaya farming system?

2.. If so, given the limitations on grazing area and stocking'-der-sity.,
is a choice between milk and draught necessary, with other benefits the
same in both enterprises?

3. If so, which alternative; milk or draught, is of greater potential
value to the Siaya farmer. A decision will require evaluation of the*
purchasing/breeding and feeding management possibilities for each

4. From the current existing situation what sequence of steps in the
development of the chosen enterprise are compatible with Siaya live-
stock owners priorities and capacities?

On the crops side the position is less complex. Ultimately a decision
on the desirability of developing the livestock enterprise will have
important implications in the choice between organic and chemical
fertilisers as a means of maintaining soil fertility. In the short
term the direction of crop development can be determined independently
without compromising a decision on the livestock side.

The strong farmer priority for a reliable food supply and the high
proportion of available farm resources being used in food production
pinpoints the foodcrops, particularly the starch staples, as the focus
for development and thus for research work. Efforts directed towards
satisfying the farmers' food priorities with less of his resources will
allow reallocation of released resources to production for the market
and increased cash incomes. A high level of risk aversion is embodied
in the existing cropping pattern and decision strategies. Innovations-
which reduce the effects of rainfall variations, as the pre-dominant
hazard, on crop production levels will allow a reduction in the extent
of insurance cropping releasing farm resources to cash crop production.
Reducing uncertainty in food supply is development orientation in
harmony with the farmers' own priorities. Maize is the preferred starch
staple and, at the same time, the crop most vulnerable to rainfall varia-
tions. Research to improve the productivity and reliability of the
maize crop has a key role in implementing a development policy designed
to reduce uncertainty in food supply and encourage a shift of resources
into cash crop production. Section 4 details present maize management
and sets out a research programme designed to give technology for
improving maize productivity which is appropriate to the circumstances
of Siaya farmers.

The analysis of the farming system allows some comment on criteria to
aid the orientation of research efforts in sorghum, cassava and cotton.
The limitations imposed by present resource constraints will be similar
to those to be discussed in some detail for maize; largely problems of
capital and labour scarcity. Comments are restricted to an outline of
the crop characteristics which would be particularly acceptable to
Siaya farmers.

1. Sorghum. As we have seen sorghum is a multi-purpose crop in Siaya
farming, it is used for brewing, as a supplementary starch staple and
as an insurance crop being less susceptible to rainfall failure than
maize, and finally as a cash crop. Serena is well favoured for both
brewing and food, its short maturity period gives the added flexibility
in management which is of such value to Siaya farmers. It shortcoming,
like that of Ohunjo, the local short term variety, is poor storing
quality. If short maturity and good storing characteristics are gene-
tically compatible, breeding work to incorporate storing quality into
a, serena type variety would allow a single variety to cover the needs
of Siaya farmers. It would reduce the need for the longer term sorghum
presently grown for their storing qualities and which, because they are
long term, necessarily take priority for farmers' resources at the cri-
tical period early in the long rains planting time; February to April.
Resources released from these purely insurance crops would be re-allo-
cated into cash earning opportunities.
2. Cassava. Cassava remains a valuable crop in the farming system as
a complement in ugali making to both maize and sorghum and as a rotation
crop. In addition to improved yields valuable characteristics to Siaya
farmer are:

Cab The ability to form tubers very rapidly
b The ability to remain in the ground in a palatable state for
a three period or better.
(c) Leaves which are acceptable as vegetables to Siaya families.

Here the emphasis in the supplementary characteristics is to give the
farmer a management tool which increases his ability to react to food
supply contingencies arising as a result of poor growing conditions
for the preferred grain staples.

3. Cotton. Given the pre-dominance of reliable food supply as s pro-
duction objective among Siaya farmers cotton planting will always be
subordinated to the establishment of food grains. Improved producti-
vity of the food grains will allow an increase in the area devoted to
cotton, and a degree of earlier planting as the area of food crops
grown to meet fixed family needs is reduced in response to lowered
S risks of failure. Cultivation and planting the smaller area absorbs
fewer resources. However cotton would still have to compete for the
released resources and earlier planting time with other cash crops, inc-
J" luding food crops produced in surplus for sale on the local markets.
In the Northern and Eastern areas of the zone where relatively sure
short rains crops can be grown, in addition to the long rains ones,cotton
is unlikely to become a viable cash crop. Picking the crop is labour
intensive and it must compete for labour with the establishment and care
of the short rains crops. A further disadvantage is that boll opening
in the rains downgrades a high proportion. of the lint. In the South
and West on the otherhand, if food crop production from long rains crops
can be made more reliable, removing the need for highly speculative
short rain plantings of food crops, cotton might spread further as a
cash crop subject always to its competitiveness with other possible
cash crops. Because food supply is a priority cotton is always likely
to be planted in May or later. Screening varieties for tolerance to
late planting would fit selections more closely to the priorities and
resource capacities of Siaya farmers. Although the UK varieties were
bred to carry over a short (3 week) drought period it is unlikely that
this principle could be extended to justify short rain plantings in
Siaya, the dry period from December to February is too long.



