The team for this demonstration was:
From the Agricultural Research Institute, Ilonga:-
Breeder D. Sperling, Plant Protection A. Mushi,
Agronomy F. Brockman, Soils J. Nyonyi,
Sorghum and Millets G. Mitawa.
From the Department of Rural Economy, Faculty of Agriculture:-
Economists: C. Bartlett, E. Manday.
The crop scientists guided the formulation of crop
management questions for the farmer survey organised
by the economists. The economists evaluated the data,
posed some possible economic solutions and discussed
these with the crop scientists who evaluated their
technical feasibility and reviewed the research proposals
CONTENTS Page No.
1.1. Background ---------------------------------------------------- 1
1.2. CIMMYT: Eastern African Economics Programme ------------------- 1
1.3. The first Tanzanian demonstration ----------------------------- 1
2. THE ZONE: SOME DESCRIPTION
2.1. Reasons for selection ---------------------------------------- 3
2.2. Rainfall in the zone--------------------------------------
2.3. Soils in the zone -------------------------------------- -- 4
2.4. Population and Settlement ---------------------------------- 5
2.5. Infrastructure and Economy ------------------------------------ 6
2.6. Markets and distribution -------------------------------------- 6
3. THE FARMING SYSTEM
3.1. Farmers' Priorities and Preferences --------------------------- 9
3.2. Crops, varieties and their end uses --------------------------- 9
3.3. The cropping pattern in the zone ------------------------------ 11
3.4. The food and cropping calendars ------------------------------- 12
3.5. Crop production methods --------------------------------------- 13
4.1. Land Availability --------------------------------------------- 16
4.2. Land Use ------------------------------------------------------ 16
4.3. Labour Availability ----------------------------------------- 17
4.4. Labour Use ---------------------------------------------------- 17
4.5. Sources and use of working capital ---------------------------- 18
5. FARMING HAZARDS
5.1. Introduction -------------------------------------- -------- 21
5.2. Identified hazards in the zone -------------------------------- 21
6. THE INFLUENCE OF HAZARDS AND RESOURCE LIMITATIONS ON MANAGEMENT
6.1. Introduction ------------------------------------------- 26
6.2. Insurance embodied in the cropping pattern ----------------- 26
6.3. Manipulation of the crop calendar ----------------------------- 27
6.4. Crop husbandry ---------------------------------------------- 28
6.5. Risk of cash losses ----------------------------------------- 30
6.6. Decision making ----------------------------------------------- 30
7. FARMERS PROBLEMS AND THREE CONTRIBUTORY SOLUTIONS
7.1. Introduction ------------------------------------- 32
7.2. Starch staple supplies in 1976 -------------------------------- 32
7.3. The search for appropriate solutions ------------------------- 33
7.4. The introduction of a short-term (85-90 day) maize ---------- 33
7.5. Improvements in Sorghum production -------------------------- 36
7.6. Wider introduction of appropriate cassava varieties -------- 36
7.7. The choice among these alternatives --------------------------- 38
8. A MAIZE RESEARCH PROGRAMME FOR FARMERS IN THE ZONE
8.1. Introduction ------------------------------------- ----- 40
8.2. Experimental background -------------------------------------- 40
8.3. Variety selection --------------------------------------------41
8.4. Variety screening and comparison ----------------------------- 41
8.5. Agronomy work ------------------------------------------------- 42
9. SOME COMMENTS ON THE DEMONSTRATION
9.1. The team approach ------------------------------------------- 46
9.2. Survey problems ------------------------------------------- 46
9.3. Logistics and Computor processing ----------------------------- 46
9.4. Conclusion ---------------------------------------- 46
ACKNOWLEDGEMENTS ------------------------------------------------------- 47
REFERENCES ------------------------------------------------------------ 47
The Research section, Crop Development Division, the Ministry of Agriculture,
Tanzania and the Department of Rural Economy, Faculty of Agriculture, University
of Dar-es-Salaam cooperated with Cimmyt in the demonstration of a farming
system based approach to planning adaptive agricultural research in the drier
parts of Morogoro and Kilosa Districts. Chapter .1. of the report describes
the Cimmyt programme and the background to this first demonstration in Tanzania.
Chapter .2. describes the natural and economic characteristics of the zone in
which the demonstration was mounted. Chapters 3-6 describe farming in the drier
areas of Morogoro and Kilosa Districts.
Farming in the zone.
Farmers in the zone put priority on maintaining a reliable and palatable supply
of foods for their families. Preferences are for maize garnished with cowpea
leaves, pods and seeds. Beans and other green vegetables, such food crops
occupy 74% of the area cultivated. Cotton is the major cash crop contributing
30% of farmers' cash earnings. Maize surplus to household requirements is
second in importance contributing 12%. Off farm labour, Crafts, trading and
Relatives working away contribute 32% of farmers' cash income. Average farmer
cash incomes are of the order of shs.900/- per family per year.
Crops are planted from December through to March and these months are also the
most difficult for food supplies, particularly after a season of low rainfall.
Land is readily available for farming in the zone with some 20 persons per
square kilometre. However farmers' are selective in their use of soils and there
is a relative scarcity of the better soil types and lowlying locations valued
for their moisture retentiveness. Labour and cash to command hired services
limit the scale of farming. The peak period of labour demand is January -
March while land is being prepared and the newly established crops are weeded.
40% of farmers hired tractors and 27% hired labour. 65% of all cash spent on
inputs was used to relieve labour bottlenecks in these ways. Farmers identified
uncertain rainfall as the most important hazard in farming in the zone, this is
evidenced by probability estimates of monthly rainfall from stations in the
zone. Farmers priorities for a secure food supply, their limitations of labour
and working capital and their need to guard against the consequences of rainfall
failure, dictate their decision making and present management strategies.
A framework for a maize research programme for the zone
Farmers in the zone will be particularly interested in changes in management
which ensure their supplies of the preferred starch staple maize. To be
acceptable changes should be capable of being implemented with the labour supply
and working capital at the disposal of the areas' farmers. Chapter .7. of the
report sets out three potential improvement strategies:
(1) The introduction of a maize tasselling in 45 days.
(2) Changes in sorghum varieties.
(3) Wider introduction of appropriate cassava varieties and management
The introduction of a short term maize variety into the existing farming system
would have many benefits (Page 34).
(1) It would make best use of the reliable period of rainfall from
mid March to mid May, giving a more reliable source of the
preferred grain staple.
(2) It would enhance the ability of the farmer to react to contingencies
of rainfall uncertainty.
(3) It would meet the need for early new food more effectively, giving
cobs 2-3 weeks earlier than existing varieties.
(4) It would release lowlying areas 2-3 weeks earlier than present
varieties which would:
(a) Increase the probability of a successful bean crop following
the maize in these areas.
(b) Increase the number of lowlying areas where a bean crop could"
(c) In the better lowlying areas it opens the possibility of
growing other crops with a slightly longer season than beans.
(5) The later planting time for the main crop will spread the existing
labour peak allowing an increase in the scale of the farmers'
(6) The increase in reliability of maize as a source of starch staple
will reduce the need for sorghum as an insurance crop, allowing
re-allocation of resources to higher value cash crops.
A maize taking 45 days to tasselling is technically feasible in the zone
and an experimental programme is designed around such a variety, to take
advantage of the potential benefits outlined and yet to be within the
farmers' capacity to absorb. The recommended experimental programme (Page 40)
focuses on variety selection and screening, with plant density and spacing,
weeding intensity and the use of limited levels of purchased insecticide and
and fertilisers as the experimental variables in the agronomy work.
The National Crop Research Planning Committee, meeting in Mbeya in October 1976,
gave approval to CIMMYT to discuss the demonstration of an inter-disciplinary
farming systems based approach to planning adaptive agricultural research with
the Directors of Agricultural Research Institutes in Tanzania.
1.2. CIMMYT: Eastern African Economics Programme
CIMMYT is urging a cooperative effort, between biological scientists as experts
in crop potential, farm economists as experts in farmers' priorities and circum-
stances, in the planning of adaptive research. Much has been heard of the need
for technology which is appropriate to the farmers' situation. It is particularly
important in the case of small farmers who usually have their improved technology
selected for them by governments. An edifice expensive of both financial and
manpower resources, often covering extension services and training, credit
schemes, input supply, processing facilities and infrastructure, is built around
selected technology as the core of a development programme. Unless the techno-
logy selected is appropriate to their circumstances it will be unacceptable to
farmers and both the research effort in designing it and the development effort
in promoting it will have been wasted.
CIMMYT biologists and economists discussed the organisation of a cooperative
effort in planning adaptive research programmes as procedures were tested in a
number of pilot studies. By 1976 CIMMYT felt sufficiently confident of the
procedures, and convinced of the importance of more relevant technological
packages, to focus the Eastern African Economics Programme on this work. The
programme with a mandate to operate in six Eastern African countries has two
(1) Wherever research administrators manifest interest the programme demon-
strates a set of procedures which allow a cooperative interdisciplinary effort
in research planning.
(2) If research administrators are convinced the approach can improve the
efficiency of their research planning, CIMMYT will advise on how, and how quickly,
the approach could be integrated with present research planning procedures.
Given CIMMYT's mandate, its main interest is in helping research administrators
improve the effectiveness of adaptive research in maize and wheat. However the
approach is essentially system oriented and can be used to design research work
on any crop grown in the farming system under investigation. For example in
the case of the drier areas of Morogoro/Kilosa problems are identified at the
system level and several potential solutions are outlined. The detailed maize
research programme aims to provide an appropriate solution to the identified
problems; that is a solution within the reach of, and consistent with the
priorities of, the majority of farmers in the area.
1.3. The first Tanzanian demonstration
A meeting was held in November 1976 at ARI Ilonga between Dr. Haulle, the ARI
Director and his staff, Dr. Bartlett and Mr. Manday, economists of the Depart-
ment of Rural Economy, Faculty of Agriculture, University of Dar-es-Salaam
and Dr. Collinson from CIMMYT. The meeting expressed interest in the
approach and an interdisciplinary team was nominated. Crop scientists from
ARI Ilonga would provide the biological input and economists from the Department
of Rural Economy and CIMMYT the economics. It was agreed that a demonstration
of the approach be mounted in the drier areas of Morogoro/Kilosa Districts. A
proviso made by the meeting was that the involvement of the crop scientists in
the demonstration should not encroach on their existing commitments.
The sequence of procedures followed in implementing the demonstration is given
(1) November 1976: Initial meeting to nominate the team and agree the drier
areas of Morogoro/Kilosa for the demonstration.
(2) Identification of the zone as the Eastern parts of Morogoro District and
western parts of Kilosa District. Later limited to areas receiving an average
annual rainfall of under 1000 mm. (see map)
(3) November 1976: Pre-survey of the zone.
(a) A description of the local farming system to allow evaluation of
farmers' priorities and resource constraints and the way these are
reflected in their cropping pattern and management practices.
(b) Checking the similarity of farming through the areas included in
(c) Preparing a sampling scheme and possible field organisation.
(4) December 1976: Team discussions on
(a) Present methods of maize growing in the area.
(b) Problems of maize growing under climatic conditions of the area.
(c) Useful questions to be explored with farmers in the course of the
(5) December 1976: Questionnaire design based on Pre-survey and Team
(6) January 1977: Field survey of 107 farmers from six villages in the zone
(see map) with 20 students from Faculty of Agriculture as enumerators.
(7) March May 1977: Computor processing of data, data analyses.
(8) June 1977: Report outline prepared.
(9) September 1977: Team discussions on:
(a) Identified problems of the area's farmers.
(b) Economic conditions of farmers as a framework for appropriate
(c) Potential solutions to identified problems; their technical
(10) September/October 1977: Report drafting.
(11) October 1977: Team discussions on draft and on skeleton research proposals
for maize improvement for farmers of the area.
(12) November/December 1977: Report finalisation and submission to Chief
Research Officer, Ministry of Agriculture, Tanzania.
Some detailed comments on lessons learned from the implementation of the
demonstration are made in the final section of the report. It should be
.mentioned here that implementation was unduly prolonged for two reasons: This
was the first demonstration for which data processing was computorised, and
some initial snags had to be overcome. Dr. Bartlett was out of action through
July and August with hepatitis. Also, despite good will on all sides, the degree
of integrated team effort envisaged in the approach cannot be achieved by an ad
hoc grouping of professionals from physically disparate institutions. The
approach calls for a team working within the same terms of reference from a
research centre allowing interdisciplinary interactions on a day to day basis.
2. THE ZONE: SOME DESCRIPTION
2.1. Reasons for selection
The primary reason for selecting the drier areas of Western Morogoro District
and Eastern Kilosa District was convenience. It is adjacent to both the ARI.
