M.S.L. Kumwenda and H.D.C. Msiska
Ministry of Agriculture
Department of Agricultural Research
Chitedze Agricultural Research Station,
P.O. Box 158, LILONGWE, Malawi
Abstract
Introduction
Review of research to date
Materials and methods
Results
Discussion
Conclusion
References
On-station research work has developed promising animal feed resource technologies for increasing milk production and these are being evaluated on-farm with both research resources and those of farmers. Preliminary results show that the inclusion of cottonseed cake in maize bran-based dairy rations would raise dairy farmers' gross margin/cow. If the programme will be successful, the utilisation of research results will eventually lead to more feed resources of high quality resulting in increased livestock productivity with the end result of attaining food security and improved farm income of the smallholder farmers.
Crop production systems and animal feed resources in Mzuzu
Pastures
In Malawi, smallholder dairy farming was initiated by the Food and Agriculture Organisation (FAO) through the agency of the United Nations Development Programme (UNDP) in the early 1970s. The programme was set up in order:
1. to provide fresh milk for the increasing population
2. to reduce imports of milk by-products, and
3. to provide alternative sources of income to farmers
The national aim of the dairy development programme is to achieve self-sufficiency in milk and milk products. Recent figures show that production And consumption of milk And milk products remains low; per capita consumption from the national herd is estimated at approximately 6 kilogrammes (kg) per year compared to 28kg per year-for the Eastern and Southern African region as a whole, and to 100 kg per year for European countries (Anonymous, 1987). At the same time, there is an urgent need to improve national human nutrition levels. The pressure on land is increasing, and the population continues t increase rapidly at the annual growth rate of 3.0% (Anon, 1988).
The government of Malawi has responded to these problems by promoting smallholder dairying. The dairy programme is concentrated in places designated as milkshed areas. There ar three such milkshed areas: Blantyre in the Southern region wit 700 dairy farmers, started in 1970; Lilongwe in the Central region with 400 dairy farmers started in 1972; and Mzuzu in the Northern region with only 80 farmers, started in 1977 (Nkhonjera, personal communication).
Mzuzu milkshed area is situated on the Viphya plateau at an altitude of 1200 metres. The area has a mean maximum temperature of 21°C; and a mean minimum temperature of 18°C. The rainfall which is influenced by the South East Trade winds blowing across the lake Malawi ranges from 1200 - 2300 millimetres (mm) annually. About 70% of the rain falls between December and March while 125-250 mm falls between May and October in most places of the milkshed area (Anonymous, 1982). This extended rainfall season provides an ideal environment for evaluating late maturity legumes which tolerate acidity (Msiska, personal communication). The weather for this area is also conducive for silage making; and it makes the pastures remain green throughout the year. The area is characterized by Brachystegia woodland savanna vegetation while the soils which are fairly acidic (pH.5.1) are mostly deep dark red uniform sandy-clay ferrisols.
Major crops grown in the area include maize, beans, citrus fruits, and bananas; whilst cassava, groundnuts, potatoes, coffee and finger millet are grown in small hectarages. Table 1 contains estimated average cultivated areas of the crop enterprises for a few selected farms in Mzuzu.
Table 1: Estimated average cultivated areas of crops enterprises for selected farms in Mzuzu
|
Crop enterprise |
Average cultivated area (ha) |
|
Maize |
2.5 |
|
Beans |
1.7 |
|
Fruits (citrus) |
2.2 |
|
Bananas |
0.7 |
|
Groundnuts |
0.4 |
|
Cassava |
0.5 |
|
Fingermillet |
0.3 |
|
Potatoes |
0.2 |
|
Coffee |
0.4 |
Sources: Cusack, T.J. 1988. Report of a Reconnaissance Survey in Mzuzu milkshed area (unpublished).