This section of the report examines the implications of the Siaya farm-
ers situation for each of the components which usually features in a
research programme designed to improve maize management, and in the exte-
nsion recommendations which emanate from such a programme. Components
covered are; variety, the use of purchased inputs, method of planting,
time of planting, thinning and weeding. Each component will be examined
on the basis of present practice in the Siaya system, determinants of
present practice, the implications for the system of changes in practice
and the conclusions on the part the relevance of each component to a
research programme oriented to the needs and circumstances of Siaya far-
mers. The final part of this section outlines a maize research progra-
mme for the Siaya zone based on the conclusions drawn for each component.

1. Present Varieties

The present maize varieties grown have been described in section 3;
Rachar a 130 150 day white variety is preferred for ugali making. The
longer maturing hybrids (not specifically named by a.y surveyed farmers)
and the rather shorter maturing H 511 and H 512 were also preferred by
farmers for ugali. The shorter term Nyamula, a yellow variety, was
used frequently by farmers due to its ability to produce grain earlier,
or to give a crop despite relatively late planting, though it is not
preferred for ugali due to the yellow appearance of the food. These
are the varieties being used and the ones for which seed is readily

The survey showed 29% of farmers using purchased hybrid seed. This
figure is rather higher than data produced by Gerhart (1975) for his
Zone 3. Covering an area including Siaya (except the North of the
District) and South Nyanza, he found 15.8% of farmers using hybrid in
1973. Hesselmark (1975) found 13% in 1974 in the same zone as
Gerhart. Repeating this work in the 1975 season Hesselmark (1976)
found 35% of farmers using hybrid seed. The increase is usage in 1975
and 1976 is possibly the result of a.major sales campaign, supported by
demonstration plots mounted by Kenya Seed Company partly- as a result of
the 1973 and 1974 findings. The impetus of this campaign is confirmed
by farmers responses concerning year of first use in the survey done
for this study. Although 29% of farmers interviewed used some hybrid
seed in 1976, 39% reported having used hybrid at some time, of which
half (24 farmers out of 48) tried it for the first time in either 1975
or 1976. 10 farmers, 42.%of farmers who used hybrid before 1975 had
stopped using it. This is not so dramatic as Gerhart's (1975) findings
for his Zone 3. He reports 15.8% using hybrid in 1973 yet he also
reports a total of 32.9% having tried hybrid in the zone, including
1.1% trying it for the first time in 1973. The implication here is that
a massive 56% of farmers who had tried hybrid in Zone 3. before 1973 had
reverted to local varieties. Gerhart (1975) asked the question 'why
do you not plant hybrid maize'. A similar question was asked in the
current survey..and the sets of answers are compared in Table 17. As .
Gerhart notes total responses exceed :the number.of non-adopters sin-ce
some farmers gave more than one answer. Of responses in the'1976 survey

Table 17. Responses to the question why farmers do not use hybrid seed

Response 19735
season season

Cost too high 52 77

Availability difficult 6 1

Never heard of it 14 7

Performance 20 24

Doesn't fit in with local practice 7 8

criticising performance,10 specifically mentioned too long to maturity
or susceptibility to rainfall shortages. In both surveys costs of seed
and complementary inputs required were the major reason advanced for