Ilonga and the Faculty of Agriculture, Morogoro. The proximity of the area
minimised costs of transport for the demonstration and facilitated contact
between team members working in different institutions. With most farmers in
the area growing maize it is also consistent with CIMMYT's mandate for the
improvement of maize and wheat production.
2.2. Rainfall in the zone
Rainfall is the main climatic feature of interest in an area bounded by the
1000 mm isohyet. The area's rainfall has several characteristics:
(1) A long dry season of from five to six months varying from year to year,
and ranging between mid May and mid or even late November.
(2) Very variable rainfall in the six wet months, of which only April gives
good rain in all years. In December, January, February March and May rainfall
levels are low in a significant proportion of years.
(3) In the December to February period showers are often heavy but also
(4) There is considerable rainfall variation according to topography, higher
rainfall is associated with ridge-lines both within the zone and on its periphery.
There are few meteorological stations within the 800 nun isohyet. Table .1. giv'H
average monthly rainfall for three stations marked on the map. Morogoro and
Mvomero within the 1000 mm isohyet ARI Ilonga just outside it. The table also
shows monthly rainfall probabilities in four years out of ten, for the three
stations, monthly radiation data for ARI Ilonga (1964-70) and estimated monthly
radiation data for Morogoro based on an annual total radiation of 2000 mm.
SKETCH MAP: DRIER AREAS
OF WEST MOROGORO AND EAST
1000 mm Isohyet
as zonal boundary
Source: Based on maps in the Morogoro Regional Development Plan.
Table .1. Rainfall data for the zone. (mm)
Mvomero station 96.37021, 60 19'S 370 26'E 500 metres a.s.l.
Ann. Av. 919 mm.
Month J F M A M J J A S 0 N D
Average mm 106 87 168 221 92 19 10 6 14 30 60 106
Prob: 4/10 84 48 132 191 75 6 3 3 1 9 23 65
Ilonga ARI. 96.37032 6 46'S 370 02'E 500 metres a.s.l.
Ann. Av. 1071 mm.
Month J F M A M J J A S O N D
Average mm 136 130 216 213 68 10 9 13 17 36 81 142
Prob: 4/10 115 108 192 185 48 5 5 7 5 19 29 116
Rad. EO(mm) 171 149 166 131 122 126 138 148 167 205 201 183
Morogoro Agricultural Off.
96.3700 60 51'S
370 40'E 575 metres a.s.l.
Ann. Av. 907 mm.
The short season and particularly the high
nating influence on farm management in the
from year to year have a domi-
2.3 Soils in the zone
Very little information is available on soils in the zone, an ILACO/NEDCO report
of 1967 covers the Mkondoa Wami basin. Soils do vary a great deal from light
sandy loams in the Mvumi Mvomero area, to thick clayey soils in the Wami -
Dakawa area and alluvial deposits near Kilosa. The variation in soils is not
Month J F M A M J J A S O N D
Average mm 95 102 167 215 92 26 15 11 18 29 61 78
Prob: 4/10 70 73 118 180 77 13 5 6 6 9 32 60
Est: EO(mm) 180 160 180 140 130 130 140 160 180 210 200 190
reflected in the cropping patterns found in sample villages except for the
localised use of low lying valley bottom soils for rice and early plantings of
maize often followed by beans. These special locations are scattered over the
study area and used by farmers wherever available.
A soil analysis was recently carried out and reported by Uriyo and Singh (1976),
extracts from the results for sites within the zone are summarised in Table .2.
T e .. Some chemical Properties of soils in Morogoro/Kilosa
low rainfall areas
e Soil Organic Total Available Base (%)
PH C N C:N P Saturation
Wami 6.8 2.07 0.13 20 29.0 71.0
Ilonga 6.3 2.13 0.10 21 25.8 73.5
Mvumi 6.0 1.68 0.12 14 30.0 53.0
Humbeti 7.2 2.41 0.12 20 34.0 99.0
Mvomeru 7.0 1.86 0.15 12 34.0 92.0
The results reported support the view that soil quality is good for crop produ-
ction. The pH range of 6.0 7.2 is favourable for most crops grown in the area.
The C:N ratio reported suggests favourable mineralisation will take place and a
release of Nitrogen during the early part of the growing season. Subsequent
applications of Nitrogenous fertilisers may however give a positive response.
Available Phosphorus ranges from 26-34 ppm. Under normal conditions a P value
of 25 ppm is considered critical. Below this value applications of phosphates
would give a positive response, above it little or no response can be expected.
The results indicate that farmers in the study area may not benefit from the use
of phosphatic fertilisers. The percentage base saturation reported seems
favourable for good crop production except for soils around Mvumi where basic
dressings might raise it above a desirable 70%. The National Maize Research Programme
Village Trials support the soil analysis with applications of Phosphates eliciting
little or no response.
Extrapolating at a rate of 3% per annum from the 1967 census the estimated
population of Morogoro District in 1977 was 417,000 and of Kilosa District &
260,000. An average household size of 4.1 gives a combined household total for
the two Districts of 165,000. Land area in the two Districts is 33,541 sq.km,
that is about 5 households and 20 persons per square kilometre. This is a low
density and an indication of a low man/land ratio; approximately 20 hectares
per family in the Districts. Since villagisation in 1975 the population of the
Districtshas been concentrated into some 300 350 villages. The settlement
pattern, formerly dispersed with families living on their own holdings, is now
nucleated. In Morogoro District the population is concentrated into the highland,
higher rainfall, areas. With more than 1000 mm annually these occupy 30 40%
of the land area and carry 75% of the population. Settlement is more evenly
distributed in Kilosa District. Very few villages in either District are sited
below the 800 mm isohyet. A very rough estimate of the population in the areas
of the two Districts below the 1000 mm isohyet, that is our study area, is
260,000 in 46,000 households nucleated in 130 villages.
Most of the population of the zone are known as Lowland Luguru or Kaguru. However
immigrants from other area- particularly Makonde from as far away as Mozambique,
Nyamwezi and Nyakyusa have been employed in the sisal estates and have intermarried
with local people who no longer have a clearly separate identity. Social tradi-
tions, including eating habits, are fairly homogenous throughout the study area.
2.5 Infrastructure and Economy
Morogoro is the only significant con drbation, located on the Eastern edge of
the zone at the foot of the Uluguru mountains. As a meeting point Cor communi-
cations from West, North West and Southern Tanzania it has recently attracted
new industries which have served to cushion the unemployment consequent on the
* contraction of the sisal industry in the late 1960's. The state farm sector
still has substantial holdings in the zone, both in sisal and beef ranching.
A small but expanding proportion of the zone population is employed on these
large scale farming enterprises and in the processing and small manufacturing
industries in Morogoro town. The large majority of the population in the zone
are small farmers growing subsistence and cash crops.
The central railway line runs through the zone and major highways from Dodoma
and Iringa pass through to meet at Morogoro en route for Dar-es-Salaam. Feeder
roads in the zone are on the whole badly maintained, especially the Morogoro/
Kilosa link through the Mkata plains which is often impassable in the wet season.
2.6 Markets and distribution
Each village has a market but only one of the six study villages reported a
regular market day: Mvomero which has a market day on Thursdays. A significant
amount of produce changes hands within the village. All major crops are sold
through government established channels. These are parastatals organizations
with an active buying season from July to March of the following year. Tanzania
Cotton Authority (TCA) is the buyer of cotton, a major cash crop in the zone.
It also acts as an agent for the National Milling Corporation (NMC) and General
Agricultural Produce Export Corporation (GAPEX) buying all crops in Kilosa
District. NMC purchases in Morogoro District acting there as agent for TAC and
GAPEX. Poor feeder roads inhibit market activity for much of the year.
Official prices are announced prior to the buying season for all produce, the
same prices apply nationwide. Official prices for the 1976/77 buying season are
set out below. The same channels will also sell grains and beans and re-sale
prices are included in Table .3. Buying price is paid into store.
A significant amount of produce, especially maize and paddy but also legumes,
changes hands through unofficial channels at 150 200% of the official buying
prices. There are strong seasonal price variations in these unofficial markets.
It was reported that maize prices use rose to shs.200 per bag for sale to urban
areas, and shs.150 per bag in the rural areas between October and the start of
the new season harvest in April/May the following year.
The channels for the distribution of farm requisites are government controlled.
Two input distribution programmes are active in the zone:
(1) Cotton input distribution. The TCA handles the distribution of requisites
for use on cotton; the main items are seed, distributed free, fertilisers,
insecticides and sprayers for their application. Distribution is done through
Table .3. Official prices 1976/77 season
* Prices for cassava refer to
it is used for animal feed.
dry and peeled weight.
the buying posts where the crop is marketed by farmers.
Farmers pay cash at
50% of the market rate for items purchased, the balance is deducted at sale of
the crop. TCA officials report a good rate of recovery on input credits.
(2)A National Maize Programme is administered by local agricultural staff now
under the Prime Minister's office. District Extension staff receive their seed
requirement from Tanzania Seed Company and fertilizer requirements from Tanzania
Fertiliser Company. Each village has a small go-down at which point requisites,
including insecticides for stalk borer control, are made available to farmers.
Upto 1974/75 season farmers had been receiving maize seed and fertilisers free
in an emergency programme to boost maize production. With the initiation of the
World Bank sponsored National maize Programme in 1976 farmers were charged
shs.1.00 per kilo of composite seed and 50% of the market price for fertilisers.
In addition to intensifying inputs privately owned tractors provide a hire
service for farmers in the zone.
Product Weight per Grade Buying Selling
bag (kg) price (shs) price
Cotton A 2/05
(per kilo) B 1/00 n.a.
Maize 90 I 72 105/50
Paddy 90 I 90 113/25
Cassava* 50 25
Sorghum 100 95 114/00
Beans 90 I 180 458/85
90 I 17/50 458/85
Bullrush millet 100 I 95 127
Finger Millet 100 I 90 114
As with all areas of the country farmers of the zone have access, through their
villages once these are Registered, to short term loans for financing the
purchase of farm requisites from Tanzania Rural Development Bank (TRDB).
The value of marketed produce sold through the official agents in the six
survey villages provides an indicator of the relative importance of local crops
as a source of cash to local farmers. There is no way to estimate the value
of local intra and inter village transactions. Table 4 summarises the value
of produce marketed for the 1976/77 buying season which closed in March 1977.
Estimated values of produce marketed
Table 4.in six study villages
channels in six study villages
(Source TCA files)
Product Value (shs) Product Value (shs)
Cotton 751,487 Sim Sim 6,022
Maize 113,352 Cashew 2,092
Sorghum 49,070 Kapok 1,991
Sunflower 47,997 Castor 1,800
Paddy 14,201 Pigeon Pea 212
3. THE FARMING SYSTEM
3.1. Farmers Priorities and Preferences
Farmers in the zone place priority on maintaining a reliable and palatable supply
of foods throughout year. Adequate nutrition for their families requires that at
least a starch staple and a legume or vegetable are always available. It is
clear from the cropping pattern followed and their manipulation of the crop
calendar that food sufficiently is the primary objective of farm management in the
zone. With the 5 6 month dry season enough food must be harvested and stored,
at the main harvests from June August, to last the family through until food
from the new season becomes available in April/May the following year. Cash
income is a supplementary objective.
Farmers surveyed showed a general preference for maize as their starch staple
(79% of farmers). A minority of 10% of farmers preferred to eat ugali made
from sorghum and 11% expressed no preference, 16% of farmers prefer to mix
cassava into the flour when eating sorghum based ugali. Farmers expressed pre-
ferences for a wide variety of legumes and vegetables to accompany ugali.
Table 5 summarises the responses.
Table Percentages of surveyed farmers expressing preferences for a
range of vegetables
Cowpea Cowpea Green Cassava Beans
Vegetable leaves pods Gram leaves and
Maize 1st pref. 34 12 6 7 31
2nd pref. 27 22 4 20 16
Sorghum 1st pref. 22 7 9 4 19
2nd pref. 9 14 2 8 19
10% of farmers either did not respond or responded with 'no preference' in
relation to maize ugali, and 38% in relation to sorghum ugali.
3.2. Crops, varieties and their end uses
(1) Maize. Maize is the main starch staple in the zone. 77% of farmers surveyed
reported it as their main source of food. It is particularly dominant on the
Western side of the zone, in Kilosa District, characterized by better rainfall
and few locations suited to rice growing. Two main varieties are grown. The
local variety accounts for 70% of the area grown and is widely planted early in
the season, in January, with the aim of obtaining early new food supplies. It
stayed in the field an average of 41 months, including a drying off period. Ilonga
Composite, the second variety grown, is popular because of its higher yield when
rainfall is good. It stayed in the field for an average of 42 months. Both
varieties have maturity period of about 4 months but are left to dry in the
field. Much of the local variety is eaten green out of the field and is taken
off rather quicker than Ilonga Composite. Seventy three farmers answered detailed
questions on preferences for maize varieties; The only notable difference in
scoring the two varieties was better storage of the local variety. Maize is the
second cash crop in the zone, named as an important source of cash by 44% of
farmers surveyed, as many of 67% of farmers in two of the six survey villages
(2) Sorghum. Three main varieties of sorghum are grown, all tall types.