Apart from dairy cattle, farmers in Mzuzu also keep poultry, goats and rabbits. For dairy production, two systems of grazing are practiced. One system is based on zero-grazing (cut-and-carry) in which dairy animals never leave the stalls and fodder/crop residues are taken to them throughout the year. This system is not common in Mzuzu because the majority of the farmers graze their animals on natural pastures. The other system is open grazing. In the case of Mzuzu, this system is inefficient because animals have to walk long distances to grazing areas from the homestead. They are taken back to the homestead for milking. However, the majority of farmers practice both systems, especially those with some established improved pasture plots in addition to natural pastures.
Feed resources available in the area fall into three classes:
1. Pastures:
(a) Natural2. Agro-industrial by-products and crop residues and
(b) Improved planted
(a) Natural grazing
This represents the major feed resource in the milkshed area. Grazing on natural pastures is commonly practiced; as a result not much effort, is made to establish improved pastures. Farmers consider that pasture establishment is not essential since there is unlimited natural grazing land.
(b) Improved established pastures
Very few farmers have improved pastures; such that they practice both zero-grazing and 'open' grazing on natural pastures. For those farmers who have improved pastures, grass is cut-and-carried to the stalls for a short period of time. After having exhausted these resources animals are taken out to natural lands. The reason for this is that farmers establish small plots of pasture which do not match with the size of their herd.
The use of Napier (Pennisetum purpureum) is proving most productive for those farmers who are able to utilise it properly; however, many farmers are wasting this resource by subjecting it to overgrazing or letting it overgrow and later cut for bedding material in the stalls. Other grasses grown in this area include Rhodes grass (Chloris gayana), Hamil panic (Panicum maximum) and Guatemala (Tripsacum laxum). In this area Rhodes grass does not seem to be persistent and is not productive enough under present smallholder management conditions. Legumes commonly grown in the area include Glycine (Neonotonia wightii), Greenleaf (Desmodium intortum), Silverleaf (Desmodium uncinatum) and Stylo (Stylosanthes guianensis). However, farmers rarely have these forage legumes on their farms. Those who have them have planted them as small plots of pure stands of one or several species.
Agro-industrial by-products and crop residues
Feed resources included in this category are:
(i) Maize bran
(ii) Cottonseed cake
(iii) Groundnut cake
(iv) Maize stover and groundnut haulms
Other agricultural products: In this category there are banana leaves and pseudostems; and small quantities of potato vines. A commercial concentrate, dairy mash from Limbe in the southern part of the country when available is used in the feeding systems.
Feed shortages during the dry seasons and sometimes during the wet seasons constrain dairy production in the Mzuzu milkshed area. A large portion of this area is under natural pastures whose nutritive values are very low. In addition, the main constraints for natural pastures are:
- seasonal and low productivity including the conspicuous absence of legume species,
- lack of proper management, including serious overgrazing
- the traditional communal land tenure systems which hamper efforts to improve the grasslands: and
- the inability to effectively carry out destocking
Improved pastures can increase milk production in the region. However, farmers lack the skills in managing these pastures once established. Weeds are left to grow together with the pastures; and the pastures are not fertilized. In most cases the blame is laid on the farmer, but the extension workers take the leading role in the improper management of the pastures because of lack of proper advice and guidance given to farmers. To resolve the constraints, the following could be done:
- need to device a production system that will ensure a continuous supply of forage throughout the year
- need to device suitable technology of forage conservation for the small-scale farmer, and
- need to intensify the production of pasture seeds for both promising grasses and legumes
All these should go along with training of both extension workers and farmers on the proper management of these improved feed resource.
The high costs of commercial supplements and also their scarcity at certain times limit their use in dairy rations. Once these supplements are bought the feeding is not properly done. They are fed in such limited amounts that farmers do not reap the full benefits of supplementing. Farmers also lack facilities for preparing feeds such as cottonseed cake in the form that could be utilised by the animal.