2. Characteristics of maize varieties suited to farmers
in the Siaya zone.

Diagrams 1, 2 and 3 are for Ukwala Division, South Gem location, Boro
Division and Bondo Division respectively. The solid histogram shows
average monthly rainfall, the broken lines show probability levels for
1, 2, 3 and 4 years out of five. The shaded areas show the water requi-
rements for 150 day, 120 day and 95 day maizes. It must be emphasised
that the rainfall averages and probabilities and the water requirements
are very crude estimates. Rainfall data for each Division was obtained
from local records maintained by the Ministry of Agriculture. All avai-
lable site and year records were pooled-to estimate averages and proba-
bilities. Water requirements estimates were based on work done by
crop scientists in Kenya. Brown and Cocheme (1969) P.288 maize water
requirements, Cooper (1975) P.8 H613c water requirements at Kitale, Law
(1974) Growth patterns for major Kenya maize varieties,Darrah (1973)
Effects of altitude changes on variety maturity. There are clear diffe-
rences in the ',goodness of fit' of varieties with varying lengths of
maturity in the three areas. Common features of the three areas are:
highly unreliable rainfall in February which effectively limits earliest
plantings to the end of the month or early March, and a shut off in
rainfall in December. Beyond this each area is examined briefly.

(a) Ukwala Division and South Gem

This is the part of the zone closest to the higher rainfall areas in
the north. Only one site with limited rainfall data (Akala) was avai-
lable in South Gem. Much of this large sub-location may be more appro-
priately linked with Boro Division. On average the rainfall distribu-
tion can carry a long term (150 day) variety in the long rain season
and a medium term variety (120 days) in the short rain season. On
average residual moisture from relatively reliable May rainfall, helped
,by, t1he contribution in June, will allow a 150 day maize to finish in -- -' i
the long rains. However in two years out of five June-rainfall is rela-
tively low. Accepting that residual moisture can help the crop for a
Month, long term varieties still grain filling into July will be preju-
diced in a relatively high proportion of years. Given:

i. Siaya farmers priority for reliability in food supply and
flexibility in management tools.

ii. The predominance of hand cultivation techniques in this area
implying inability to plant much land on the beginning of the rains.
(81% of all maize plantings in Simenya and Kambare, the two sub-
locations in this area, were made after March 1st, and 35% were
made after April 1st. Long term varieties would finish poorly in
a high proportion of years if planted after the middle of March).

There is a strong case for a 120 day variety for the long rains capable
of finishing off residual May rainfall even when planted in mid-March.
There may be a case even here for a shorter term variety in the long
rains to give more flexibility in planting time. Similarily, because of
the relatively poor October and November rainfall in two years out of
five, there is a case for a shorter 95 day term variety for planting in
the short rains.

(b) Boro and Bondo Divisions

Based on the same grounds of farmer priorities and resource constraints
the case for a short term variety (95 days) in these areas is very
strong. A 120 day variety is too long in the drier areas of Bondo Div-
ision, even when early planted. In Boro grain filling of such a variety
would be jeopardised in a high proportion of years due to relatively
unreliable rain in May and June. Planting after the first of March *
would increase the failure rate. Similarly a short term variety gives
a better chance of success in the short rain period, though, even in
Boro Division, as indicated by the September and October probabilities
in particular, the failure rate will be relatively high. Chances of
a short rains crop in Bondo Division are very low. As we have seen, a-
high proportion of farmers have expectations of planting maize even in
the Southern and Western parts of the zone. (94% for Omia Malo and
Nyawita, sub locations in this area). This can be done within the tra-
ditional system at very low cost but it is clear that the yield possi-
bilities from the short rains levels for this area could never support
a high cost system of maize production. Importantly, given the popu-
lation pressure, a short term maize will allow two crops from the same
land. A shorter term variety will improve flexibility in management in
ways which are of great importance to farmers in the Siaya zone.

(i) A shorter term variety will allow earlier food supplies in
seasons following a poor harvest when household stocks are low.

(ii) Early harvests from a short-term variety will save farmers
from having to purchase maize on the local markets at a time when
prices are very high.

(iii) At the same time good managers will be able to sell maize on
local markets at a time of high prices. It will be a profitable
crop for them and it will help to reduce seasonal food prices by
bringing supply closer to seasonal demand at this time.

(iv) In years when rains begin late, or start and stop again, a
short term variety will more often be able to finish properly
before the rainfall trough in June and July.

Overall a shorter term maize variety will improve the reliability of
Supply of farmers preferred starch staple and reduce the need to use
resources on the long term 'insurance' sorghums which, because of their
long maturity period, compete for cultivation resources in the early
long rains and push back the timing of many of the maize plantings.
There is insufficient information on the interseasonal yield distribu-
tions of various varieties to permit a probability analysis on yield
variability among existing varieties.

