Livestock Research for Rural Development 12 (3) 2000

Citation of this paper

 Effect of upgrading small East African goats on feed resource utilisation in the Uluguru mountains in Tanzania; a farmers' perspective


Gudrun F Ingratubun, Emyr Owen, Nicolaus F Massawe, Louis A Mtenga and E G Mtengeti

University of Reading, UK and Sokoine University of Agriculture, Tanzania


Participatory methods were used to analyse the feed resources in three highland villages in Tanzania. Farmers assessed feed availability, botanical composition, feeding method and feed intake. The purpose of the investigation was to see whether the introduction of dairy goat crosses had changed the utilisation of feed resources and to establish whether feeding was a constraint.

The major feeding system was tethering goats to eat a large variety of naturally occurring grasses, herbs, shrubs and trees and crop residues in fields after harvesting. Some dairy goat farmers practised herding when sufficient labour was available but 58 % employed additional cut-and-carry from planted forage and all added maize bran to dairy goat rations. Feed availability near the homestead depended heavily on the occupation of land by crops. Rainfall played a minor role. Feed intake and consequently production and reproduction were perceived to be reduced by rainfall related factors such as low DM content, surface water on forage, reluctance of animals to graze during rainfall and worm infestation. Forage species consumed depended on feeding system, occupation of land by crops and ecological factors. Dairy goats were offered a wider variety of species through herding and the additional cut-and-carry. Because of the need to minimise labour requirements in tethering at distant sites, local goats depended more on species on the farmland and in vicinity of the homestead.

The average farm size of both groups of farmers was 3.1 ± 0.3 ha. Households had 2.6 ± 0.2 person-equivalents available for farm work (1.2 ± 0.2 person-equivalents/hectare). Using homestead housing quality as an indicator, dairy goat farmers had significantly more capital at their disposal than local goat farmers. Dairy goat farmers were also slightly better educated and more active in farmers' associations. Dairy goat farmers invested 40 % more time (e.g. herding) in their goats. Investing more capital - to buy maize bran for example – dairy goat farmers gained more than two times higher returns to labour than local goats farmers did. Local goat farmers received almost five times higher returns to variable costs on the little capital they used for the tethering rope. Although dairy goat farmers had made changes in feed resource utilisation, these appeared insufficient to cope with the low intake of forage DM in the rainy season and the increased requirements of dairy goats.

Key words: goats, feeding, season, composition, system, gross margin, participatory methods 


In 1987, the Tanzanian Ministry of Agriculture and Livestock Development reported an annual milk consumption of 15 litres per person which translates to a consumption of only 40 ml per person daily. The National Food Strategy plans to double this annual per capita consumption to 30 litres by 2000 (Mtenga and Kifaro 1992). To achieve this, several strategies were planned, including the introduction of improved dairy breeds, the intensification of the use of locally available feed resources and exploiting indigenous animal breeds' potential for milk production. 

Highland areas seem to be predestined for milk production because of their relatively low ambient temperatures imposing little stress on animals and their reliable rainfall providing sufficient feed. Tchenzema ward on the slopes of Uluguru mountains in Morogoro region is a typical example. The elevation of Tchenzema varies between 1,250 (Mwarazi), 1640 (Nyandira) and 1,675 (Tchenzema) m above sea level and the reliable rainfall ranges between 1020 and 1351 mm annually following a bimodal distribution (Paul 1988; Kiango 1996). A dairy goat project operated from Sokoine University of Agriculture (SUA) was introduced into the area because the keeping of Small East African Goats was common and there has been no tradition of cattle husbandry on the account of the steep terrain. 

In 1988 and 1990 pregnant crosses of the Norwegian landrace and Small East African Goats were transferred from SUA to smallholder farmers in three villages in Tchenzema ward (Madsen et al 1990). Norwegian breeding bucks were supplied by SUA so that the project in Tchenzema ward could be sustained without further supply of does from SUA. At the time of the survey the project comprised 22 farmers keeping 64 does.

Experiences with the project indicated feeding to be a constraining factor for the adoption of dairy goats in Tchenzema ward and feeding was perceived to limit the productivity of does of participating project farmers (Madsen et al 1990). However, no detailed information on the feeding constraint was available. Since most dairy goat farmers previously had kept local goats, the present study considered it interesting to see whether farmers had changed their feeding strategies according to the new breed of goats. 