Most experiments on dairy production in Malawi have been conducted to investigate alternative sources of nitrogen (N) to be fed with maize bran. Maize bran is one of the most easily obtainable and cheap feeds available in Malawi. It is low in crude protein (10% CP) and if fed alone, it cannot meet the nutritional demands for protein of high milk producing animals. One of the agro-industrial by-product feeds that has been found to give favourable milk yields when fed in combination with maize bran, is cottonseed cake. Cottonseed cake is a valuable high protein feed for mixing with carbohydrate feeds for cattle. It is high in fat content. It is one of the richest feeds in phosphorus, containing 1.0% or-more of that important mineral (Topps, 1961).
All previous work on the utilisation of maize bran alone and in combination with other concentrates such as cottonseed cake has been done on government research stations. With high quality grass forages fed ad libitum, a ratio of maize bran to cottonseed cake that has been found most economical is four to one, respectively. A series of feeding trials involving different proportions of maize bran and cottonseed cake has shown that as cottonseed cake levels in the maize bran-based dairy rations increase, concentrate feed costs also increase. The aim of varying the proportions of maize bran and cottonseed cake was to come up with a higher crude protein level of the proportion than that of maize bran alone and at the same time arriving at the maize bran and cottonseed cake ratio that is cost-effective. Although the current mixing proportion of maize bran and cottonseed cake is four parts of maize bran to one part of cottonseed, this is not the final recommendation. Further studies are in progress to investigate the proportions of feeding maize-bran and cottonseed cake to lactating cows that could be cost-effective for small-scale dairy farmers. Although CP levels are increased with subsequent increases in levels of cottonseed cake (Table 2), the performance of cows in terms of milk production varies considerably depending on the quality of the forages fed. Table 2 contains the chemical composition of one batch of samples of different proportions of maize bran and cottonseed cake.
Table 2: Chemical composition of maize bran and a mixture of maize bran and cottonseed cake (% of DM)
|
Component |
Proportions of maize bran and cottonseed cake* | |||
|
|
1 |
2 |
3 |
4 |
|
Dry matter |
90.0 |
90.3 |
91.3 |
93.0 |
|
Crude protein |
10.9 |
18.8 |
20.3 |
25.1 |
|
Neutral detergent fibre |
26.6 |
27.2 |
26.9 |
29.7 |
|
Acid detergent fibre |
8.2 |
9.9 |
10.7 |
11.7 |
*1 = 100% maize bran, 2 = 80% maize bran plus 20% cottonseed cake, 3 = 60% maize bran plus 40% cottonseed cake; and 4 = 50% maize bran and 50% cottonseed cake
Source: Kumwenda and Munthali (1988)
The development of high quality feed resources in the form of improved forages is one of the research objectives of the pasture research programme in Mzuzu as well as in other parts of Malawi.
Past research work carried out at Lunyangwa Research Station and Choma Veterinary Livestock Multiplication Centre identified grasses (Napier, Bana and Giant panic cv Ntchisi) and legumes (Silverleaf desmodium, Tinaroo glycine, Cook stylo and Macrotyloma) which could be successfully grown on the acidic soils of Mzuzu. Work has also been conducted on agronomical practices of managing the improved pastures. These involved the use of fertilizers and manures. Table 3 contains dry matter yields of Rhodes grass with no fertilizer or manure applied, manure only, fertilizer only and manure plus fertilizer from the work done at Choma.
Table 3: Rhodes grass yields (dry matter 1st year) by four treatments
|
Treatment |
Yield (kg DM/ha) |
|
Control |
1090 |
|
Manure only |
1720 |
|
Fertilizer only |
3420 |
|
Manure + fertilizer |
3030 |
Source: Katuma (unpublished)
Results in Table 3 demonstrate that fertilizers and manure can play significant roles in increasing pastures dry matter production under these acidic soil conditions.
The available technology for pasture production is described in the Pasture Handbook for Malawi (1983 edition). A summary of crude-protein values and organic-matter digestibility of prominent pastures in Malawi is given in Tables 4 and 5.
The crude-protein content of young and improved grasses is higher than that of unimproved forages (Table 4) and that of legumes is even higher. Therefore, the inclusion of legumes in grass pastures would improve the overall nitrogen content of the forages. The digestibility of improved grasses is also much higher than that of unimproved grasses (Table 5).