3. Other desirable varietal characteristics

Yield remains the key criterion for variety selection, the selection
should out yield the local Rachar under management conditions achievable
by the local farmers. Yield is qualified, on the grounds set out above,
by reliability in season to season yield levels, and the increased
flexibility in management given by a short term variety, of particular
value under farmers' circumstances in the Siaya zone. Further chara-
cteristics require consideration in selection work. Any selection should
be wlhite-giving a preferred:appearance to the ugali. Selections should

1 Monthly average rainfall (mm) pooled for
im 270 available years and sites solid line
2. Probabilities for critical months showing
240 number of years in five rainfall will be
0 lower than levels indicated by the broken

210 3. Approx. water requirements 150 da,
for 3 varieties of maize.
120 da'
180 95 dai

150 4

1 3 ---------

120 4 -----

4k I /il ; 1 i 1 ; 1 1,
0! -1 ,-.-i

3 1

30 1 ------

Feb Mar Apr May Jun Jul Aug


Sep Oct Nov Dec

Diagram .2.

nm 270









----- 4



/T T! i- r'-'-r-i- r -"
4 IIII I I i
/I I 1 1



1. Monthly average rainfall (mm) pooled
for available years and sites; solid

2. Probabilities for critical months
showing the number of years in five
rainfall will be 'less than' the
levels indicated by the broken lines.

3. Approx. water requirements for
3 varieties of maize.
150 day ;!;;;,,
120 day
-----4 95 day
4- ----

i I -- -
2!;-- ---3



Sep Oct Nov Dec

Jun Jul Aug

Jan Feb Mar



Diagram .3.










-4- -- 4


1. Monthly average rainfall (mm) pooled
for available years and sites; solid

2, Probabilities for critical months
showing the number of years in five
rainfall will be 'less than' the
levels indicated by the broken lines.

3. Approx. water requirements
for 2 varieties of maize. 120 day
+ 95 day

4- -------



Aug Sep

4 ------I

3 ---











: '1

store well; the seeds should stay on the cob and be at least as resi-
stant to storage insects as the local variety but should also be easy to
shell and hand pound into flour. Costs of the seed will be an important
element in the farmers' evaluation of the change. A majority of non
adopters gave cost as their reason for not using hybrid seed, however,
it is unlikely that the cost of seed alone will deter farmers from
using a hybrid, but as will be argued later it is the cost of
fertilizer as a complementary input which is a major deterrent to adop-
tion. A choice between hybrid and composite will depend mainly on the
efficiency of the seed distribution in the zone, this, from available
information,will require improving by helping stockists to finance the
holding of seed stocks. If performance on other criteria is not signi-
ficantly different a composite variety would be preferable.


The purchased inputs, in addition to seed, relevant to improved maize
growing in Siaya are fertilizer and insecticide against stalk borer and
storage pests. Of major concern, because of the high and the urgent
need for an alternative means of fertility maintenance, is fertilizer.
11% of the interviewed farmers used'purchased fertilizer in 1976, all
these were using hybrid seed. 62% of farmers were using animal manure.
The 46% of Siaya farmers buying purchased inputs had an estimated average
Soutlay of "s 260.00 per farmer. Using it as an example,the level of
fertilizer application on the majority of the Kenya Seed Company 1976
demonstrations in Siaya was 500 kg/ha single superphosphate and 250Kg/ha
ASN. At current prices the cash outlay for these applications is
KS 1,100.00 per hectare. This is something over 4x the average total
working capital outlay of the 46% of farmers current purchasing inputs
in Siaya and is probably close to the Gross Cash income of a significant
proportion of Siaya farms. This level of outlay is covered by an
increment of 18 bags of maize per hectare as a response to the fertilizer
applied at current fertiliser/maize price relationships. Some authori-
ties would not expect much demand for fertilisers from small holders,
particularly in areas of high yield variability unless they could
earn a 50% return on their outlay. (CIMMYT 1976) This implies a yield
increment of 27 bags per hectare in our example. A ratio of 2:1 for
the minimum acceptable return is also quoted, implying, for our.example'
a response of 36 bags of maize per hectare.