The present study had the following objectives:

·      Assessment by local goat farmers and dairy goat farmers of the natural feed resources, feeding methods, seasonal feed supply and feed intake leading to the identification of nutritional constraints for goats

·      Identification of reasons for the change in resource utilisation within the socio-economic environment:

·      Characterization of the farming system with respect to dairy goat farmers and local goat farmers

·      Assessment of the objectives and the profitability of the two goat enterprises


It was considered that the findings would be applicable to other relatively densely populated sub-humid areas in the tropics. 


A mixed participatory approach was chosen. This involved participatory methods such as Semi-Structured Interviews and Feed Calendars letting farmers assume an active role in analysing their problems. "Our role is to facilitate their analysis" (Chambers 1996). One to two facilitators and one note-taker took part in conducting the participatory methods.  

The Semi-Structured Interviews, characterised by open-ended questions and casual talk (Theis and Grady 1991), used a checklist on the household structure, farming system, goat production with its objectives and constraints and goat feeding including forage species and periods of feed shortage. It also involved direct observation on the farmers’ wealth through housing quality.

The Feed Calendars (adopted from IIED 1994) made a view over the whole year possible despite the fact that the length of the survey was only five weeks. For the Calendar of Feed Availability and Determining Factors, beans were used as counting tools: five beans symbolizing the greatest amount available, three beans an average availability, which was decreasing down to zero when no bean was allocated. First, farmers agreed on the factors influencing feed supply. Then they allocated beans to these factors. Finally the resulting overall feed availability was assessed by another allocation of beans. While looking at the results, relationships between the availability of the different feeds, the influencing factors and the overall feed availability were discussed. 

Structured Individual Farmers' Interviews, focused through the earlier participatory methods following a questionnaire, were conducted to get quantifiable information on goat feeding methods, production factors and gross margin parameters.

The concluding participatory activity was a Calendar of Ranked Feed Intake and Intake Species Composition. First, the  farmers identified the forage species that contributed significantly to the goat's ration. There was a fixed number of beans ( determined according to the number of species included in the Calendar (two times the number of all species) being the same for every month. Farmers allocated these beans to the relevant species for every month. When all species were consumed in equal amounts, two beans were allocated to each species. If only a limited number of species, and in different relative amounts, were consumed in a particular month, the number of beans to be allocated was agreed on accordingly. The scores of the individual species could be translated into percentages of every monthly ration. The overall feed intake was scored separately following the same scoring system as in the Calendar of Feed Availability and Determining Factors. These made the periods of low intake visible to farmers and facilitators.


Table 1: Methodical techniques and their function






Getting to know each other, creating rapport, identifying key questions and indicators, adjusting further PRA activities, direct observation of wealth




Group meeting

‘Calendar of Feed Availability and Determining Factors’

Amount of feed available for every month, feed supply determining factors, identifying feed shortages






Production factors, profitability of the goat enterprise, goat feeding details




Group meeting

‘Calendar of Ranked Feed Intake and Intake Species Composition’

Relative amounts of feed intake and every feed per month


The completely randomised sample consisted of forty farmers spread over the three villages, half of them keeping local goats and half of them keeping dairy goats. In every village group meetings were held separately for local goat farmers and dairy goat farmers.  

Results and Discussion

Common natural feed resources

Common feeds eaten by both types of goats comprised a large variety of species from the natural vegetation. Farmers differentiated 12 grass species dominated by Cymbopogon caesius, Pennisetum purpureum, Commelina beghalensis and Panicum trichocladum, 27 herbs including a number of legumes such as Glycine wightii, 8 bushes like Pavonia uretis and 14 shrubs and trees such as Morus alba and Albizia gummifera. On harvested fields goats consumed numerous weeds consisting of 5 grass species like Snowdenia petitiana and 11 herbal species such as Bidens pilosa and Gallinisoga perviflora and crop residues like maize stover, bean and pea haulms, sweet potato vines and cabbage stalks. Some of the off-farm species were found on farmland as well.