Table 4: Crude protein contents of natural grasslands and improved forage species in Malawi
|
Species |
Crude-protein content (% of DM) | |
|
Natural grasslands | ||
|
- dry grass |
- wet season |
11.0 |
|
|
- dry season |
2.0 |
|
- "Dambo" grass |
- young |
6.0 |
|
|
- mature |
3.0 |
|
Improved grasses | ||
|
Rhodes |
- young |
13.8 |
|
|
- mature |
4.2 |
|
Green panic |
- young |
14.4 |
|
|
- mature |
4.1 |
|
Ntchisi panic |
- young |
12.3 |
|
|
- mature |
4.1 |
|
Improved legumes | ||
|
Tinaroo Neonotonia |
- young |
23.1 |
|
|
- mature |
14.4 |
|
Endeavour stylo |
- young |
19.4 |
|
|
- mature |
13.6 |
Source: Ministry of Agriculture (1983). Pasture Handbook for Malawi, Lilongwe, Malawi
Table 5: Digestibility of some natural grassland species and improved forage species
|
Type of forage |
Organic-matter digestibility at the end of wet season (% DM) | |
|
Natural grasslands | ||
|
|
Hyparrhenia spp. |
31 |
|
|
Sporobolus spp. |
30 |
|
Improved grasses: | ||
|
|
Giant Rhodes grass |
63 |
|
|
Ntchisi panic grass |
61 |
|
|
Common guinea grass |
62 |
|
|
Bushmine panic grass |
65 |
|
Mature forage (Standing hay) | ||
|
|
Buffel grass |
35 |
|
|
Rhodes grass |
40 |
|
|
Napier grass |
35 |
|
|
Joint vetch |
58 |
|
|
Siratro |
47 |
|
|
Neonotonia |
52 |
Source: Ministry of Agriculture (1983). Pasture Handbook for Malawi, Lilongwe, Malawi
Crude protein and organic-matter digestibility are all higher for improved pastures than natural pastures.
Bana Napier grass which grows very well in Mzuzu, could be one of the grasses to be pushed to farmers. Although data is not available in terms of its nutrient content, work done in Kenya shows that Bana Napier grass has crude protein percentages ranging between 8.6 - 14.0, acid detergent fibre 3.3 - 5.7; and digestibility of dry matter 56-72%. It yields at least two tons of dry matter per hectare and could support 0.4 kg/day of weight gains for heifers and 10kg of milk/day (Anonymous, 1985).
Another forage technology that has shown to provide a cheap feed resource of high nutritive value to dairy animals is the undersowing of forage legumes in cereal crops such as maize (Dzowela, 1987). The undersown legumes are utilised together with maize stovers during the dry season in situ or in feeding stalls.
Although the utilisation of maize bran and cottonseed cake in dairy rations has proved a promising technology on station this technology has rarely reached the farmer. Best-bet forages are available. These too have not been widely taken up by farmers. It is not surprising that in the Mzuzu milkshed area, milk yields per lactation are still low, and calving intervals are long. These are associated with poor nutrition. There is also critical feed shortages during dry seasons. The reasons for the failure of farmers to take up research generated technologies are not clear.
In view of the foregoing, it was proposed to undertake a study of smallholder dairy farmers. Initially the programme is aimed at evaluating the economic benefits and on-farm milk yield response to the utilisation of maize bran and cottonseed cake and introduce and evaluate the best-bet forages as animal feed resources. The overall objective is to promote or speed up the adoption rate of these technologies.
This paper reviews on-station research done in the use of maize bran and cottonseed cake in dairy rations and development of forages as animal feeds. On-farm evaluation work of these on-station generated technologies is outlined and the possible utilisation of the research results in dairy production and soil fertility maintenance contexts of the smallholder producers is discussed.