It is not so much the profitability levels which concern us here, altho-
ugh these require careful evaluation over a number of seasons, it is the
high capital outlay. By far the majority of Siaya farmers do not have
anything like these cash sums available. The little working capital they
do layout is mainly on machinery and labour hire for long rains planting
and weeding. Nor does the Siaya farmers' situation lend itself to a
S credit scheme for fertilizer purchase.

1. Most crops sales are surplus to food requirements and sold onto the
local market. The cost of credit administration in terms of manpower
and funds, particularly in controlling repayments in this type of marke-
ting situation, is prohibitive when aiming to reach the majority of
Siaya farmers.

2. The risk preferences of even the minority of Siaya farmers currently
using working capital are unlikely to encourage them to double their
present cash outlays. Even this would only provide fertilizer and seed
for a half acre of improved maize at the rates of application example.

These circumstances have important implications for an adaptive maize
research programme for the zone. Fortunately short-term varieties are
usually more efficient in converting plant foods absorbed to grain and
should respond better than longer term varieties to low manuring levels.
The selection of shorter term varieties for the zone should be done
under a low level manuring regime which a majority of Siaya farmers can
be expected to achieve in the near future. Animals in Siaya are esti-
mated to produce about 2.5 tons of farmyard manure per year per hectare.
In all farmer populations there are a significant proportion of late
adopters, perhaps a level equivalent to 4 tons of farmyard manure per
hectare per year could be considered as a guide to the availability of
manure with say 5 tons per hectare per year a constraint on the levels
examined in experimentation.


Due to the timing of the farmer survey, after farmers had harvested
"" their maize, observations on plant populations and spacings are limited.
No data was obtained on the spacing of intercrops. The limited data
collected during the survey refers to 119 fields which had carried
maize in the 1976 long rains this is compared to data from Gerhart
(1975) and from Hesselmark (1976).

1. Present practice

71% of the maize grown was intercropped (110.7 ha out of 155.3 ha).
29% was grown in pure stands. This is exactly the same proportion
recorded by Gerhart in 1973. 52% of farmers (62) made some pure stand
plantings and a further 21% (26) made plantings of maize mixed with
sorghum and no other intercrop. Of the 119 maize fields examined (after
harvest) 48% were row planted, 29% were hill planted that is the seed
was placed but not in rows and 23% were broadcast. Gerhart recorded
47% of farmers row planting in his study.

Table 18 sets out the data obtained on the spacing being used and the
resulting plant populations for row planted and broadcast maize. Weighted

Table 18. Data obtained on plant populations in maize

Sub:location Omia Nyawita Kambare Simenya Total

Row planting
no. of observations 16 24 17 13 70
Average interrow(cms) 85.0 80.0 80.0 87.5 82.5
Average interplant(cms) 40.0 35.0 55.0 29.5 40.0
Plant population (ha) 29,400 35,700 22,700 38,750 30,300
no. of observations 11 4 9 3 27
Plant population (ha) 29,400 10,000 18,550 23,300 22,200

by the number of observations (assuming these reflect the proportions
planted in rows and broadcast) the average number of plants per hectare

recorded was 28,045. This is close to Hesselmarks (1976) figure of
28,200 for the whole of his Zone 3, and is equivalent to 11,200 plants
per acre. Usually 2-3 seeds were planted per hole,.47% of these fields
were thinned and an average of 1.6 plants found per surviving stand.

2. Implications for a maize research programme

Intercropping still predominates in the management of the maize crop.
It is known to produce a higher value of crops than pure stands in many
traditional farming systems. It is recognized as having value in insuri-
ng a return to resources employed in cultivating planting and weeding
by mixing in crops less susceptible to hazards, in the case of Siaya,
poor rainfall conditions. Despite the predominance of intercropping a
significant proportion of Siaya farmers, 73% if those mixing only maize
and sorghum are included, plant in pure stands. Research on intercrop-
ping is highly complex, the extra variables greatly increase the facto-
rial permutations and thus the experimental work and the time required.
The methodology associated with intercropping experimentation needs
further development, and, unlessmonocropping is clearly unacceptable to
farmers, short term adaptive research can more easily be undertaken on
this basis. In the Siaya zone, despite the high level of intercropping,
a high proportion of farmers are already planting maize in pure stands
and payoffs to research on a pure stand basis will be more readily

Despite a continuing prevalence of intercropping maize populations in
the area are quite high, averaging over 28,000 plants per hectare.
There is no reason why the close spacing usually associated with recom-
mended practices should be difficult to accept for Siaya farmers. How-
ever, given only a low level of manuring is feasible,the interactions
between the level of fertility and plant population will be an important
aspect of the agronomy work,.