Mbangala, the main variety, has a hard grain which stores well, it is popular for
brewing and acceptable in ugali. There is a good local market for the sale of
the variety for brewing. Mbangala stayed in the field for an average of 51 months.
Jebele is a variety with a softer grain which stores poorly. It is popular for
eating fresh out of the field because of its particularly sweet taste. It stayed
in the field an average of 61 months. The third variety known as Nailon is
shorter term, taking some 4 months to maturity. It is grown by farmers as an
early food maturing in May or early June, or as a late, March planted, crop.
Sorghum was sold by 23% of farmers surveyed,rising as high as. 38% in one village.
(3) Rice. Rice is grown across the zone in the wet valley bottoms. It is a
major crop in parts of the Wami valley where suitable locations in the river
flood plain are relatively extensive. However less than 5% of farmers surveyed
reported it as their main source of food. The main variety grown is Supa which
is readily sold. The local variety Kihogo is more difficult to sell but preferred
locally for taste, it is reported to be less drought resistant than Supa. 27%
of farmers surveyed reported sales of rice, the proportion rose to 70% in
Dakawa village in the Wami flood plain.
A very small amount of Bullrush millet is grown in the zone and is used for
brewing. Root crops are not a major source of food although both cassava and
sweet potatoes are grown in small quantities as supplementary starch staples.
Bananas are also grown throughout the zone in small numbers.
(4) Legumes and vegetables. 1976 was a very poor year for legume production.
Farmers indicated a shortage of seeds, a widespread failure of the cowpea crop
and a failure to establish the normally late planted bean crop due to a sudden,
premature, end to the rains. In addition the survey procedure explicitly oniitted
consideration of the household plots which are now seen to be important not only
for the production of cowpea leaves and pods, but also for cassava leaves, early
supplies of green maize and for sweet potatoes.
The main legume crop grown is cowpens contributing 45% of supplies reported by
farmers in 1976, Beans contributed 27%. A wide variety of minor legumes supple-
ment these two; Green Gram, Pigeon pea, Fiwi Bean, Groundnuts and Bambarra.
A little sunflower is eaten as a legume.
(a) Cowpeas: The variety Chora is grown for its seed while Tambaazi is parti-
cularily valued for its leaves, usedas a vegetable. Both varieties are indeter-
minate in growth habit.
(b) Beans: The preferred variety in the area is called'Kenia' of the Canadian
wonder type, a second variety has a smaller red seed and develops a slightly
climbing mode of growth. A third variety Iringa has a streaked black and yellow
seed. There appears to be no difference in storage quality or palatability
between the three varieties. The large red Kenia appears to be more saleable
and this is one reason why farmers prefer to grow it (Bartlett 1977).
All these legumes, and cassava leaves, are used as a relish to garnish the basic
staple dish of ugali.
(5) Cash crops. Crops sales were the most important source of cash income for
86% of the farmers surveyed, and accounted for 62% of cash income received by
survey farmers. Table 6 summarises the number of farmers earning from each of
the five major cash crops and the relative contribution of each crop to cash
earned from crop sales.
Table 6. Crop sales as a source of cash income
Other crops (bananas, coconuts, cassava, mangoes sugar cane and groundnuts)
contributed 7% of cash earned from crops.
3.3. The cropping pattern in the zone
The typical farm unit in the zone devoted 74% of its land to food crops produ-
ction and 26% to production of cotton or sunflower. Food surplus to family
requirements is also sold and, in some cases, family food was sold to meet an
urgent need for cash. Table 7. summarises the estimated crop acreages over
the sample expressed as an average farm, it also shows the numbers of farmers
growing each crop and the average area per grower.
Table 7. The cropping pattern in the zone
Legume crops are grown as intercrops in maize and sorghum plantings and in
household plots as pure stands. Crops may be inter-planted into a ripening or
partially failed crop. Sunflower planted into ripe early sown maize is common.
Two separate crops per year may be grown only on valley land near rivers where
the soil remains moist well into the dry season. Beans, small patches of
tomatoes and onions, cowpea leaves and other vegetables may be harvested from
these areas well into thedry season and are often planted after an early
established maize crop.
Crop Cotton Maize Rice Sorghum Sunflower
% of farmers 75 44 27 23 9
% contribution to 47 19 14 9 4
cash from crops
Sample % Farmers Average
Crop Average growing p.grower
Maize .56 88 .64
Sorghum .28 51 .55
Rice .20 37 .54
Total food grains 1.04 100
Cotton & Sunflower .36 73 .49
Total Crops 1.40 100
Figure .1. Calendar for farmers growing and buying selected foods in 1976
M A M J J A S 0 N D
X Average dates of
0 Average dates of
There are significant variations of emphasis across the zone influenced by rain-
fall and the availability of locations suitable for rice growing. Table 8 shows
the percentage of fields in different crops in the six survey villages.
Distribution of main crop fields in six survey villages
Table 8. (% of village fields)
Village Maize Cotton Sorghum Rice Sunflower Cassava
Mvumi 66 18 10 7
Mangae 18 33 49 -
Dakawa 28 20 15 26 11
Mvomero 34 26 20 11 1 8
Chanzuru 30 15 11 22 3 6
Mamoyo 45 24 8 7 6 3
Total 37 23 19 12 4 3
3.4. The food and cropping calendars
The strong interdependence between the food requirements of the household and
farm management is seen in the agricultural calendar for the zone. Food needs
are a major decision critrion for the timing of crop establishment within the
I iaiework of natural conldiL ions and fiarmii-ir' liabour supplies. Figure .1. sets
out the average dates reported by growers for beginning the harvest and for
finishing home produced foods, also the average dates for start and completion
of purchases by those households with insufficient produce from the farm.
Figure .2. shows the periods over which growers planted and harvested the major
crops, and distinctive crop varieties. Neither figure is complete for legume
and vegetable crops due to the survey emphasis on starch staples. Figure .1.
shows January to May as the difficult period both for starch staples and for
relish crops. Farmers try to minimise the uncertainty of food supplies over
this period by early plantings of maize, beginning to give green maize in
April, and holding reserves of sorghum, mainly Mbangala, in case maize supplies
are inadequate. Families utilise their cassava to eke out supplies of maize
and sorghum from October and a few farmers bought cassava in October to help
prolong their home grown starch supplies. Bartlett (1977) reports that the
late planted bean crop is an important source of relish from September/October
onwards, supplemented by cowpeas and vegetables matured on residual moisture
in the lowlying areas. Although there is variation within the zone depending
on the balance between maize, sorghum and rice, in most areas farmers grow
maize as an early food crop as well as a main staple, they may use rice or
sorghum later in the season helped out by cassava. Milk and meat supply little
food for farmers of the area- less than 3% of farmers surveyed owned cattle, in
each case a single animal, and 7% owned sheep or goats.
Planting of local maize begins at the start of the rains in December,with cowpea
and green gram sometimes grown as an intercrop. Rice is also an early planted
crop. Failure of these early plantings is common due to cessation of the rains,
army worm or flooding of the low lying areas in which they are made. Many
fields have to be replanted. Local maize continues to be planted into March.
Planting of Ilonga Composite is more concentrated in February, it is felt to
Planting and harvesting periods of the major crops.
Crops & Varieties
J F M A M J J A S 0 N D
X Marks the main concentration of plantings
0 Marks the main concentration
I -e--- I
-X- -9----- i-
be less drought tolerant than the Local variety and to need the best of the
rainfall. Jebele sorghum is planted in December and early January to ripen and
be used fresh at the beginning of July. Mbangala the main sorghum crop is
planted late January and February. The distribution of main crop pantings is
summarized in Table 9 below.
Table 9 Percentages of reported plantings on the major crops;
by fortnightly intervals
Nov Dec Jan Feb Mar Apr
Crop 1 2 1 2 1 2 1 2 1 2 1 2
Maize 8 9 33 2 30 7 11 1 -
Sorghum 8 10 42 8 24 3 3 -
Cotton 4 4 49 23 15 4 -
Rice 6 3 3 25 25 19 8 8 1 -
During December and January planting is intermitant. It will be intensive for
a day or two after rain and then cease. Continuous planting begins in February
when the main rains start. Priority is given to maize, the workload is very
heavy with maize, sorghum and cotton all to be established. Farmers are parti-
cularly active if there has been widespread failure of the earlier plantings.
Some farmers attempt to improve their cash earnings by late planting of sun-
flower after the main labour peak is passed. Beans are planted in May and June
in the valley areas and some maize and cowpeas may also be tried on residual
moisture. Planting of crops ranges over a full six months in the area.
Cowpeas planted in January start to be harvested in early April followed by
early maize initially eaten green. Some green rice may be harvested in May and
eaten after roasting and pounding. The main rice harvest takes place from early
June to .early August. The main sorghum crop is the last of the grains to be
harvested during July and August. Sunflower and cotton are also harvested
during July and August. Harvesting of the indeterminate cowpea varieties planted
between January and May continues through to August. Green grams is also
harvested over a wide period. Beans planted in May are harvested in August and
Pigeon Pea is the latest harvested legume picked in August and September.
3.5. Crop production methods
(1) Land Preparation. Rotation is not practiced on farms in the zone, more
than 80% of 1976 plantings were to be followed by the same crop. Fallowing is
not generally practiced on the flat valley-bottom lands on which most crop
production takes place. Close to the villages hillside land is used, this will
be farmed for a few years and then rested for an indeterminate period. However
only 4% of all fields were reported fallowed in the previous four years. The
land is cleared during October and November, bush growth is cut and burned
together with weeds and stubble from the previous crop. Cultivation of the land
starts with the first rains. Hoeing is the main method of cultivation, though
on a significant proportion of the land hoeing is restrictedLto;scraping off the
weeds. Farmers are making extensive use of tractors for primary cultivation,
particularly in preparing for the maize crop, 40% of farmers surveyed hired
tractors. Table 10. summarises the various methods of cultivation'used by
farmers in the zone on 106 maize and sorghum fields observed in more detail.
Table 10. Methods of land preparation (% fields)
Method Tractor Tractor Scrape Deep
Harrow hoe hoe
Maize 42 4 16 38
Sorghum 9 41 36
Scraping with the hoe is a system of minimum tillage probably motivated by the
urgency with which planting has to be done early in the season in conditions
of intermitant rainfall.
(2) Planting Methods. Maize is planted in rows on the flat as is the majority
of the sorghum; However 40% of the sorghum fields observed were planted in hills
at random. Rise is mainly broadcast. Crops on household plots are planted in
a close random pattern. The seed rate for maize is generally 3-5 per planting
station. While the median rate for sorghum was reported as 5-10 per station,
40% of farmers asked sowed more than ten seeds per hole.
Farmers were asked to report their interrow and interstand distances on maize
and sorghum and the average results are summarised in Table 11 below. This
is the population aimed at and the stands per ha figures in the table reflect
the implicit assumption that plants survive at all stations.
Average spacing and stand populations for maize
Interrow Interstand Stands
Crop distance distance per ha.
Maize 79 53 23,900
Sorghum 90 55 20,200
Farmers were also asked to estimate the plants per stand and these estimates
averaged 2.4 for maize and 4.3 for sorghum. Fifty four of the farmers felt
that higher populations would reduce yields, five specified moisture shortages
and four pest and diseases as the cause of such reductions.
(3) Crop combinations. The timing of fieldwork was such that farmers were
reporting on crops harvested from the fields, in the case of some intercropped
legumes, eight months previously. It is likely that this, together with the
very poor season for both cowpeas and beans with an apparently high failure
rate, led to under reporting of intercrops. However it may be that the area
of intercropped is less extensive than had been thought.
