Previous Page Table of Contents Next Page

The use of diagnostic surveys in directing on-farm research: The experience of the smallholder dairy feeding systems project in Tanzania

Gilead I. Mlay
Department of Rural Economy
P O Box 3007, Morogoro, Tanzania

Results and discussion


The survey was conducted in the Hai district of Kilimanjaro Region in February/March, 1984. The main objective of the survey was to identify those constraints related to livestock feeding that can be addressed by researcher designed innovations within the bounds defined by available resources and the institutional set-up.

The survey results indicate that natural grass, maize stover and bean haulms are the main feeds used. High transport costs are involved in moving the crop byproducts from the crop fields to the homesteads where livestock are kept. The shortage of concentrates and/or high costs involved limit their use. Molasses is available in large quantities but high transport costs have limited its use up to the time the survey was being conducted. A project is currently underway to construct storage tanks/selling centres in the villages and this should lead to a substantial reduction in costs. Research should focus on improving the nutritive value of the byproducts or the introduction of legumes. The economics of baling the byproducts and providing large volume transport to user centres should also be investigated.


Technological packages recommended to farmers are, in general, a result of research efforts at research stations. It is true that testing of such packages takes place in localities where they are intended to be used, but the general practice has been to test these in researcher-controlled environments only. Involvement of the farmer is rare.

There is no doubt that the effectiveness of technological packages recommended to farmers will be dependent on the extent to which the following factors are taken into account during their formulation (Hardaker, et al 1985):

* The institutional setting and policy environment in which farming is conducted;

* The economic environment of farmers, including long term market prospects for inputs and outputs;

* Attitudes and personal constraints of farmers, including their desires or otherwise for change;

* The knowledge that the technological package is addressing already identified breakable constraints and exploitable opportunities within the existing systems.

The increasing interest in adaptive, on-farm research is a reflection of the failure of the more traditional approaches to solve the problems of low productivity in peasant agriculture. Under the umbrella of farming systems, methodologies have been proposed for conducting on-farm research (Zandstra, et al, 1981; Collinson, 1982).

Diagnostic surveys play a key role in identifying constraints that can be addressed by research and in providing base line information about the target farmers. The smallholder Dairy Feeding Systems Project in Tanzania has a large component of on-farm research and a diagnostic survey was used to identify constraints and tentative innovations for on-farm testing.

Choice of the study area

Kilimanjaro Region was chosen as the study area because its climate is well suited to dairy production. The region also has the highest and most advanced smallholder dairy production in the country. The land constraint and the production systems used in the region pose special problems for livestock feeding. The survey and the subsequent on-farm experiments-were limited to the Hai district because of constraints on time and resources.

Description of the study area

Hai district is one of the six administrative districts of Kilimanjaro Region in the northern part of Tanzania. The district covers approximately 2308 km2 and has a population of 172 317 and a population density of about 74.7 per km2 (1978 census).

Two main zones can be identified; the highland zone and the lowland zone (Japan International Cooperation Agency, 1977). The highland zone enjoys milder temperatures and heavier, more reliable rainfall than the lowland zone. Most smallholder farmers in the district reside in the highland zone in permanent homesteads. Coffee and bananas are their main crops. Livestock raising -in the highland zone is of the stall feeding type, with most animals being dairy cattle. Many parts of the lowland zone were developed only as recently as the 1960s. The main crops grown in this zone are maize, beans and finger millets. Livestock is mainly beef cattle. Most farmers residing in the highland zone also own farms in the lowlands and commute daily in the crop growing season to work on the farms.

Objectives of the survey

The main objective of the survey was to identify constraints related to livestock feeding that can be addressed by researcher designed innovations. To achieve the main objective, information was sought to describe a farmer in terms of livestock numbers and performance, crops grown and yields, purchased inputs and marketed output, use of home grown feeds and byproducts and constraints to increased production.


Data were obtained from primary sources using two techniques. Informal interviews were conducted in December 1983 involving district and village officials and extension officers. In addition the managements of a sugar processing company and a livestock feeds company were interviewed. The objective of these informal interviews was to obtain information on:

* Institutional and policy settings that have a bearing on dairy production in the district;
* Inputs and output marketing arrangements;
* Livestock feed situation in the Kilimanjaro Region;
* Perceived constraints to smallholder dairy production in the district.

A formal survey was conducted in February/March 1984 using a structured questionnaire.