It was reported from discussions from farmers that there is an increasing
adoption of row planting. Under increasing pressure from a labour bott-
leneck during the weeding period farmers have begun to appreciate that
row planting facilitates rapid weeding. Although presently only 48% of
fields are row planted with a further 29% hill planted, only 23% of
fields are now broadcast. Present practice does not place any obvious
limitations on the planting techniques which it may be desirable to
incorporate in experimental work.


1. Present time of planting of maize

Two sets of observations were obtained for the time of planting of maize.
Each maize planting was recorded for each farm and also planting time on
the 119 selected fields were recorded. Table 19 sets out the incidence
of all maize plantings and the proportion of area planted in each month.
Out of the 119 fields examined in more detail 33 or 28% had to be repla-
nted, these were all completed before the end of April. As has already
been intimated farmers appreciate the value of early planting. It is
beneficial to all the crops they grow. 28% of farmers had started culti-
vation in December and 47% by the end of January in order to be ready
for the rains. However it is not feasible for them to plant all their
crops at the onset of the rains, nor, given frequent false starts to the
rains is it desirable. Priorities competing for the limited labour and
draught resources ensure that planting will be stretched over a two month
period on most farms. This staggering of planting helps to stagger

Table 19. The incidence of maize plantings on sampled farms

MONTH Feb Mar Apr.. May

Number of plantings 52 66 35 16

% of area planted 33 40 21 6

demands for labour for other operations, in particular weeding, and
allow the farmer to manage a larger area of cultivated land. Given the
likely trend towards more hand cultivation the spread of planting is
likely to be wider as more labour is required for the hand operation.
-. Further, as hand cultivation on the heavier soils is very difficult to
do before the ground is thoroughly wet., the trend may be towards later
planting in general.

2. Implications for the maize research programme

It is because this resource constraint in seed bed preparation and
weeding is very widespread in small family farming that time of plant-
ing components have been a great stumbling block in extension recomme-
ndation. Analysis of the farming system has suggested that risk of
crop failure in maize, and therefore the use of resources to insure
against such failure can be reduced by the use of a shorter term maize.
Given local rainfall conditions a 95 day variety will increase relia-
bility in yield levels and improve the flexibility of action available
to managers. Both these desirable features would be further enhanced
if the chosen variety was insensitive to the time of planting effect.
It is recommended that tolerance of late planting is a further, supp-
lementary selection criterion.


1. Present practice

As has been discussed in section 2.5the evidence, both the subjective
responses of farmers and their actions in supplementing family labour
by hired resources, indicates a labour peak at a time when the weeding
of early crop plantings and the establishment of later plantings over-
lap. March and April were indicated as the busiest period of the year
for farmers in the Siaya zone. In this context all farmers weeded
their maize once and 61% of farmers weeded it a second time. Distri-
butions show the timings of first and second weeding on 119 fields
of maize in Table 20. (Some farmers mentioned two months in responding)
A short, 2 ft long hoe, weighing some 1g kg is used for weeding.

2. Implications for a maize research programme

The fact that labour for weeding, and capital for hiring supplementary
labour for weeding are in short supply in the system has important
implications for a maize research programme. It is true that higher
yields will allow food maize requirements to be met off a small area
planted. However, because the farmer is averse to risk he is unlikely
to reduce his planted area in anticipation of higher yields. He will

Table 20. Distribution of first and second feedings on
119 fields of maize

MONTH Feb Mar Apr May Jun

Incidence of 1st weeding 2 32 60 23 8

Incidence of 2nd weeding 1 15 44 18

continue to cultivate the same area at least until he has been assured,
by experience, that his yields are higher and/or more reliable. Fur-
S their, in the case of the Siaya farmers; surplus food crops being their
-, major source of cash income, they are likely to maintain their planted
area regardless of yield levels, to reAlise larger cash surpluses.