Most crops in the sample taken were planted alone, 30% of maize plantings
carried an intercrop and 36% of the sorghum plantings. In the case of sorghum
the predominant intercrop was maize. Maize and sorghum was the most common
intercrop, followed by maize or sorghum, with cowpeas. In 57% of the reported
cases of intercropping the second crop was planted 1-6 weeks after the main
crop, although in most cases farmers reported that the decision to mix the crops
was iilado belor'e pl- 1l i u I he ma i ii crop. Oni 15% of tlhe intercropped fields tile
decision to intercrop was made after the main crop was planted, indicating it
as a response to changing circumstances, often the poor performance of the main
(4) Weeding and Thinning. Weeding is done with a longhandled hoe, except for
rice in which a small short handled hoe is used. Most farmers weeded twice for
both maize and sorghum, though 32% reported weeding maize and 23% sorghum three
times. The most usual time of first weeding was 1-2 weeks after planting for
maize and sorghum with 25% farmers reporting after three weeks for maize and
23% and 18% reporting after 3 weeks and 4-5 weeks for sorghum. 83% of farmers
reported practising thinning including 84% on their maize crops and 77% on their
sorghum. Observations on fields planted after the survey suggests
that proper thinning is not widely practised and the data on plants per stand
for maize and sorghum would seem to support this impression. The survey data
provides evidence that the time of a tractor gives farmers more timeliness in
establishment vis-a-vis planting rains and better weed control. Table 12 sets
out evidence from 99 farmers reporting on maize management and production.
Table 12. Evidence of advantages from tractor ploughing
Method of Scrape with Tractor Deep
cultivation hoe hoeing
(22 farms) (37 farms) (40 farms)
Number of 2 or 3 or 2 or 3 or 2 or 3 or
weedings less more less more less more
% of each Good or
27 23 51 19 33 20
production Poor 41 9 24 5 25 22
Overall 70% of farmers using tractor cultivation reported good or average results,
compared to 52% using alternative cultivation techniques. Yet 76% of tractor
hiring farmers reported only two weedings compared to 61% of farmers using hand
cultivation. It has been suggested that improved water penetration on tractor
cultivated land will give better establishment, more rapid stand emergence and
so better weed control.
(5) The use of purchased inputs. The use of purchased inputs is dominated by
the cotton crop. Fifty six of the 75 cotton growers used insecticide, that is
52% of the farmers surveyed. These cotton growers tend to use insecticide on
other crops, particularly beans. Four farmers reported purchasing fertilizer;
two used it on maize and three on cotton. Twenty six percent of farmers
reported purchasing sorghum seed. With the exception of cotton insecticide the
level of use of intensifying inputs is low.
4. FARMERS' RESOURCES
4.1. Land Availability
Currently land is not a factor limiting farm production. As we have seen the
man/land ratio is low in the zone with an estimated 20 hectares available per
farm family. Families are on average cultivating about 1.5 hectares each
season giving a ratio of 13:1 between availability and use. However three
factors qualify this apparent abundance- a significant part of the zone is
alienated land and not available for the smallholder farming population,
secondly, and perhaps more important, farmers are selective about the type
of land they use, finally the consolidation of the settlement pattern into
villages has concentrated the population, land close to the villages is
relatively intensively utilised. There is no information on the proportions
of the different soil types in the zone.
4.2. Land Use
It is clear that farmers utilise the dark soils in lower lying areas relatively
intensively and continuously. Exploiting the moisture retentive character of
these locations helps farmers to reduce the uncertainties of the early rains
as well as prolong their growing season. The 327 plots recorded on the 107
farms were visually classified into 23.5% Black heavy soils 69% Black/sandy
soils and 7.5% Red soils mainly Red Sandy soils (6.5%). There were some
discernible preferences for crop location between the heavy and light black
soils- these are Summarised in Table 13.
Table 13. Soil type preferences for selected crops
Crop. (% plantings) Black Black Red
Heavy Sandy soils
All plantings 23.5 69.0 7.5
Sorghums 13.0 76.7 8.3
Local maize 19.0 74.2 6.8
Ilonga Composite 33.3 63.3 3.4
Cotton 29.2 63.8 7.0
Rice 47.5 47.5 5.0
Cassava 18.2 81.8
Sunflower 27.2 72.8 -
Others 35.7 42.9 21.4
Plantings of the minor crops were few but among the main crops there is a
tendency to keep the sorghums off the heavier soils and to put Ilonga Composite
on the heavier land. Rice on the heavier land and cassava mainly on the sands
are more predictable practices. 59% of farmers were operating wholly on sandy
soils, the majority on black sands, 31% had a mixture of heavy black and sandy
soils, 9% were working wholly black heavy soils and one farmer reported only
red soils. Maize research work indicates the red soils as fertilizer responsive
and the black soils as non-responsive, bearing out the farmers' logic.
4.3. Labour Availability
The families surveyed had an average of 5.62 members with a wide spread of
family sizes from 2 to 7 persons. Most families had 2, 3 or 4 members aged
over 15, with an average of 3.36. Seventeen households reported 24 family
members working permanently off the farm and therefore not available for farm
work. Eleven households reported 15 family members having had temporary off
fa-rm employment during the year. Though the majority of farms relied on
family labour resources, 27% of farmers hired casual labour to help them out
at peak work periods, mainly for weeding, and 37% hired tractors to help in
seedbed preparation. No specialisation of work between men and women was
observed in the area. All labour can be considered as available for all
4.4. Labour Use
In aone visit survey of this type it is impossible to record details of labour
use. It is possible to diagnose labour peaks by reconciling two types of
information; farmers opinions on busy periods in the agricultural year, and
farmers actions in supplementing, their resources during these periods.
Farmers opinions on the busiest month of the year, together with the incidence
of labour hirings by 27% of farmers are shown in Tables 14 a & b.
Tae 1. Farmers opinions on the busiest months and the incidence
of labour hiring.
Month J F M A M J J A S O N D
% indicating 42 23 20 5 2 3 1 4
% of labour hirings 15 21 23 19 ( 13 )9
There were a further 40 hirings of tractors (40% of all hirings
of both labour and tractors) for land preparation in December,
January and February. Farmers were also asked what their main work was
during this busy month and, as a separate section of the questionnaire, what
work labour was hired for the results are summarised in Table 14b below.
Table 14b. Opinions on peak operations reconciled with hirings
Type of Land Preparation Weeding Harvesting
work and planting and grading
% farmers indicating 51 43
as main work
% all hirings 5
tractor and labour
Farmers were also asked to name their second most busy month and the major work
done then. Ilarvesting an, d Igradi iig or sorl i n;g IFe rCviurd more pIromilnn I Vly; wil I
45% of responses spread over the period May to September. Later weeding, spread
over the March to May period had 40% of responses.
It is clear that the greatest problems for farmers are in the January, February,
March period of the season with an overlapping of land preparation and planting
and the weeding of the earlier plantings. An estimated 65% of working capital
(including credit advances) used on farms went to ease these bottlenecks, all of
it paid out as cash. The capacity of family labour resources, and the farmers'
ability to supplement this capacity by hiring resources during this period is
the present major constraint on the amount of land farmed by the household in
4.5. Sources and use of working capital
No major item of capital equipment was found on any of the farms surveyed.
Sprayers, purchased at the subsidized price of shs. 125.00 were the most
expensive equipment owned and used by individual farmers. The only other on
farm equipment was handtools.
(1) Sources of working capital. The 1976 season marked a change in policy on
the part of Government over the supply of inputs. It is not clear whether the
change in policy; farmers paying cash albeit at subsidized rates, for their
maize inputs was implemented during the 1975/76 season in the study area. It
is not therefore clear just how far farmers were dependent on their own cash
incomes as a source of working capital. There was considerable political
exhortation to grow cotton and to use insecticides in the 1975/76 season and
again, it is not clear how far this influenced farmers' decisions on the use of
inputs or their responses to survey questions concerning inputs. Seventeen
farmers of the 107 surveyed did not estimate their cash incomes, the estimates
of the remaining 84% of the sample are set out below in Table 15.
Table 15. Estimated cash incomes of zone farmers
Income less than shs 500 shs 1000 shs 1500 shs 2000 over
level shs 500 1000 1500 2000 2500 shs 2500
Sof 41 29 13 4 6 7
Average cash incomes were of the order of shs 900.00 per family with 70% of
families estimating their cash earnings at under shs.1000. Most cash came from
crop sales and 30% of all cash income came from cotton. Table 16 summaries
the reported cash sources.
Table 16. Reported sources by % contribution to total cash income
Source % contribution % farmers
Crop sales 62% 89%
Off farm labour 12% 23%
Crafts and Trading 11% 22%
Relatives 9% 19%
Beer sales 3% 7%
Livestock 1% 3%
(2) Expendirltre on ii inipits. Eighty one farmers purchased inputs of one
type or another, that is 76% of those surveyed. Actual cost data was only
collected for labour and in order to make some estimate of the level of outlay
it has been assumed that other area related inputs were purchased for one acre
plots. Details of the proportion of farmers purchasing different types of
input and the assumed costs for valuing these are set out in Table 17.
The range of inputs purchased and their
% farmers actual or
purchasing imputed cost
Labour 25 200
Tractor services 37 100*
Insecticides 52 26*
Sprayer 17 125
Maize seed 26 20*
Sorghum seed 18 20*
Fertiliser 4 50*
Costs imputed at the acre level, subsidized where
The 76% of farmers purchasing inputs spent an estimated shs.178.00 each, that is
17% of their average cash income. An expected relationship is that willingness
to spend on inputs rises as cash income rises and this trend is discernible
amongst those farmers as demonstrated in the table below.
Average levels of spending by farmers in different cash
Income Category Under shs.500 Shs.l000 Shs.1500 Shs.2000 Over
shs.500 1000 1500 2000 2500 shs.2500
Farmers in each
category (%) 41 29 13 4 6. 7
buying inputs (%) 68 96 83 100 100 100
Average level of
expenditure (shs.) 128 182 159 243 253 302
Expenditure or percentage
of reported.income (%) 32 24 13 14 11 10
It is unexpected to find the proportion of cash income spent on farm costs
falling as income rises. It is possible that the lower income groups are
being encouraged to layout disproportionate levels of cash by the carrot
and stick of credit and political exhortation. The average level of outlay
of 17% of cash incomes by those buying inputs is about the expected proportion
at these levels of cash income. To be within the reach of the incomes of the
majority of farmers in the zone the ceiling on the costs of an improvement
package should be of the order of shs.150.00.
A further point to be made is the high proportion of working capital (65%)
being spent on relieving labour bottlenecks. This compares to 35% being
used to intensify yields and of this 31% is for cotton insecticide and
sprayers under a strong campaign from government. Given the relative abundance
of land in the area and evidence of labour bottlenecks in the January to
S March period limiting the scale of family from operations, the use of working
capital in labour and tractors is the more logical path. It may be that the
new nucleated settlement pattern, the high overheads of getting to and from
unused land and thus a tendency to use lands adjacent to the villages more
intensively will improve the attractiveness of using intensifying inputs in
the future. Anyway it is clear that the profitability of using inputs to
improve yields will have to be competitive with and no more difficult to
manage than the use of .labour and tractor hire both to extend the cultivated
area and to allow farmers to catch planting rains more easily.
5. FARMING HAZARDS
5.1. Introducitr ion
Hazards of rainfall failure, pests and diseases, and of the uncertainty of both
crop selling and food buying prices in local markets, create risks for farmers,
particularly for those operating subsistence and semi-subsistence farms. The
lower the level of family income the closer the family are to subsistence
level the greater the risk from a given set of hazards. Clearly, for families
operating at the subsistence level, crop losses may mean starvation and death.
Similarily the range and extent of hazards vary with the farming environment.
Tn marginal areas rainfall failure is likely to be the dominant hazard, and a
severe one. In wetter areas the set of hazards may be less severe and among the
set crop pests or crop and animal disease may be dominant.
Uncertainty is the key characteristic of such hazards; their timing and severity
cannot be forseen. The probabilities of such hazards occurring over a period
of time can be calculated from experience and so their frequency can be antici-
pated. From experience farmers know the incidence of hazards in their own
farming environments and have evolved management strategies, embodied in their
enterprise pattern and husbandry practices, to protect them from the effects.
The closer farmers operate to the subsistence level and the more hazardous
the particular farming environment the more will hazard aversion dominate farm
management and decision making.
When .fanrmnr consider trying oIu iillovItion tIhey assess the likely impact on
their livelihood. They make their assessment on the basis of their present
situation: They relate the costs of the changes involved to their present
cash income and expenditure levels. They relate the new management require-
ments.to present management patterns of their farming system. If the inno-
vations require extra cash outlays which increase the losses at risk due to the
effects of hazards to an unacceptable level, or, if the changes in.management
disrupt management strategies considered vital as insurances against the
effects of hazards, the changes will not be adopted. Innovations which threaten
farmers' hazard management are inappropriate and research which works towards
such innovations is unlikely to be fruitful. Indeed, such innovations can be
positively harmful, increasing the dependence of the farming population on govern-
ment assistance when hazards are experienced at all severely.
The study has identified uncertain rainfall as the major hazard facing farmers
in the drier areas of Morogoro/Kilosa and attempts to evaluate how farmers
manage their farms to reduce the effect of rainfall variations on their liveli-
hoods. The evaluation aims to guide research efforts for farmers of these
areas towards innovations which, at worst, do not disrupt farmers' hazard
management strategies, and which, at best, give improved insurance against the
possible effects of rainfall uncertainty in these areas.