Survey design

A single household keeping cattle was used as the sampling unit in this study. A list of households keeping livestock was obtained from each village and these were used to construct a sampling frame. Thirty households were sampled from each of the five divisions of the district, forming a sample size of 150 households. A questionnaire was constructed to obtain the following information:

* Demographic characteristics of a household;
* Number of farm plots and distance from the homestead;
* Livestock types and numbers;
* Crops grown and their yields;
* Livestock feed sources and prices where applicable;
* Milk production and marketing;
* Perceived constraints to crop and livestock production;

Information was sought for the year 1983. The interviews were conducted by the researchers.

Results and discussion

Land constraint in the district precludes the use of technologies that make large demand for land. Table 1 presents results of farm size distribution across sampled households. All farmers sampled have farm plots in both the highland and lowland zones. The classification of farms by zone has been used because of the distinct differences in the crops grown in the two zones.

The results show that the majority of the sampled households have farms of less than a hectare in each of the two zones, but the farms on the lowland zone are on average larger than those in the highland zone. However, there is a large variation in farm size across sampled households as indicated by the standard deviation figures. There is a wide variation in distance from the homestead to the lowland farms. The survey results indicate an average distance of 18.8 km. This has serious implications for the cost of transport of farm products from these farms to the homesteads.

Table 1. Frequency distribution of farm size across households.

Range in hectares

Percent of households

Highland zone

Lowland zone


























Standard deviation




N = 107
Source: Diagnostic survey.

Livestock production

Livestock production is an integral part of the farming system in the district. Land scarcity has contributed significantly to the high degree of dependence between the crop and the livestock subsystems. While the crop byproducts are extensively utilized as livestock feed, the manure from the livestock is in turn used on the banana/coffee and vegetable plots to maintain soil fertility.

Livestock numbers and herd size distribution

The 1978 census indicates that numerically, cattle are the most important livestock in the district, followed by sheep, goats and pigs respectively (Table 2). Traditionally, cattle have been kept to meet domestic milk needs and customary requirements such as payment of dowry. However, the rapidly growing demand for milk (as reflected by the high parallel market price of Tz shs 20.00 compared with the official price of Tz shs 10.00 per litre), coupled with the declining farm income from coffee resulting from falling real prices and persistent out-breaks of coffee diseases, has seen a rapid growth in commercial milk production by smallholder farmers in the district. The local Zebu is a low milk producer and is being replaced by improved breeds.

Table 2. Number of domestic animals in Hai district, 1975.



Average per household













N = 107
Source: Adapted from Mlambiti, M.E., et al, 1982. Economic analysis of the Kilimanjaro Region - Tanzania. IAF No. 85. West Virginia University, p.13.

Table 3 presents cattle herd size distribution across sampled households. It is clear that among the sampled households the replacement of the local Zebu by improved breeds is almost complete. Farmers interviewed indicated that the herd top size of cattle is constrained by feed availability and the associated high costs of acquisition.

Table 3. Cattle herd size distribution by type across households.

Number of animals

Percent of households with:


Improved cows

Improved bulls

Improved growers

Improved calves









1 - 2







3 - 4







5 - 6




























Standard deviation







N = 107
Source: Diagnostic survey, 1983.

Cattle breeds and breeding practices

A wide variation of pure breeds and cross breeds was observed among sampled farmers. This variation is a reflection of what was available on the market when the farmer purchased his heifers, or the bull service/semen available at that moment for upgrading the local zebu. Often farmers keep more than one breed, as seen in Table 4.

Table 4. Distribution of improved breeds across households, 1983.


Percent of households









Two pure breeds


Three pure breeds


Pure breeds and crosses


N = 107
Source: Diagnostic survey.

Most farmers indicated that, given a choice, the preferred breed is Jersey. The justification given is its relatively low feed requirement and the high butter fat content of the milk.

The survey results indicate that 63.1% of the farmers use natural breeding, 10.7% artificial insemination and 26.2% use both methods. Although natural breeding methods are commonly used, 92.6% of the sampled farmers do not own bulls (Table 3). Such a service is obtained either from neighbours or cooperative farms, at a cost. The low utilization of artificial insemination is due to the shortage of semen and equipment and the small number of veterinary extension staff who, in most cases, are not provided with transport. As a result the artificial insemination service has become very unreliable.

Information was also sought on the source of replacement stock. The survey results indicate that 58.4% of the sampled farmers raise their own replacement stock, 18.8% purchase replacement stock from government farms and other farmers and 22% utilize both sources. Scarcity of improved replacement stock on the market accounts for the large number of farmers raising their own replacement stock.

Feeds and feeding practices

Stall feeding of livestock is the most common practice in the highland zone of the district. Since agriculture is so intensive, there is hardly any land suitable for open grazing. The main sources of roughage are natural grass, maize stover, bean haulms, banana pseudo-stems and leaves, and established pasture plots. These are fed ad libitum when available. Minerals, in the form of naturally occuring sodium salts (magadi), are commonly sprinkled on roughage to improve palatability.