Experimental animals and feed resources
Experimental measurements
Evaluation of improved forages on - farm
Utilisation of maize bran and cottonseed cake in feeding trials: Farmers and location of trials
Farmers in the four bulking groups (dairy farmers associations) were selected and these were divided into two groups (1) a group of farmers with ten animals in total utilising pasture feed resources supplemented with maize bran, (2) a group of farmers with ten animals in total utilising pasture forage resources supplemented with a mixture of maize bran and cottonseed cake.
The criteria for farmer selection was based on:
a. a full time average farmer
b. willingness of farmer to participate in the research management to collect reliable data and exchange experiences with other animals
c. accessibility of his/her farm holding
Each treatment involved ten animals. The animals were those that calved at about the same time. The animals were between two and three lactations averaging seven years in age. Most of them were 3/4 Fr × MZ (Friesian × Malawi zebu). Farmers provided the forage feed resources, whereas the researchers provided maize bran and cottonseed cake. The farmer got these feeds free. The supplements were fed at milking time at the rate of 1kg per 2.5 kg of milk produced per cow. All animals were supplemented with maize bran two weeks before calving. The mixtures of maize bran and cottonseed cake was in the ratio of four parts to one part, respectively.
Farmers had their buckets calibrated for the quantities of the supplements to be fed based on milk production. Most farmers had calibrated cups for measuring milk. Those who did not have, had to use cups that they used for their drinking water; these had to be calibrated too. Daily milk yields were measured and the amount of supplements offered and consumed were recorded.
Feed and forage samples were taken for chemical analyses. Labour costs were estimated. The farmers were given folders with forms in which records were entered. The forms had columns for recording any activity done by the veterinarian and an Artificial Inseminator.
Gross margins were calculated to assess the economic benefits of utilising maize bran and cottonseed cake.
Best-bet forage resources will be established on the farmers, fields involved in the feeding trial and feed budgeting to ensure a continuous supply of forage throughout the year and suitable technologies for forage conservation for dry season feeding will be developed. Production of pasture seeds for both promising grasses and legumes will be another activity for the researchers in the milkshed area.
This programme has just started and the results included in this paper are for one phase of a series of trials, of which the second trial is in progress. Data is collected for a period of five months for each trial.
Data collected on milk production and concentrate consumption, estimates on transport and labour costs were used to calculate gross margins of the two treatments. However, milk production on these farms was affected by several factors that were difficult to quantify.
Major problems with on-farm testing include;
a. several levels of factors cannot be implemented within a farm
b. limited number of animals
c. the animals may vary in breeds/crosses, age and number of lactations
d. the basic management by the farmer vary considerably, and
e. success of the testing depends on willingness of the farmers
With the first feeding trial, a few problems were experienced. Some farmers gave the feed for the animal on trial to other animals in the herd. Some irregularities in recording were experienced. It was not possible for researchers to visit every farmer at least every week. The dairy extension assistants (DEA) were given the responsibility of closely monitoring and supervising of the trials, since each bulking group has its DEA. DEAs' visits to participating farmers was very irregular. However, there was good cooperation from the farmers and a lot of interest in the programme. Tables 6 and 7 present results of gross margin analysis based on current production organisation in the milkshed area. The prices used are those that are currently in force. Revenue on milk sales assume that all milk was sold to Malawi Dairy Industries (MDI) which infect is not the case.
Table 6: Gross margin for the dairy with animals fed maize bran.
|
Animal numbers and performance | |
|
Average number of cows in milk |
3 |
|
Average milk yield per cow/day (litres) |
9.7 |
|
Average lactation length (months) |
4 |
|
Revenue | |
|
Milk (3,533 litres) at Malawi Kwacha (MK) per litre 0.45 |
1589.85 |
|
Variable costs | |
|
Concentrates: 30 bags of maize bran at MK 4.70 per bag (MK) |
141.00 |
|
Transport costs of feed at MK 0.50/bag |
15.00 |
|
Labour costs (family) |
2.50 |
|
Total variable costs (MK) | |
|
Gross margin (revenue-costs) MK |
1431.35 |
|
Gross margin per cow (MK) |
477.11 |
Table 7: Gross margin for the dairy farmer with animals fed a mixture of maize bran and cottonseed cake.