5.2. Identified hazards in the zone
(1) Rainfall Uncertainty. It does not require a farmer survey to verify the fact
that rainfall uncertainty is a dominant hazard to farming in these drier areas,
and, as we shall see, a major influence on farm management. Data from two
stations in the area demonstrate the high variability of monthly rainfall.
No records are available for the Mkata plains area at the centre of the dry
areas and figures for Mvomero are used to discuss the rainfall pattern there,
ARI Ilonga figures example the situation on the peripheral areas surrounding
the drier areas proper, these areas have a higher average annual rainfall
and lower year to year variation. Rainfall from June to October is negligible
throughout the area, November, December and May are very variable, March and
April are almost always reliable though in some areas March, like January
and February in others, show a rainfall trough and low reliability in the
middle of the tainy season. The trough is a vestige of the bi-modal distri-
bution nearer to the Equator. The tables below show the maximum levels of rain
likely to occur over the growing season months in six, four & two years out of
ten. Very rough evapo-transpiration estimates and crop requirements for Et/Eo=
.50 and .75 are included. For example; for maize growth at flowering Et/Eo =
1.00 is required. Table 19 shows this to be 180 mm in January in the drier areas
which receive a maximum of 107 mm in January six years in every ten.
Rainfall probabilities in the drier areas Mvomero (mm)
P level Years Nov Dec Jan Feb Mar Apr May
.60 6 54 86 107 89 176 238 94
.40 4 23 65 84 48 132 191 75
.20 2 5 25 35 22 87 146 53
Et/Eo = 1.00 200 190 180 160 180 140 130
Et/Eo = .75 150 142 135 120 135 105 98
Et/Eo = .50 100 95 90 80 90 70 65
Tae 2. Rainfall probabilities in the peripheral areas -
ARI Ilonga (mm)
.60 6 62 172 142 154 218 222 67
.40 4 29 116 115 108 192 185 48
.20 2 9 59 69 74 148 154 24
Et/Eo = 1.00 201 183 171 149 166 131 122
Et/Eo = .75 151 137 128 112 125 98 92
Et/Eo = .50 101 92 86 75 83 66 61
The main points to be emphasised from the tables are:
(1) In the drier areas, in the majority of years, rainfall
February will not provide crop requirements of Et/Eo = .75,
required for periods of fast vegetative growth. Only April
in January and
below the level
reliably meet full crop requirements in all years. March rainfall can meet
maximum requirements in four years out of ten and 75% of Eo in six years out
(2) In the drier areas December and February provide adequate planting rains
four years out of ten, January six years out of ten.
(3) May rainfall can provide 75% Eo in four years out of ten and 50% six years
out of ten as an aid to grain maturity. May will only give 100% Eo in two
years out of ten and is usually inadequate for a crop in a period of maximum
For the peripheral areas with higher levels of rainfall the pattern is very
similar but the higher levels of rainfall, with the exception of May which
is similar to May rainfall in the drier areas, keeps the levels useful for
crop growth at a higher probability. The real anomaly of the drier areas is
that while December or January or February can provide adequate planting rain
in most years, only April can provide maximum crop water needs reliably every
year. Even March falls short of maximum requirements six years out of ten.
Using the planting rains that often fall in December or January will almost
certainly put the crop under heavy water stress in February or March. Even
using planting rains in early February crops must be vulnerable to water
stress in March as their requirements increase. Perhaps the only reliable
season is from early March to early or mid-May, giving a period of some ten
During the survey farmers were asked to specify their three major farming
problems. 104 out of 107 farmers answered the question but many responded
only once. Answers showing the major problems specified and the accumulation
of answers on all problems are given in Table 21 below.
Table 21. Incidence of answers on farming problems
The expected dominance of rainfall uncertainty is clearly shown. The pattern
of answers reflects the unreliability of the rains at the start and finish of
the growing season and the rainfall trough in the middle of the season. The
uncertainty of the timing of the planting rains shows up strongly, specified
by 53% of those answering as the major problem.
Answers on All
Problem major answers
Rains starting late 55 83
Rains failing during Jan/Feb. 18 67
Early finish to the rains 25 76
Too much rain 1 8
Other problems 3 17
No problems 2 4
(2) Pests and Diseases. Further questions were asked of farmers specifically
on the pest and disease problems they feel are important to them. While
reporting their opinions here, this is not to deny that other pests and
diseases may have a significant effect on crop production. It does suggest
that innovations are likely to be more readily acceptable for those pests and
diseases felt to be important by farmers. To put the following paragraphs in
perspective it must be re-emphasised that few farmers rated pests and diseases
as important as rainfall uncertainty as a problem in farm management. Farmers
were asked for the pests they considered most important in relation to their
food crops in the field and in store, and their most important animal problem.
Their responses are set out in Table 22 below.
Table 22. Responses to questions
on pests on food crops
. A similar attempt to identify diseases
of which 27 reported 'no problem'. 34
mentioned rust on sorghum. No attempt
frequency of pest and disease attacks.
(3) Supply and Pric4
prices are published
sold unofficially in
e Uncertainty in Lo
before the buying
local markets and
important to farmers brought 85 responses
farmers mentioned maize streak and 11
was made to assess the severity or
cal Markets. Although official market
season many food crops are bought and
between neighboring farms, no price
indications are available for these transactions. It is known however that in
the months before the new harvest in the area certain foodstuffs may be
obtainable or obtainable only at very high prices. The official farmer price
for maize in the survey area for the 1977 year was shs. 65.00 per 90 kilo bag
and the official selling price (ex-maize board) shs. 105/50. On the unofficial
market, in the early months of the year, maize for food cost shs.200/- per bag
Major field pests 104
Army Worm 60
Maize stalk borer 31
Sorghum shoot fly 9
Major storage pests 106
Maize weevil 8
Major Animal pest 104
Baboons and Monkeys 16
in Morogoro and Kilosa towns, and shs.150/- per bag in the rural areas. Thus
the value of home produced maize, to farmers who produced enough to last into
these months, was 230% of the price the farmer could have realized by selling
to the Board. This level of seasonal variability creates high risks and
farmers are unwilling to depend on purchased supplies. Paying shs.105/50
per bag for six months and shs. 150/- per bag for the other six months it
would cost the average farmer some shs. 1650/- to keep his family in maize
for the year; some 180% of reported cash income levels.
These then are the hazards faced by farmers in the zone. We go in section
six to discuss the farmers' management strategies in the face of the hazards
of the area, particularly rainfall uncertainty, and his resource endowment
pattern which bounds the type of strategy he can adopt.
6. THE INFLUENCE OF RAINFALL UNCERTAINTY AND LABOUR AND CAPITAL
LIMITATIONS ON FARM MANAGEMENT.
Farmers in this zone are limited in what they can produce by rains which fall
with some regularity for only 21-3 months in the year and are variable in their
The management strategies which they adopt to combat this major production hazard
of rainfall uncertainty, and the risk of food supply failure it presents, are
restricted by their limited supplies of family labour and the small amounts of
cash available to hire extra labour or tractor services.
Many of the management strategies for reducing the effects of rainfall uncertainty
have evolved over generations of farming under hazardous conditions. Such
strategies are inherited by the present generation of farmers who are not always
wholly aware of the insurance roles of particular practices. Other strategies
give flexibility to react to contingencies as they arise and are more obviously
Strategies can be divided into three categories:-
1. Enterprises included in the cropping pattern primarily or partially for
their insurance role.
2. Manipulation of the relative size and timing of crop enterprises.
a) Adjustment of the area cultivated of various crops. For example when
food stocks are low a larger area of a drought resistant, more certain,
staple may be grown.
b) The sequence of crop establishment may be varied to give priority to
short maturing crops or drought resistant crops. This may include the
manipulation of varieties, for example the use of a shorter variety
when early food is required, or when initial plantings are suffering and
seem unlikely to provide enough staple, shorter term varieties may be
planted to give a higher probability of maturity in the remaining period
3. The use of specific husbandry practices to improve the probability of
production; intercropping is a common example.
Strategies identified in the drier areas of Morogoro/Kilosa can usefully be
discussed under these heads. It must be emphasised however that evidence on
the use of some strategies, even though reported by farmers in presurvey dis-
cussions, is difficult to document in survey due to the confounding of the variety
of crop end uses, hazard management, individual management capabilities and
investigational problems. As in any farming population some farmers are better
managers and are better able to exploit the available strategies. Poorer managers
lack the ability to manipulate the crops, varieties and practices effectively..
6.2. Insurance Embodied in the Cropping Pattern
In the Morogoro/Kilosa areas sorghum is an example of a crop included in the
cropping pattern by some farmers as an insurance against a failure of the preferred
food grain maize, The clarity of the insurance role is obscured by two factors:
1. Brewing is an important end use, sorghum is a base for making local beer.
2. 10% of farmers expressed a preference for Ugali made of sorghum and a
further 11% were indifferent whether their Ugali was sorghum or maize based.
Despite these obscuring factors 45% of farmers, more than double the population
preferring sorghum, ate sorghum in the course of the year.
Two other crops cassava and perhaps also rice can be seen as playing an
insurance role in the cropping pattern. Although only 11% of farmers were
recorded with fields of cassava it is clear from the number of house plots with
cassava, and the number of farmers familiar with cassava varieties that it is a
widely used food. On the other hand no farmers reported cassava as their, main
source of food. Rice is in a different category; 37% of farmers sampled reported
growing rice and 70% of these reported it as an important cash crop while 12%
of these reported it as their main source of food. It seems probable that, fot
farmers with access to the low lying heavier soils, being both a food and a good
cash crop, rice also has an insurance role for growers. The low lying areas j
with a high water table are liable to flooding in the heavier rains of March
and April. Maize would not survive waterlogging. These areas benefit by the
early, sporadic, rains in December and January, which raise the water table and
allow relatively reliable early planting of rice as a water tolerant crop.
6.3. Manipulation of the crop calendar
Characteristic of the area is the wide spread of planting dates, and as a
consequence, the wide spread of harvest dates. 90% of reported plantings are
spread over the three month period from mix-December until mid-March. Harvesting
is spread over a longer four and a half month period between the second half
of April and the ehd of August. Maize harvesting dominates the period April
and May and cotton and sorghum the harvesting done in August. The spread of
planting dates over such a long period is caused by the need to juggle crop and
variety maturity periods and family needs with the family's capacity to prepare
seedbeds and plant. It also contains an element of risk aversion: When earlier
plantings can be seen to be unsuccessful further food crop plantings take
priority over cash crop establishment. The general priority given to food crops
is shown with over 90% of cotton plantings made after the first of February
compared with 32% of plantings of starch grain crops.
The general picture of priorities given by farmers in discussion is listed
1. Early establishment of Jebele; it takes longest to mature and needs the
full length of the rainy season, and is used green from the field as an
2. Establishment of maize to give early food from the new seasons plantings
with the local variety having preference. Use of maize as an early food
was widely reported by farmers.
3. Continuing establishment of sorghum and the establishment of rice, also
requiring a relatively long season. The major sorghum plantings are of
Mbangala, a shorter term variety; a good storer popular for beer making and
for ugali; surpluses are sold when the crop is good.
4. Establishment of cotton and sunflower.
The relatively long maturity periods of the varieties of both sorghum and maize
grown make the early plantings very vulnerable to the markedly unreliable rains
in December, January and February particularly in the drier areas. Similarily
late plantings are vulnerable to the unreliable rains in May and the dry period.
At this altitude Ilonga Composite has a maturity period of 115-120 days depending
on the season, the local maize is similar. Figure 3 plots estimated water
requirements for a 115 day variety at alternative times of planting against
monthly probabilities of rainfall level six years and four years out of ten.
The diagrams illustrate the vulnerability of any variety to early planting. Mid
December plantings need maximum water at flowering in February when rainfall is
highly uncertain, similarly January plantings come under heavy stress at
pre-flowering in a high proportion of years. Mid February planting gives the
best possibilities for meeting the requirements of a 115 day variety enhanced
by the lower radiation levels in April and May and lower crop requirements in
March. Clearly the sorghum varieties, because of their relatively long maturity
periods, must be planted early in the season despite the heavy risk of water
A further important aspect of the crop calendar is the late planting of beans,
from late April to mid June, in very specific locations; the heavy wet valley
bottom lands. This late planting ensures the beans fruit in dry weather and
avoids many of the diseases problems associated with beans maturing in the rains.