Feeding methods 

Tethering on common land with natural vegetation, along road sides and on harvested fields was identified as the main feeding system. Local goat farmers allowed their goats to tether for 7.9 ± 0.1 hours whereas dairy goat farmers tethered their goats for only 6.4 ± 0.2 hours. Farmers tethered their animals to save labour (77 % of respondents) and to avoid crop damage (43 % of respondents). Herding was the minor feeding method, practiced either seasonally outside the main cropping season and during the school holidays or throughout the year when children who did not attend school were available for herding. There was a marked change in feeding method from local goat farmers to dairy goat farmers to make up for the dairy goats’ higher requirements. Dairy goat farmers employed markedly more herding (20 % of respondents) than local goat farmers did (X2: P<0.01). The main advantage stated was the provision of feed selection possibilities for the goats (54 % of respondents) and thereby enlarging the accessible feed resources. Additionally cut-and-carry of planted fodder such as Tripsacum laxum, Setaria sphacelata var. splendida and Sesbania sesban and crop residues was practiced by 58 % of the dairy goat farmers if feed from tethering/herding was not sufficient. All dairy goat farmers purchased maize bran and fed on average 1.3 ± 0.2 kg / day to lactating goats and 0.8 ± 0.01 kg / day to non-lactating goats.

Seasonality of feed supply, feed intake and species intake composition 

The Calendar of Feed Availability and Determining Factors showed that all farmers perceived that farm land being available for grazing and rainfall increased the feed availability near the homestead. Feed intake, added from the Calendar of Ranked Feed Intake and Intake Species Composition, was seriously depressed at the start of the short rains and at the peak of the long rains. An abundant feed availability near the homestead increased feed intake. During periods of feed shortages near the homestead, labour demand increased considerably, due to the longer distances to the tethering sites. Figure 1 shows the described cause-effect relationships.


Figure 1: Flow chart of feed intake and feed availability determining factors

Farmers' assessment of the species intake composition varied between dairy and local goat farmers due to the different feeding systems applied by both groups of farmers. The species composition varied between villages for ecological reasons. As an example, only the assessment from farmers in Mwarazi is presented.


The species composition given by the dairy goat farmers in Mwarazi comprised a great number of species, but showed relatively little seasonal variation (Figure 2.1). Crop residues and weeds ranged between 10 and 35 % of the ration, the peak of 35 % being reached in July. This was probably due to the high percentage of herding in this group of farmers, which allowed access by goats to a greater number of species. Since goats were mainly herded in open spaces with natural vegetation and being some distance from the village, there was little dependence on the farmland near the village, resulting in the small seasonal variation.






Local goat farmers in Mwarazi relied heavily on few grass species from off-farm land (Figure 2.2). As soon as crops were harvested weeds and crop residues took over and made up until 65 % of the ration in April and from August to November, when maize was planted again. Apart from the latter the species composition was seen as being very uniform throughout the year.




This analysis of the Feed Calendars showed that the overall quantity of feed was not the predominating problem. However, there were two feeding problems. The first problem was the lack of space for tethering in the vicinity of the homestead leading to an increased labour demand and, like in Tchenzema, even to a decreased intake. This was obviously caused by the respective land being occupied by crops. The second problem was the depressed intake at the start of the short rains and the peak of the long rains resulting in a prolonged kidding interval and relatively low milk production. The cause of the depressed intake was the rainfall and its effects. It was assumed to be a combined effect of the low DM content of the feeds as the majority of feeds ranged between 18.5 to 20.5 % DM in the diet at the beginning of the short rains (Madsen 1987; Jähnig 1996). Low intakes were also attributed to the surface water on the forage (Mtenga and Mtengeti pers. comm.; Devendra and Burns 1983), the rainfall as such (Semenye and Hutchcroft 1992) and a high internal parasite burden (Peacock 1996; Moulton 1988; Shekilango, pers. comm.). There might also have been some interaction in terms of the low intakes making the goats more susceptible to worms (Semenye and Hutchcroft 1992).

The two goat systems in Tchenzema ward 

Dairy goat farmers and local goat farmers described their goat systems by the following characteristics (Table 2). Economic parameters were also calculated.