1. Natural grass

The reliability of natural grass as a source of roughage is restricted to wet seasons. Limited supplies are obtainable within the highland zone on unusual patches of land and along river banks. The lowland zone has a higher potential to supply grass but the poor condition of roads in the wet season limits its exploitation.

2. Crop residues

It is argued that the shortage of grass in the dry season is one of the major causes of drastic deterioration of livestock nutrition. Inevitably, crop residues, mainly maize stover and bean haulms become very important in the dry season. These are obtained mainly from the farms in the lowland zone.

The survey results indicate that 86% of the sampled farmers use both maize stover and bean haulms, while 9.7% use maize stover only and 4.3% use bean haulms only. While 67.7% of the sampled farmers depended entirely on crop residues from their own farms, 33.3% supplemented their crop residues with purchases from other farmers. The main constraint in obtaining these residues is transport. The shortage of transport, coupled with high hiring costs, has caused many farmers to rely on family and/or hired labour to head-carry the residues to their homesteads.

Due to the wide variation in the modes of transport used and in the sizes of loads carried by head, it was not possible to estimate the amount of residues stocked per household per year. Estimates of costs associated with purchases and transport of crop residues for the year 1983 were obtained from only 47 farmers. These ranged from TZ shs 200.00 to TZ shs 5200.00 with a median of Tz shs 1755.00.

3. Established pasture

Farmers in the study area recognize that natural or established pasture is an economic input and are willing to, and do, pay for it. Despite the scarcity of land, an increasing number of farmers are growing their own pasture to supplement other sources. The majority of surveyed farmers had established pasture along footpaths and farm boundaries and in only a few cases had they set aside land for pasture establishment.

Many farmers indicated a willingness to replace part of their coffee plantations with grass if such an opportunity was provided. The grass species grown are Pennisetum purpureum (elephant grass), Tripscam laxum (Guatemala grass) and Setaria spp.

4. Standing hay

Use of standing hay is not a common practice in Hai district and only 13% of the sampled farmers used it in 1983.

5. Concentrates

The main, formal, retail sources of concentrate feeds are cooperative shops and the Tanzania Farmers Association. These sources supply maize bran, wheat pollards, cotton seed cake and dairy complete meal. Molasses is also available in large quantities from the Tanzania Sugar Company, located about 20 km south of Moshi. However, there is a critical shortage of concentrates which means that their use is below requirement. In 1983 only 59.8% of the sampled farmers fed concentrates to their livestock.

While molasses is available in large quantities, high transport costs have restricted its use. Work is in progress to construct storage tanks in selected village centres for stocking a molasses/urea mixture. The project is being supported by FAO/UNDP Dairy Development Services. To date, 19 centres have been established and are operational and this means that many farmers are now able to purchase the mixture in small quantities within walking distance of their homesteads.

Dairy Cattle Performance

The performance measures investigated in the survey include daily milk yield, lactation length and calving interval. For the purpose of this paper only milk yield performance is reported. As shown in Table 5, the majority of sampled farmers realized an average daily milk yield of between 5 and 10 litres. The mean daily milk yield was 7.3 litres. Farmers interviewed indicated that the lowest milk yield is in the dry season when grass is in short supply.

Table 5. Frequency distribution of average daily milk yield, 1983.

Range in litres

Percent of households per day











Standard deviation


N = 107
Source: Diagnostic survey.


Assuming that the survey results represent the true, overall picture, then the constraints related to feeds are:

* The high cost of transport of crop residues and grass from the lowlands to the homesteads;

* The low nutritive value of feeds used, particularly in the dry period.

With the re-introduction of cooperatives, there is a need to investigate the cost implications of baling and large-scale transportation of bales to the utilization areas. Research should be directed towards improving the nutritive value of crop residues since they form a very significant part of livestock feeds in the district. The results show that the established pastures do not include legumes and their establishment under the existing farming system should be studied.


Hardaker J B. Anderson J R and Dillon J. 1985. Perspective on assessing the impacts of improved agricultural technologies in developing countries. Australian Journal of Agricultural
Economics 128(2 and 3):87-108.

Collinson M P. 1982. The use of farming systems research for understanding small farmers and improving relevancy in adaptive experimentation. Intercropping. Eds. Keswan C L and Ndunguru B J. IDRC. 186e pp. 121-125.

Japan International Cooperation Agency, 1977. Kilimanjaro Region integrated development plan; summary report.

Zandstra H G. Price E C, Litsinger J A and Morris R A, 1981. Methodology for on-farm cropping systems research. IRRI.

Previous Page Top of Page Next Page