The late planting and therefore late harvest of beans reduces the storage
period through until legumes or cowpea leaves become available again from the
following seasons plantings. The bean crop is often planted after an early
maize crop reflecting the intensive use made of the heavier, water retentive
valley bottom soils by those who have access to them. Some farmers also
reported planting maize and cowpeas at the end of the wet season in these valley
6.4. Crop Husbandry
Several facts of crop husbandry reflect the risk avoidance priorities of local
(1) Intercropping. The survey recorded a low level of intercropping, partly-
a reflection of the very low production of pulses reported by survey farmers.
However intercropping cowpeas and to a lesser extent beans and gram, particularly
into maize, is observed practice by farmers in the area. Specific intercropping
practices to help avoid crop failures are the mixing of maize and sorghum,
reported by 13 farmers, and the planting of beans or cowpeas into rather poor
maize and even sorghum crops towards the end of the rains.
(2) Locational Factors. We have already described the importance of the low
lying areas to farmers in the zone. They use these areas with a high water
table and good moisture retention to extend their effective growing season. Rice
is planted in these areas early in the season. Early maize plantings are made
in such areas to produce an early food supply. Beans and sometimes maize and
cowpeas are planted here late in the rains to take advantage of residual moisture.
(3) Wide spacing of stands. Farmers aimed for a population of 23,900 stands of
maize per hectare and 20,200 stands of sorghum. This compares with 37,500 stands
of maize recommended from experimental work. Wide spacing allows plants access
to more soil area and one reason for it is to increase the likelihood of a crop
in dry years by improving the chances of the plants resisting water stress. The
large number of seeds placed per hole reduces the rate of failure in germination.
'16IJE 3: WATER REQfJIREMENTS lP'OU A 115 DAY MAIZE VARIETY
DIA(GRAM 1. PLANTED') MIVI) IDE CEER
DEC. JAN. FEB. MAR. APR. MAY
DIAGRAM 2. PLANTED MID JANUARY
6 Years of 10
4 Years of 10
3. *0 Maize
DEC. JAN. FEB MAR. APR. MAY
DIAGRAM 3. PLANTED
DEC. JAN. FEB. MAR. APR. MAY
(4) Use of minimum tillage techniques. Particularly early in the season,
because of the intermittent character of the rainfall, planting of the crop
must be done to the day. We have seen a significant proportion of fields
prepared by merely scraping the weeds off, particularly for sorghum (41% of
plots). When the crop has to be established urgently to catch soil moisture
to ensure reasonable germination the low labour requirements of this scraping
technique mean a larger area can be established at one time. Bartlett (1977)
has reported moisturing conserving techniques in the preparation of the low
lying lands for the late planting of beans. Weeds grass and crop residues
are cut off at ground level, with a panga, and left on the ground until planting
when it is cleared off.
6.5. Risk of cash losses
A final area influenced by the risks arising from the uncertainty of the rain-
fall is the amount of cash farmers are prepared to layout on the purchase of
inputs. Traditional methods of production in the area have no cash costs;
farmers use their labour on the land and only the effort is wasted if the crop
fails. When cash costs are incurred by the use of purchased inputs the results
of crop failure are more significant to the farmer. He will be averse to
placing his limited cash income at risk and places a ceiling on how much he is
prepared to spend. The level of this ceiling can be raised to some extent by
the use of credit but this is limited; he still knows, because of the chance
of crop failure, that he may have to meet his credit obligations out of his
usual cash income. As we have seen, to be within the reach of the majority of
farmers in the zone the cost of any recommended husbandry changes should not
6.6. Decision making
Three types of decision situation relating to the uncertainty of crop output
levels caused by unreliable rainfall can be envisaged.
(1) Long term evolution of the farm. Where the weather pattern changes tlhe
farmer, by experience over a number of seasons, observes that he does not need
to fall back on his insurance foods or insurance management practices as
frequently. He may feel his risks are reduced and change the level of iinsrance
cropping, reduce the area of sorghum planted or even drop sorghum altogether
in the case of Morogoro/Kilosa farmers. This will allow a re-allocation of
resources to more profitable crops; an increase in the area of cotton grown, or
in the number of farmers growing sunflower, by releasing labour and land tied
up in insurance crops. It is possible that maize asserted itself as the main
'* starch staple in the long period of good rainfall in the early 1960's.
(2) Season to season decisions. At the beginning of a new season the farmer
will be influenced in his priorities by last seasons harvest, his present food
stocks and his expectations of purchases prices for food in the local market.
Low food stocks may mean apriorityis given to early planting of shorter term
crops or varieties. High food stocks may mean that preferred crops or even
pure cash crops are planted early. Both the cropping pattern and the agricu-
ltural calendar are likely to be varied due to this category of decisions.
(3) Within season decisions. As the season progresses and he sees how his
initial plantings are doing in the field he may modify his plans. For example
he may have planted early maize in low lying areas. December, January and
February rains may have been a typically heavy giving a high water table in such
areas and his early maize plantings suffering from waterlogging. lie must then
decide what to plant and where to plant it to be sure of securing his starch
supplies through the next dry season. This flexibility in management to react
to contingencies of the season is particularly important in areas with rain-
fall characteristics of the Morogoro/Kilosa zone.
Important aspects of farmers' hazard management in the drier areas of Morogoro/
Kilosa are confounded by resource constraints which also play their part in
dictating their cropping pattern, crop calendar and crop husbandry practices.
A research programme for farmers of the area will be strongly influenced by
both resource constraints the farmers operate under and the hazards they face,
particularly rainfall uncertainty which they need to be able to offset by
flexible management strategies.
7. FARMERS PROBLEMS AND THREE CONTRIBUTORY SOLUTIONS
In general the main problem for farmers in the zone is reliable production of
enough starch staples and relishes to meet family needs through to the new
harvest time in April/Mny. We Iave pointed out that 1976 was a bad year Ior
the legume crops, with poor early plantings and very few successful late bean
plantings due to an early, abrupt end to the rains. Farmers'opinions also
reflect 1976 as a poor year for maize and sorghum, certainly poorer than the
previous 1975 season. Table 23 summarises their responses when asked to
assess their grain harvests as good average or poor.
Farmers' assessments of their maize and
Table 23. in 197 and 1975
in 1976 and 1975
% of farmers reporting
Year Crop Good Average Poor
S Maize 22 31 47
Sorghum 15 45 40
Maize 45 32 23
Sorghum 53 28 19
These self assessments suggest that 1976 was below average in the distribution
of years in farmers' experience.
7.2. Starch staples supplies in 1975/76
In the context of 1976 as a poor season a high proportion of farmers in the
zone failed to maintain their supplies of maize and sorghum through to the new
maize harvest, starting in April/May. Due to the timing of the survey qluestion-
ing related to the relatively late finishers of the 1975 harvests (January to
June 1976)and the early finishers of the 1976 harvests (July-December 1976).
Table 24. sets out the percentages finishing in each quarter of those farmers
reporting on maize and sorghum.
Finishing stocks home grown maize and sorghum
Harvest 1976 (% of farmers reporting)
quarter maize sorghum
finished (74 farmers) (38 farmers)
Jan-March 47 61
1975 April-June 19 16
July-Sept 15 5
1976 Oct-Dec 19 19
Backing up these rough figures on the time of finishing homegrown stocks of
maize and sorghum- 33% of all farmers made purchases of maize during 1976 and
13% made purchases of sorghum. Most purchases were made in the three months
before the 1976 harvest and must be related to the 1975 harvests which only
a minority of farmers assessed as poor (23% maize, 19% sorghum).
Despite a variety of management strategies to try to ensure a reliable supply
of foods, maintaining supplies through to the new harvests, at the earliest
in April/May, is a problem for a significant proportion of the farmers of the
'on0e. The problem period for grain supplies lasts as long as October to April
in very bad years, even in an average year many families have to eat sorghum
instead of the preferred grain maize over this period.
7.3. The search for appropriate solutions
A variety of potential solutions exist:
(1) Bringing new season harvest dates forward to get earlier new food. This
may be tackled by:
(a) Growing earlier maturing varieties.
(b) Planting drought resistant varieties earlier.
(c) Soil management which allows existing varieties to survive
more easily when planted earlier despite drought periods.
(2) Allowing more reliable production of a storable starch staple to prolong
(a) Improve methods of storage.
(b) Grow crops and/or varieties best fitted to water availability
in the zone.
Farmers will be most interested in solutions to this, their priority problem.
Changes in management which raise yields in the years of good rainfall by the
use of high levels of purchased inputs are not a solution and may, in years of
poor rainfall, merely exacerbate the problem. If the variety concerned is
susceptible to drought, not only will food shortages be more severe but the
outlay on the inputs purchased will be a loss to the farmer. The changes in
management most acceptable to farmers in the zone are those which:
(1) Contribute to a solution of their priority problems; a reliable food
(2) Can be implemented with the resources- the labour supply and available
working capital, the farmer has at his disposal.
(3) Improve the productivity of the resources farmers' utilise in producing
foods, allowing a re-orientation of resources released from producing insurance
foods to producing cash crops for the market.
Three possible contributory solutions are examined in some detail.
7.4. The introduction of a short-term maize, taking 45 days to tasselling
Maize is the preferred starch staple of farmers in the zone yet its establishment
tends to be subordinated to that of sorghum because it is a less reliable crop.
The sorghum varieties grown, with their longer maturity periods, have a clear
need for earlier planting. Later planting is however consistent with the greater
vulnerability of maize to water stress given the erratic rainfall in a high
proportion of years in January and February. However the present Local and
longaa Composite varieties have a 115 120 day maturity period and if planting
is delayed beyond mid February they became vulnerable to the cut off in the
rains in mid May. With their maturity periods the present maizes can give food.
I':rli'or ill In Ithen, s ni h l hn h I e I'Ini-ghiiiIIi ;Ind Folrlmers in need of early food from the
In'w hii rvs t wilI Imiiak an Oarl y iii.ize pl )an; i ng 1lalf the farmers surveyed had
planted maize by mid January. Early maize plantings tend to be in the lower
lying locations to help protect plants from water stress in the erratic January/
February rainfall. Such early plantings are often followed by beans planted
from May to July to exploit the water retentiveness of these areas. However
the rate of failure of such plantings is known to be high.
The most reliable period of rainfall is early March to Mid May, a period of 70-75
days. A variety tasselling in about 45 days planted on the first good March rains
would make the best use of this reliable period of rain, and would bring significant
advantages to farmers as a more reliable source of their preferred starch staple.
(1) The shorter term variety would enhance the ability of the farmer to react
to contingency situations a late start to the rains, a marked gap in the
rains and would allow them a crop from early and mid March plantings despite
an early finish to the rains in May. These three contingency situations were
mentioned as key hazards by a high proportion of farmers.
(2) Maize is already used to supply early food in the new season.A variety tasselli
in about 45 days would produce food 3 weeks earlier than present varieties meeting
this need more effectively.
(3) The shorter variety early planted in these areas would free the wet low
lying locations some 3 weeks earlier than the existing varieties which has
several possible ramifications.
(a) Increased probability of getting a bean crop from these areas.
(b) Increase in the number of low lying areas where a bean crop could
(c) In the better low lying areas it opens up the possibility of growing
other crops with a slightly longer growing season than beans.
(4) The introduction of a short-term maize crop capable of giving a reliable
supply of the favoured starch staple from an early or mid March planting will
have two more important alnd intLracting repercussions on the farming systems
both of which should improve the scope for cash crop production by Morogoro/
Kilosa farmers in the zone.
(a) The later planting date will spread the demand for labour into late
March/early April (weeding the maize planted). This will improve a
farmers flexibility at a key time of the year by:
either Allowing him to reduce his cash outlay on labour and tractors to
give him the required area of crops (2/3 of all farm expenditure)
permitting the use of cash for other purchased inputs.
or Allowing an increase in the area cultivated by maintaining his
present outlay on hired labour and/or tractor services.
or Allowing more timely planting of an increased area of cotton or
other crop grown for cash.
(b) The increased reliability of the maize crop will reduce the need
for a back up with sorghum and the sorghum area may be reduced. This will
free labour being used on Sorghum establishment and weeding in the January/
February peak periods creating the same alternative opportunities for
resource use as set out in (a) above.
An 85 day maize is technically feasible at this altitude in these latitudes and
several potential selections are already under trial at ARI Ilonga. Indications
from work of the National Maize Research Programme augur well for finding a
short term variety and agronomic practices appropriate to the conditions of the
drier areas of Morogoro/Kilosa Districts.
(1) There is data to show that the short term varieties outyield the standard
varieties at mid March planting dates.
(2) Yields from some short term varieties under observation are good after only
60 days of rain.
(3) Increased plants per stand and fewer planting stations are a substitute for
high numbers of planting stations. This is close to farmer practice in the
(4) Evidence is growing that some shorter term varieties with limited vegetative
growth can carry very high plant densities (twice normal) which compensate
in yield terms, with the longer term, more vegetative, varieties planted at
'normal' densities, even at earlier dates of planting.