The effects of asking Siaya farmers to increase their weeding inten-
sity as part of recommendations for improved maize management are
likely to be counter productive; either they will be pressured into
cultivating reduced areas or into using scarce capital funds to hire
extra labour for weeding. Any maize improvement programme would hope
to see those scarce funds moving into the purchase of manure or ferti-
liser. In the medium term the enhanced flexibility in management pro-
vided by the short term varieties will allow a further spreading of
the demand for labour resources, at this stage increased intensity of
weeding may be feasible. In the shorter term it is recommended that
weeding levels in the experimental programme are held to those now bei-
ng achieved on Siaya farms; a maximum of two weedings in the two month
period after planting. At the same time varieties should be screened
for weed suppression from rapid early growth and large leaf areas.


A five year programme with the probability of initial recommendations
emerging after three years work might have three overlapping phases.

Variety selection over years 1, 2, 3
Agronomy work on the three best
adapted varieties over years 2, 3, 4
Management refinement over years 3, 4, 5.

The programme assumes effective administration and supervision and would
be multi-locational within the zone on soils typically used by farmers
for growing maize. The first phase of variety selection, would be wide
based in Year.l. with perhaps 10 sites across the range of rainfall
found in the zone; 800 1500 mm, treating sites as replicates. It
would be repeated in the long and short rain seasons. In the second
year repetition of the variety selection trials would be limited to per-
haps five sites and factorials would take place at the same sites with
perhaps two replicates at each. Refinements, beginning in year 3, would
be done at a single site which is relatively easily administered.

1. Variety selection

Readily available white composite and hybrid varieties with maturity
periods of between 90 and 120 days in the Siaya area would be compared
with the local variety in a three year programme covering both long and
short rain plantings. This programme would be implemented at a mana-
gement level as close as possible to that now being achieved by farmers;
time of planting in mid March, 2.5 tons of FYM per hectare per year, a
population of 28,000 plants per hectare and a maximum of two weedings
in the two months following planting. Recording should cover rainfall
data at each site, physiological features including speed of growth/
degree of ground cover, time of tasselling, time of grain setting and
time of grain harvest, also the incidence of pest and disease attack in
the field. Major selection criteria will be yield of course, and relia-
bility of yield using inter site data, comparisons between long and
short rains and inter year data for assessment, 20 observations on each
selection would be available at the end of the first year from long and
short rains plantings at 10 sites. Parallel with the first years field
work checks should be run among the farm population on palatability and
ease of processing. Checks should also be run on the resistance to
storage pests of the entries in the variety trials. These checks, toge-
ther with records on in field pest and disease attacks would serve as
supplementary selection criteria.

2. Agronomy work; varieties and management

The best three varieties should be selected and used as one factor in
two sets of factorials, each set examining two management factors in
relation to the varieties. If the local variety is only marginally
inferior to the third best selection it should be included as a variety
in the factorials rather than this selection. Entries may differ at
higher and lower rainfall sites. The varieties entered may change in
years 3 and 4 as a result of the additional experience with varieties
in the selection trials. If entries are changed in year 4 it will be
desirable to continue the agronomy work into the 5th year.

(a) The first set of factorials should examine the interactions between
PYM use at three levels; perhaps 0, 2.5 tons/ha and 5.0 tons/ha, plant
population at three levels; perhaps 25,000, 35,000 and 45,000 per hec-
tare and the three variety selections. In this set time of planting
should be held at the median time at which farmers are now planting;
say mid March. Similarily weeding should be limited to twice at inter-
vals of 3 4 weeks.

(b) A second set of factorials should test the three selections for
time of planting effect with three treatment levels of say 1st and 20th
of March and 5th April. Also for weed tolerance with perhaps 1, 2 and
3 weedings within the first six weeks of growth. In this set plant po-
pulation and manuring should be held at farmers levels; say 28,000
plants per hectare, and 2.5 tons/ha PYM.

There should be no non-experimental treatments which are not consistent
with present farmer practice in any of the selection or agronomy work.

3. Management refinements

As a result of the farming system analysis more detailed work is justi-
fied on three aspects of management. Clues ori the orientation of this
work should emerge from the initial agronomy work in year 2.

(a) Examination of the optimal timing of the limited weeding input the
farmer can manage.
(b) Examination of the rates/frequency/placement alternatives in the
use of FYM within the bounds of the 4 tons/ha/year judged to be availa-
ble to farmers.
(c) Examination of the response to artificial with a budget constraint
of about S 400.00 per hectare. Lower levels of inorganics will probably
be found to be consistent with the shorter term varieties being advoca-
ted and the later planting dates.