Table 2: Characteristics of the two goat systems (herd size was 5.5 ± 0.6 for both categories of farmers)


Local goat farmers

Dairy goat farmers


1.Manure production

2. Capital asset

1.Milk production

2.Manure production

3.Capital generation (sale of breeding stock)




Kidding interval months

11.3 ± 0.4

12.6 ± 0.9

Kid mortality, %

6 ± 3 %

17 ± 3 %

Milk yield, litres/day


1.5 ± 0.2 l/d; 449 ± 55 l/lactation; 425 ± 53 l/a






2.Low feed intakes in rainy season


2.Low feed intakes in rainy season

3.Uncertain availability of oil seed cake

4.Costs of feedstuffs




Gross value of outputs, TSh



Variable costs, TSh



Gross margin, TSh






Return to variable costs*



Return to hours of labour**



*Return to variable costs =[(gross value of outputs) - (variable costs)] / (variable costs) 

**Return to labour =[(gross value of outputs) - (non-labour costs)] / (hours of labour input) 

The gross value of outputs of one doe plus followers was six times higher for dairy goats than for local goats. However, since variable costs (maize bran accounting for 83 % of variable costs) were even 20 times higher for dairy goats as compared to local goats, the gross margin for dairy goats was only four times higher amounting to 72,141 TSh and 18,035 TSh, respectively. The relatively high gross margins were strongly influenced by the high value farmers attributed to the goats' manure which was badly needed for crop production. 

The annual hours invested into one adult goat plus followers was 511 hours for dairy goats and 292 hours for local goats. The returns to labour were more than twice as high for dairy goats as for local goats. But returns to variable costs were almost five times higher for local goats than for dairy goats

Availability of production factors and social aspects favouring change of the feed resource utilisation 

The available production factors of the surveyed farms and social aspects of households are summarized in the following table:


Table 3: Production factors and social aspects


Local goat farmers

Dairy goat farmers

Farm size

3.1 ± 0.2 ha

Land-labour ratio, person – equiv. / ha

1.2 ± 0.2

Housing quality



Education, years

5.4 ± 1.1

3.8 ± 1.0

Membership in farmers' association

0.3 ± 0.2

0.7 ± 0.3


Land size was very variable but relatively large and comprised a high number of scattered plots. It did not differ significantly between the two types of the goat enterprises. Soil fertility was declining. All farmers were mainly growing maize, beans, sweet potatoes and coco-yams as staple food crops and vegetables as cash crops in mixed cropping terraces.  

The labour:land ratio was low compared to ecologically similar areas in East Africa having 2.5 to 11 person-equivalents / ha (Bartels 1995; Jakob 1995). As labour was scarce, the high labour input as implied by herding - and cut-and-carry to a lower degree - was only given in the dairy goat enterprise, where returns to labour were high and when children at a suitable age were available. However, many dairy goat farmers and all local goat farmers substituted the high labour input by the comparatively lower capital input of the tethering rope.

Capital as assessed by using the observed homestead housing quality as an indicator, was scarce, showing the clearest differentiation between dairy and local goat farmers (P<0.05). Since capital was scarce, purchased inputs were restricted to a minimum and were mainly employed by dairy goat farmers, maize bran representing the bulk of the variable costs. Dairy goat farmers purchased feeds because milk production, their main objective, relied on these feeds under the existing technology and the value of the outputs justified the costs. This also showed dairy goat farmers' greater awareness of the importance of feeding.  

The higher education obtained by dairy goat farmers in connection with their engagement in various farmers' associations even before the onset of the project might have made them more ready to engage in the dairy enterprise and also change their behaviour with respect to goat feeding. By the employment of herding, cut-and carry and maize bran feeding they widened the traditionally utilised feed resource to lessen the feed constraint in the rainy season. However, the reduction in feed constraint was insufficient as evidenced by the long kidding interval and the relatively low average milk yields. Nevertheless, only 58 % of dairy goat farmers planted fodder, and the area planted with forage was still limited despite the fact that there was land available that would allow more forage crops to be planted.


·        There were two quantitative feeding problems reported by dairy goat farmers and local goat farmers: Firstly there was not sufficient feed in the vicinity of the homestead due to the occupation of land by crops. Secondly feed intakes were reduced at the start of the short rains and the peak of the long rains. This was connected with diarrhoea. The rainfall complex, consisting of an interaction between the high water content of the forages, the surface water on the forage, the rain falling on the goats and internal parasites, was the assumed cause. Farmers were mainly aware of the internal parasites largely ignoring the connection to feeding.

·        The low intakes indicated a need for a feeding management practice that allows for the incorporation of dry forage into the ration in the rainy season. Harvesting and manual box baling of dry forages and crop residues as maize stover and bean straw could help to solve the problem as baled grass/stover/straw could be easily stored and properly budgeted for incorporation in the ration during the rainy season. Techniques for box-baling of dry forages have been described in detail by Massawe (1999).