Farmers expressed some views on the type of maize plant they preferred in the
course of the survey. Rat damage in the field was said to be a major problem.
Rat damage is unlikely to be correlated with the short plant height associated
with short term varieties but with proneness to lodging, caused by weak stalks
and a poor rooting system. Resistance to lodging would be a desirable characte-
ristic of a selected variety. Good storage characteristics are a clear need for
farmers of the area. Flint rather than softer grain types are preferable and
good husk cover would minimise the transfer of insects from the field to the store.
Good hust cover would also reduce proneness to ear rots which would be heavy
with varieties planted early for early food, maturing in heavy rains. The ear
rot problem would be less acute when such early plantings are eaten green which
is local practice with much of early planted crop. Farmers seemed to prefer
large cobs, and, while small cobs are not necessarily a feature of short-term
varieties, several prominent ones do have small cobs. Two further points raised ,.
in team discussions will be important features of an experimental programme based
on a short term variety. First the question of whether such varieties will be
able to survive erratic rainfall in December or January as well as the local
variety. Second, given the weeding bottlenecks in the system at present, the
possibly greater need for early weeding of short term varieties may aggravate
already prominent labour peaks. These two points, together with others, are
taken up in more detail in the proposed experimental programme centred around
short term maize selections in Section 8 below.
Finally it seems unlikely that planting the late bean crop some three weeks
earlier would expose it to new insect and disease complexes. Most of the insect
and disease problems with beans in the area are related to flowering and fruiting;
these stages of growth would still occur after the end of the rains, except very
occasionally when rains are prolonged into June.
7.5. Improvements in Sorghum Production
Alternative strategies for the improvement of sorghum productivity in the farming
system seem economically feasible but may present technical problems.
(1) The promotion of a 6 61 month long term sorghum (the same maturity period
as Jebele) to be planted in December, but with the qualities of Mbangala. In
addition to drought resistance it would require good storage qualities, good
qualities for local methods of brewing and be acceptable to local families for
use in ugali. It would take over the role of the main sorghum crop from Mbangala
and, with the longer maturity period, would aid the spread of labour peaks in
both the cultivation and harvest period.
(2) The promotion of a short-term sorghum with similar varietal characteristics,
again to take over the rolQ of Mbangala in the system. In addition a short
term sorghum may provide more reliable source of early food than a short term
* maize. As with the short- rm maize the sorghum would give the farmer greater
flexibility in labour mana ment, particularly when reacting to contingencies
arising such as a late state to the rains or a failure of initial plantings.
Superficially an early mat ring variety, if planted early for early food,
because of more rapidly in, easing water requirements, would appear susceptible
to the vagaries of rainfall in February when it would be flowering. However, if
rooting depth is the key to drought resistance and a good rooting depth is
genetically not incompatible with early maturity, this may be a false impression.
The key to storability is a hard or corneous endorsperm and this is not gene-
tically incompatible with either early or late maturity. However, most of the
presently available varieties under test have a soft endosperm.
Screening for consumer acceptability either for beer or ugali is feasible and is
particularly important in sorghum. Flour often changes colour when mixed with
water, from white to yellow or brown, and it is important that the resulting
ugali can be kept overnight.
A short-term variety grown for early foods would run into significant drying and
related disease problems when maturing in the rains. Grain moulds and head
sprouting are potentially very severe. This is particularly true on the short-
term varieties now under test which have closed tight panicles, much more prone
to the wet than the traditional varieties which are open panicled.
Proneness to disease due to maturing in the rains would seem to question the
technical feasibility of using a short-term sorghum for early food. It would
not influence its usefulness as a main sorghum crop, planted later and maturing in
the dry season. The improved varieties presently available are all short term,
between 80-120 days, nothing of the order of 6 62 months.
Nor is there sufficient variability in the local populations to justify selection,
these varieties have been selected very closely over a considerable period of
time. To find a long term variety with the desirable characteristics new
material would have to be brought in from outside the country.
7.6. Wider Introduction of appropriate cassava varieties
As a dual purpose crop cassava would be particularly useful in two ways:
(1) The roots used for flour during the October to April period when farmers
are often thrown back on food purchases or the use of sorghum as a non-preferred
starch staple. Cassava improves the digestability of Sorghum ugali.
(2) In this same period greens are often scarce and cassava leaves, being an
acceptable, though not preferred vegetable, would contribute to a better supply
Cassava is particularly useful with its ability to store in the ground allowing
use over a two to four year period as required.
To function as an acceptable supplementary starch staple cassava establishment
would have to slot into the system in a way which did not compete for resource
with other, preferred crops. It would clearly be unacceptable and uneconomic
to recommend cassava for the water retentive bottom lands already prized and
used relatively intensively by farmers. Nor would it be acceptable to attempt
to urge establishment during peak months of January, February or March.
Establishment would have to be either very early or very late in the rains,
in either case particular qualities would be required and technical questions
of feasibility need answering. The two alternatives are perhaps December or
April establishment. Subject to other technical reasons for preferring late
establishment, from the point of view labour peaks early establishment in
December would probably be preferable. There are other good possibilities for '
extending cropping into the latter part of the rains. Technical questions
concern varietal characteristics, soils and pest or disease problems, some
general points are raised and then some specific to the alternative times of
planting. There is no ongoing cassava work at ARI Ilonga.
(a) Are the soils of the area which are not heavily utilised suitable for
cassava growing n.b. the red sandy soils?
(b) 'Kigoma' is the most popular variety for both leaves and roots. Is
this variety known to researchers and are there alternatives with
similar root and leaf palatability characteristics?
(c) The ability to shoot leaves through the dry season would be a very
valuable varietal characteristic if it exists.
(d) Pigs were reported as their most devastating field pest by 48% of
farmers, high resistance to pigs would be a valuable character,
perhaps even if not compatible with the taste.and cookability of
(e) For use of the leaves as vegetables a high resistance to mosaic is
desirable as, once curled, leaves are not acceptable.
(f) High yield is desirable but not a paramount criterion in the selection
of an appropriate variety.
(2) December planting:
(a) Would planting in December at the onset of the rains render the crop
more susceptible to pests or disease?
(b) The ability to survive erratic rainfall as cuttings (Dec-Feb) would
(c) The ability to give edible roots within a year of establishment and
the ability to retain the roots in the ground in an edible condition
for 2J years after planting would be useful.
(3) April planting:
(a) Is third planting time likely to render the crop particularly susceptible
to pest and disease problems?
(b) The dry season usually starts at the middle of May. Due to existing
labour pressures it would be unlikely to be acceptable to establish
cassava before early April. Would a five week period of rain be
adequate to ensure establishment well enough to carry over the follo-
wing five month dry season.
(c) For this planting time edible roots after 18 months in the ground with
the ability to stay edible in the ground for upto three years are
r These cassava plantings might provide a very viable village enterprise. If each
individual farmer were to 'cover' himself in terms of planting cassava as an
insurance crop a significantly larger area would be required. On a village
basis, because not all farmers are likely to be obliged'to resort to cassava
in the same year, a smaller area per household would do the same job.
Assuming a yield of 5,000 Kg/ha of dry cassava and a need for 20% of families
in the average 500 family village to draw on reserves for a three month period
each year, each village would have to establish 5 hectares each year. This
would require of the order of one man day per family per year (500 mandays for
land preparation, planting and two weedings). Assuming a variety capable of
staying in the ground for three years the village would require a 20 hectare
block of land to devote to the crop. The system would have a degree of flexi-
bility due to the storability of the crop in the ground. If a bad year follows
two good ones the village will have some 900'family months'of dry cassava
available to meet any emergency. When cassava reached full maturity, if it was
not required for food, it could be sold by the village.
7.7. The choice among alternatives
Under CIMMYT's mandate for the improvement of maize and wheat section 8 of the
report sets out a detailed research programme to devise a technological package
around a short term maize which, if productivity improvements can be identified,
will represent an appropriate solution to the major problem of farmers of the
Morogoro/Kilosa area- maintaining a reliable supply of foods. There is a strong
case for maize being the best vehicle for the improvement of the system, certainly
of the three possibilities outlined:
(1) Maize is the preferred starch staple of 80% of farmers in the zone.
(2) Maize probably offers the best possibility of creating favourable
repercussions through the farming system as a whole.
(3) It is already a dual purpose crop; food and cash, and has a wider
market, both local and national than either sorghum or cassava.
(4) There is already a National Maize Programme, reflecting a policy for
self-sufficiency in maize, which includes Morogoro Region.
(5) With ARI Ilonga as the centre of the National Maize Research Programme
there is a backlog of experience and experimentation which is immedia-
tely relevant to a further research effort for farmers in the zone.
It would be prudent to recommend that either sorghum, with its advantages
as a dual purpose crop; for brewing and as a back up food, or cassava with
its advantage as an insurance crop which could be communally grown, is also
researched on the lines discussed. Either is complementary to the strategy
for maize improvement and both focus on meeting the identified needs of
farmers in the zone.
8. A MAIZE RESEARCH PROGRAMME FOR FARMERS IN THE ZONE
Based on the need to introduce a short term maize as an appropriate solution
to farmers' problems in the zone, and as a potential means for development of
the farming system, this section outlines an experimental programme designed
to provide a maize improvement package within the capacity of the majority of
zone farmers to implement.
In summary the benefits of a maize tasselling in 45 days will be:
(1) Increased reliability of production from the main maize plantings, conse-
quently less need for insurance crops and increased production for the
(2) Earlier food in the new season from early planting of the short-term
(3) Better prospects for crops following early maize plantings because of
earlier evacuation of the lands.
(4) By later plantings of the main maize crop a spreading of labour peaks in
January, February and early March allowing an increased scale of farming.
The experimental programme is designed around these sources of benefit. It
will compare the performances of selected short-term maizes planted in early
or mid March as a main crop and in later December/early January as an early
crop. It will seek improved management techniques, for the best selections,
techniques which are within the capacity of the majority of zone farmers to
8.2. Experimental background
Under the conditions of uncertainty and low incomes existing in the zone
farmers are unwilling to put themselves at risk by taking large steps away
from existing practice. They are looking for small changes which will improve
farm productivity without jeopardising family security. The experimental
programme seeks to identify small changes in maize management which acre coin-
sistent with this strategy. The starting point for experimentation is the
present maize management routine of the zone farmer. Control or check plots
r will be under farmer management as representative of the standard the programmed
is trying to improve on. Practices which it is considered farmers will he
willing to change will be the experimental variables in the programme. Treat-
ment levels will be bounded by the range felt to be feasible for the farmer.
Practices which it is considered farmers will not be willing to change will be
non-experimental variables. In the agronomy and selection work these will be
held at farmer levels as identified by the survey or, if necessary, by supple-
mentary investigation in the zone.
The location of the experimental work will be a particularly important factor.
Maize is mainly grown on the black sandy and black heavy soils of the area.
For the early plantings low lying areas are used to improve water availability.
Experimental work should be located in these same conditions. On station work
is easiest to manage where farmers conditions can be reproduced on the station.
However, external field trials, within the zone, will play a hajor part in the
8.3. Variety selection
Variety work would be limited to selection, it would be unlikely that a breeding
programme would be viable with only 65,000 farms in the zone. If, however, the
areai is g ivon high priority by (ovci rllnlctl or if the strategy of a short term
maize used in this way is flouli to be mIore widely relevant than for this zone
alone, longer term breeding work might be justified. An initial development
effort based on selection with subsequent monitoring of farmer reactions would
create a valuable base for a closely focused breeding programme. It is unlikely
that presently available selections will meet all desirable criteria. In the
final choice there is a pay-off, and very good performance on a more fundamental
criterion can offset non-compliance on a less fundamental one.
(1) Yield. Yield remains an important consideration with the emphasis on the
reliability from year to year. It is believed yield reliability can be improved
by growing a shorter term variety which can mature over the period of most
reliable rainfall; early March to mid May. However reliability cannot be con-
sidered in isolation; the farmer will expect any new selection to give better
results thanhis existing varieties in most years. It must improve yield levels
when planted at the identified times; in early or mid March and in late
December/early January. Bearing in mind the likely losses from early plantings
a composite would have an advantage over a hybrid.
(2) Maturity period. The solution assumes the selection of a variety tasselling
in 45 days. However one tasselling in 50 days would offer a 2 weeks shorter
period in the field and should be considered if particularly competitive on
(3) Good rooting system. As an aid to drought resistance which is important
if the selection is to be used as an early planted food crop, strong roots are
also important in preventing lodging to avoid rat damage in the field.