From the results of three years work; given positive results emerge, it
should be opportune to mount a set of demonstrations in farmers fields
comparing the local and selected varieties with and without manure.
In the demonstrations the selected varieties should have the benefit of
findings on plant population and the timing of the weedings. Local
varieties should be demonstrated under farmers' management levels.



The report documents the findings from a demonstration of a set of pro-
cedures designed to allow an interdisciplinary approach, marrying the
skills of crop scientists and economists, to adaptive research planting.
The philosophy behind the approach is that adaptive research efforts
should be based on an evaluation of the priorities of local farmers and
of the constraints within which they are obliged to operate. Only such
a basis can produce recommendations which will be acceptable to and
absorbed by farmers in the area the research is serving. In these con-
cluding paragraphs of the report we look at some of the problems which
arose in implementing the demonstration and some possible modifications
in the methodology to suit Kenyan circumstances.


The essence of the methodology is a team approach to the identification
of farmer problems and of possible solutions to the problems. For the
demonstration we liaised closely with the Research Division, the Mini-
stry of Agriculture, we received considerable help, in the field, from
Sthe extension staff of the Ministry of Agriculture following support of
the project by the Director of Agriculture, and we had first class
support in everyway from our associates in the project; The Department
of Economics and Farm Management, Egerton College. Yet, due to the
geographical and administrative isolation of these various units, it
would be difficult to consider them as a team in the sense of the word
envisaged by the approach. It is considered that much greater insights
into farmers needs and problems would be possible with a team of scien-
tists operating at the level of the local research centre, with a day
today interest in a coordinated effort to solve local problems. This
being the first demonstration a closer coordinated team effort was dif-
ficult. Two changes would considerably enhance further use of the
approach in Kenya.
1. The establishment of a post for an economist within the Research
Division of the Ministry of Agriculture.
2. An appointment of a team, including a breeder, agronomist, a plant
protection worker an extension worker and an economist, as far as pos-
sible with existing, local responsibilities, by the Director of
Research to implement the approach in areas selected for further


The major component of the approach is the farmer survey. Its impleme-
ntation is the responsibility of the economist and takes about 4 man
months of his time. Other professionals are involved in orienting the
survey work and interpreting the results, their commitment involves
only 10 15 days over a six month period. Such a limited commitment
is unlikely to interfere with their existing programmes and could rea-
sonably be undertaken by existing personnel.


A great deal has been learned in the course of implementing this, the
first, demonstration of this approach to adaptive research planning in
Eastern Africa. Three major points have already aided the orientation
of other demonstrations currently underway in Tanzania and Ethiopia.

1. The balance of effort between pre-survey and farm survey can
usefully be modified. A longer pre-survey period using a more carefu-
lly designed approach to farmers, .allots a much closer focussing of
the survey questionnaire, a shorter formalised interview during the
survey, a faster rate of work and a cheaper operation.

2. The documentation of factors in the environment seen as hazards
by farmers, and the documentation of their management alternatives in
the face of such hazards, gives very useful indications of the criteria
likely to be important in designing an adaptive research for the area.
These areas of investigation deserve more emphasis in future work.

3. A post survey visit may be just as valuable as a pre-survey
visit in verifying conclusions reached from analysis of the collected


CIMMYT wishes to acknowledge the invaluable assistance received from
Messrs Lugogo and Machooka, Department of Economics and Farm Management,
Egerton College, and from the final year farm management students at
the College who interviewed the farmers covered by the survey. CIMMYT
also wishes to thank the staff of the Research Division, the Ministry
of Agriculture for their close interest and the Director of Agriculture
and his staff at Provincial and District level for their support for
the work.












(1975) The diffusion of hybrid maize in
Western Kenya.

(1974) The 1974 Kenya Maize Farmer Survey.
(Maize and Produce Board)

(1976) The 1975 Kenya Maize Farmer Survey.
(Maize and Produce Board)

A Study of the Climatology of the
"Highlands of Eastern Africa.

(1975) Soil Physical Studies report: in:
Maize Agronomy Research Project
(O.D.M. Scheme No R2989) Annual
Report 1975.
Investigation of the important growth
stages of maize.

(1973) Altitude and environmental responses
of entries in the 1972-73 East
African Maize Variety Trial.

(1974) Agriclimatic Isohygromenes of Kenya.

(1970) Agricultural Survey of Bako, area

(1976) From Agronomic data to Farmer
recommendations -'An Economics
Training Manaual.