·        Tethering was the major feeding system only substituted with herding by dairy goat farmers - since returns to labour were higher - when children at a suitable age were available. Methods of improving production of goats by tethering have been described by Sendalo et al (1996). Additionally dairy goat farmers tended to employ cut-and-carry of some planted feeds and crop residues. All of them fed maize bran since dairy goat farmers had more capital (P < 0.05) at their disposal than local goat farmers. The higher education and engagement in farmers' associations encouraged them to make this change.


We would like to thank the Arthur Hosier Award of the University of Reading for sponsoring this study which was in partial requirement for the MSc degree in Tropical Agricultural Development - Option in Animal Production, University of Reading, UK. Our sincere gratitude also goes to Rustica Mkwidu, Wiliam Shekilango and all participating farmers at Tchenzema ward for their warm and ready assistance during the field work.


Bartels M 1995 Dairy goat adoption in. Central Kenya: Influence of labour constraints and preferences.MSc, Humboldt Universität Berlin, Germany.

Chambers R 1996 Relaxed and Participatory appraisal. Notes on practical approaches and methods. Notes for participants in workshops of the first quarter of 1996, Institute of Development Studies, University of Brighton, UK.

Devendra C and Burns M 1983 Goat production in the tropics. Commonwealth Agricultural Bureaux, Slough, UK.

IIED 1994 RRA Notes No.20. Special Issue on Livestock.  International Institute for Environment and Development London, UK.

Jähnig G F 1996 Production of Small East African Goats and Norwegian Crosses in the Uluguru Mountains in Tanzania, MSc dissertation, Reading, UK 

Jakob U 1995 Ziegenhaltung bei Kleinbauern in Burundi. Wissenschaftsverlag Kiel KG, Germany. (Sozialökonomische Schriften zur ruralen Entwicklung) Bd. 113 zugleich-Berlin Humboldt Universität, Dissertation, 1994. 

Kiango S M 1996 Studies on Factors Affecting Performance of Diary Goats and on Socio-economic aspects of Dairy Goat Production in Tchenzema and Dareda Wards in Tanzania. Department of Animal Science and Production, SUA, Morogoro, Tanzania 

Madsen A 1987 Nutritive values of grasses and leaves used for feeding goats in Mgeta, Department of Animal Science and Production, SUA, Morogoro, Tanzania. (unpublished)

Madsen A, Nkya R, Mtenga L A and Kifaro G C 1990 Dairy goats for small-scale farmers: Experiences in Mgeta highland. In Proceedings of the 17th Scientific Conference of the Tanzania Society of Animal Production. Vol. 17, p 48-57, Tanzania.

Massawe N F 1999 Strategies based on participatory rural appraisal for improving the utilisation of forages to increase profitable milk production on smallholder farms in Tanzania. PhD, University of Reading, UK.

Moulton M 1988 The potential for dual purpose goats in the village of Nyandira, Tanzania. M.Sc., Agricultural University of Norway, Norway 

Mtenga L A and Kifaro G C 1992 Dairy goat research and extension at Sokoine University of Agriculture (Lowlands) and Mgeta (highlands) areas of Tanzania. In: Future of livestock industries in East and Southern Africa.  Proceedings of a workshop held at Kadoma Ranch Hotel, Zimbabwe, 20-23 July 1992 (Editors: J A Kategile and S Mubis), ILCA (International Livestock Centre for Africa) Addis Ababa, Ethiopia 

Paul J L 1988 Farming Systems in Upper Mgeta. Franco-Tanzanian Project Report to CIMADE, SUA, Morogoro, Tanzania.

Peacock C 1996 Improving Goat Production in the Tropics. An Oxfam/FARM-Africa Publication, UK.

Semenye P P and Hutchcroft T (Editors) 1992 On-farm research on technology for dual purpose goats. Small ruminant collaborative research support Programme,  Maseno and Kabete, Kenya.

Sendalo D S C, Owen E, Mtenga LA, Romney, D L, Penning, P D and Mayes R W 1996 Tethering of grazing goats in Tanzania: effect of tethering for four or eight hours daily on intake and milk yield of Norwegian cross local goats grazing Brachiaria in the wet season. Proceedings of the 22nd Scientific Conference of the Tanzania Society of Animal Production 22, 295.

Theis J and Grady H M 1991 Participatory Rapid Appraisal for Community Development. IIED, London, UK.


Received 29 March 2000


Go to top