(4) Storability and palatability. Selections should be no worse than the
Local variety and Ilonga Composite for storing and for preparing and eating
as ugali. The early plantings for early food are unlikely to be stored for
long. For main plantings, expected to provide stored maize to carry over the
whole year, both good tip cover, related to the tendency of badly covered cobs
to carry insects in from the field to the store, and hardness of endosperm will
be valuable characteristics.
(5) Large cobs. Farmers in the zone express a preference for large cobs. It
should not however, be a primary consideration in selection. ,
8.4. Variety screening and comparison
Potential selections should be compared with each other and with Local and Ilonga
Composite under conditions as close as possible to farmer management. Time of
plantings, (both early and late) are laid down by the overall strategy of the
solution. Method of planting should be farmers' methods and the plant population
should be at the farmers' level with a relatively low number of stands and 3-5
seeds per hole. Weeding should be limited to two, the first in the two weeks
after planting, as reported by most farmers.
(1) Acceptability. Potential selections should be screened, in comparison with
Local and Ilonga Composite, for preparation, cooking and palatability when
made into ugali by local methods, by a panel of farmers' wives from the zone.
The cooking and palatability tests only need take a short time and should
precede the selection work, potential selections which are clearly unacceptable
can then be dropped.
(2) Storability. Potential selections should be screened for storability
under local methods of storage against Local and Ilonga Composite, with and
without the use of a storage insecticide.
(3) Main Crop Plantings: Potential selections should be compared for yield with
Local and Ilonga Composite when planted on the first good planting rains in
March. It would also be valuable to compare results from plantings on each
planting rain falling in February and on planting rains falling upto the 20th
March. Screening for reliability in results presents a problem of assessment
in the short-term. Comparisons should be repeated over several years to catch
the interseasonal rainfall distribution. It will be important to maintain
rainfall records at the site of each comparison. It may be possible, in order
to give more rapid results, to approach the evaluation of reliability in other
(a) Use historical data already available which can be related to site
rainfall conditions and therefore to historical rainfall patterns.
(b) Use micro-climatic variation, across the zone to proxy inter-seasonal
variation, here site recording of rainfall would be particularly
All trials should be scored for stalk borer infestation, bare tips, the incidence
of lodging and the size of cob.
(4) Early plantings. Potential selections should be compared for yield with the
Local variety in trials planted on each feasible planting rain from mid
December to mid-January. Sites for the comparisons should be chosen in low
lying areas characteristic of where farmers attempt their early plantings.
Varieties should be scored for depth and strength of rooting, stalk borer
incidence, lodging and bare tips of husk cover. At least a proportion of the
crop should be allowed to dry off in the field and eventually be scored for cob
rots. Again, it may be feasible to screen for drought resistance by simulation
with irrigation in off season nurseries.
Any longer term breeding work would seek to improve yield and remedy weaknesses
in the supporting characteristics of the best selections.
8.5. An Agronomy programme
The times of planting of the selections are laid down by system considerations.
Agronomic work will vary in seeking improved practices for the early and main
plantings, but the respective times of planting should be constant; the first
adequate rains after mid December for work on the early plantings and the first
adequate rains after the first week in March for the main plantings. Although
maize growers used three distinct methods of cultivation on their plots there
was no discernible relationship between the type of cultivation used and time of
planting. It is important that the agronomist examines seedbeds produced by
the locally available hire tractors and by deep hoeing (scraping with a hoe
was used less on maize than sorghum) to ensure that the experimental seedbeds
are not atypical of the tilth obtained by farmers. Agronomic work will centre
around three experimental variables: plant spacing and density, timing of
weeding and the limited use of purchased inputs. Each is discussed in turn.
(1) Plant spacing and density. As we have seen spacing is organised with
relatively low stand populations and 2 or 3 plants per station. This clustering
of plants allows intercropping of the maize stands and facilitates weeding with
the long handled hoes in use. It also reduces labour use in planting and in
fertilizer placement where this is stand related. Recent agronomy work suggests
that there is no sacrifice in yield by increasing plants per stand and reducing
stands. This augurs well for a recommendation close to existing farmer practice,
it requires confirmation and the limitations on such a substitution need to be
In experimental work for the zone treatment levels for interstand spacing
should not fall below 40 cms, and for interrow spacing below 75 cm. Further
checks on present stand populations are required in the field, however it is
tentatively suggested that any increase in plant populations above 25,000 per
ha. (slightly higher than stand populations reported in the survey) are
achieved by increasing plants per stand. Work should aim to identify optimum
densities within these bounding conditions.
It seems possible, given generally reduced vegetative growth of the shorter
term varieties, they will offer greater intercropping potential. This is
likely to be particularly important for cowpeas planted into maize. Work
should check intercropping potential. Cowpea planting should follow local
practice; being placed in between stands in the row, planting increases weeding
problems for farmers. The best density for cowpeas and the time of establishment
in the range 0-6 weeks after the maize crop should both be examined vis-a-vis
yields practice which should be checked from farmers' fields. Separate work
on plant densities and spacing will be needed for the main and the early maize
plantings. A major aim in management of the early plantings will be'drought
resistance. In the main plantings reliability will be achieved by the short
maturity period and the time of planting, a priori much higher populations may
(2) Weeding. One of the most critical issues in the success of the solution
pu't forward in the report is the question of weeding. It is clear from the
survey data that weeding is a major problem for farmers and it is unrealistic
to expect them to intensify their weeding inputs in the peak labour period;
January, February and early March. This will be particularly true for early
plantings of the shorter term selections, weeding will fall into the peak
months of January or February. No more than two weedings can be expected
from farmers, and agronomy work should concentrate on establishing the optimum
timing for two weedings. Recommendations should perhaps be related to the
stage of crop growth, rather than to time from emergence due to the very variable
growth rates likely, of both crop and weeds, under erratic rainfall conditions .
at this time in the season.
Moving the planting date for the main crop back into March should ease the
labour problem for weeding. Much of the labour required will be in April,
a less critical month. The agronomy work should concern itself with an optimum
weeding regime, with emphasis on the timing of the first weeding.
(3) The use of purchased inputs. We have seen that any set of inputs, to be
within the reach of most farmers in the drier areas, should not cost more than
shs.150.00. The general resource endowment pattern in the area, with a
relative abundance of land, suggests that working capital is most productively
used in extending the area farmed by each family. This is the way farmers are
currently using most of the cash spent on their farming operations.
At the same time it is well established that cotton, the main cash crop amongst
farmers in the zone, will not do well without insecticide, the returns to the
use of which are high. An attempt to spread the allocation of very limited
working capital onto major purchases of inputs for maize could be unacceptable
to farmers, or, if subject to heavy political and extension pressure, could
detract from the benefits of tractor and labour hire, or from the momentum of
the drive to encourage cotton grown with insecticides.
It is recommended that initial agronomy work with inputs be restricted to
purchased seed and to stalk borer control if counts show infestations are
seriously threatening yields. Work on storability should include the use of
insecticide and rat baits in stores, all these items are cheap and unlikely to
seriously encroach on present cash spending. However, given the new concen-
tration of population in villages it is possible that fertility levels may
degenerate or that farmers must be less selective in the soil types used for
cropping. A need may arise for fertilisers despite the low man/land ratio,
dressings will have to offer high returns for a limited outlay in order to be
taken up by farmers. Given the risks inherent in planting the early crop it
is not recommended for any fertilizer work. By corollary the comparative
reliability of main plantings of a short term variety in March should reduce
the risks of cash loss and make fertilizer use (if found to offer good returns)
a more secure investment. It is recommended that work is carried out on March
planted selections with a view to fertilizer being a second step in any maize
improvement programme for the zone. Nutrient levels and combinations should be
kept within a cost of shs. 275.00 per ha. (farmers grew an average .56 ha)
to be within the reach of farmers finding themselves having to intensify on
their-existing land, rather than extend their scale of farming. The work should
be looking for an incremental value reliably produced, from the fertilizer alone,
and not via interactions with changes in o-ther minmagecnt pr.tices which cannot
be absorbed by farmers, of the order of 8-900 kgs. p.lia. at present into store
maize prices from an outlay of shs.275.00 per ha. As most farmers grow much
less than a hectare, acre, or even half acre packs in the early years of any
fertilizer scheme, would be more appropriate.
(4) Important interactions in the agronomy work. Certain interactions are
likely to be particularly important in addition to variety/management interactions.
(a) Time of weeding/timing and density of intercrop establishment/maize
stand density. There are likely to be different optima for the
early,and main plantings. With the emphasis on drought resistance
in the early planting and in the circumstances of scarce weeding
resources, the timing of intercrop establishment may play a very
(b) In the second stage of improvement the use of fertilisers on the
main maize plantings is likely to have ramifications on all variables;
the weeding require intercropping timing and density, and the maize
The point to emphasise is how the examination of the farming system has allowed
a close focusing of emperimental effort, for the farmers of the zone. If a
selection and agronomy programme mounted within these guidelines can show
improvements in productivity and reliability, the changes in management proposed
are highly likely to be rapidly absorbed by all levels of farmers in the zone.
q. SOME COMMENTS ON THE DEMONSTRATION
9.1. The team approach
The approach being promoted by the programme is based on an interdisciplinary
team,each member bringing his particular expertise to bear on a defined farming
situation. The approach sees as ideal such a team operationing from the same
research centre, interacting on a day to day basis. For this, and for other
demonstrations, a team had to be brought together on an ad hoc basis. It was
not possible due to other commitments of personnel, physical separation and
financial limitations to create the integrated unit as envisaged by the
approach. A team operating as a unit and interacting on a day to day basis
will achieve much greater insight into the situation of the local farmer and
the sequence of procedures will be much more rapid.
9.2. Survey problems
The National Crop Research Committee gave the go-ahead to discuss the programme.
with ARI Directors in late October 1976. In a perhaps overeager reaction on
the part of Cimmyt the fieldwork for the first demonstration was completed in 1
(1) The timing of the survey was poor. Farmers had begun cultivation for
the 1977 season, questioning had to be related to the 1976 season. The gap
between doing and questioning gave high levels of none response on a signi-
ficant number of questions.
(2) Students from the Department of Rural Economy worked as survey enumerators
as part of their field practical. Teaching requirements were such that two
groups had to be trained separately and each group did about a weeks enumera-
tion. The initial two days enumeration always bring problems and the groups
hardly had the opportunity to settle into their work.
(3) Transport shortages made survey supervision difficult.
(4) A decision was made to leave household plots out of the survey work.
With hindsight it is possible to see that, though the volume of produce from
these plots is small, the timing of the production means it plays an important
part in household food supplies.
9.3. Logistics and Computor processing
Logistical problems of transport and mail arose with the closing of the
Tanzania/Kenya border. Computer facilities in Nairobi were being set up by
Cimmyt and it was decided to use these. There were some initial difficulties
in getting the work onto computer, these were greatly exaggerated by the time
taken for mail between Morogoro and Nairobi. The amount of supplementary
computer analysis undertaken was limited. Data from future Tanzanian demons-
trations will be processed on the computer at the Maths Department, University
Because of the administrative problems of the survey work the data collected
was less comprehensive than is desirable. The legumes in particular were
poorly covered. The turn round time; from initiation to conclusions has been
10 months, rather than the 5 months which would be feasible for an integrated
Leam operating from a local research centre. Despite these shortcomings it is
believed that the approach has been able to offer an appropriate solution to
the identified problems of zone farmers, and detailed guidelines for a research
effort to exploit the solution.
Cimmyt wishes to acknowledge the interest and encouragement from the National
Crop Research Committee and the Chief Research Officer, Dr. Liwenga. The
demonstration was made possible by the interest and cooperation of the Director,
ART Ilonga, Dr. Haulle, and his staff,especially those participating as team
members. Its implementation in the field and successful conclusion, including
the drafting of the report, have been largely the responsibility of Dr. C.D.S.
Bartlett and Mr. Emmanuel Manday; Department of Rural Economy, Faculty of
Agriculture, University of Dar-es-Salaam. Cimmyt thanks them for their major
Uriyo and Singh (1976), Faculty of Agriculture, Morogoro
Bartlett C.D.S. (1977), Research on beans for some Lowland areas of
Morogoro Region (Draft).
CIMMYT International Maize and Wheat Improvement Center
The CIMMYT International Maize and Wheat Improvement
Center holds all rights to the source text and shall be
considered the copyright holder for the text and images of
The CIMMYT International Maize and Wheat Improvement
Center has made this publication available to the University of
Florida, for purposes of digitization and Internet distribution.
The CIMMYT International Maize and Wheat Improvement
Center reserves all rights to this publication. All uses,
excluding those made under "fair use" provisions of U.S.
Code, Title 17, Section 107 are restricted.
Contact the CIMMYT International Maize and Wheat
Improvement Center for additional information and