|
Animal and performance | |
|
Average number of cows in milk |
3 |
|
Average milk yield/cow/day (litres) |
10.7 |
|
Average lactation length (months) |
1 |
|
Revenue | |
|
Milk (3,913.8 litres) at MK0.45 per litre (MK) |
1761.21 |
|
Variable costs Concentrates | |
|
(a) 8 bags of cottonseed cake at MK 8.50/bag (MK) |
58.00 |
|
(b) 30 bags of maize bran at MK 4.70/bag (MK) |
141.00 |
|
Transport costs of feed at MK 0.50/bag (total of 30 bags) |
19.00 |
|
Labour costs (family)* |
12.50 |
|
Labour variable costs (MK) |
240.50 |
|
Gross margin (revenue-costs) (MK) |
1520.71 |
|
Gross margin/cow (MK) |
506.90 |
* includes costs of milling and mixing the cottonseed cake.
The gross margins/cow in Tables 6 and 7 show that the farmer could get an additional MK29.79 by including cottonseed cake in his maize bran-based dairy ration. Other sets of data are being collected for the other trials. It is still early to make any deductions based on these data.
Crude protein contents of the feed and forage resources utilised by the animals in the study
Samples of maize bran, a mixture of maize bran and cottonseed cake, and improved pastures collected were analyzed for crude-protein content. Samples of natural pastures were taken but the chemical analyses are not completed. Crude protein values of the feeds issued to dairy farmers and the forage resources available on their farms are presented in Table 8. Maize bran contains less crude protein than maize bran mixed with cottonseed cake. Crude protein values for forage resources did not vary much among the different pastures. Values for Rhodes grass were less than crude protein values of other grasses (Hamil panic, Ntchisi panic, Napier and Guatemala in Table 8). In general, the crude protein values of legume forages were much higher than the protein values of grass forages (Table 8).
Table 8: A summary of crude protein contents of maize bran mixtures of maize bran and cottonseed cake, and forage samples collected from dairy farmers' fields (% of DM).
|
Sample |
Number of samples |
Range |
Mean |
|
Maize bran |
8 |
11.5 - 15.3 |
12.9 |
|
Maize bran + cottonseed cake |
6 |
18.7 - 25.8 |
20.1 |
|
Hamil panic grass |
2 |
12.5 - 15.5 |
14.0 |
|
Napier grass |
6 |
12.5 - 14.5 |
13.4 |
|
Rhodes grass |
3 |
10.0 - 15.3 |
11.8 |
|
Guetamala grass |
1 |
- |
12.5 |
|
Silverleaf desmodium |
1 |
- |
18.8 |
|
Common centrosema |
1 |
- |
15.6 |
On-station animal feed resource technologies have been generated in Mzuzu. But a very small proportion of these technologies is utilised by the small-scale farmers. This is shown by the tow levels of milk yields in the area because natural pastures are the feed resources that are commonly used. Lack of researchers extension workers and farmers interaction has been common so that the technologies developed have not found their way to the farmer. With the formation of bulking groups (dairy farmers associations) things are likely to improve. Work is done on a group basis and inputs required for utilisation of results are not a problem since the associations have funds that cater for purchases of feeds and other inputs involved in the dairy production system. Thus, these associations are acting as instruments for promoting the adoption of technologies. Farmers will be able to buy cottonseed cake and with proper feeding management and the utilisation of improved forage resources milk production levels will be increased.
The inclusion of legumes into the farming systems and the encouragement of farmers to practice zero-grazing system, feed CP levels will be increased, leading to the efficient utilisation of the feeds, and soil fertility maintenance will be enhanced.
If the testing of the on-station generated technologies will be successful on-farm level, farmers will see increased levels of milk production, reduced calving intervals and reduced animal mortality rates. This will speed up the rate of dairy development because there will be more replacement animals. At present there is a critical shortage of dairy animals in Mzuzu milkshed area.
Although there are some problems being experienced in carrying out this programme, there are prospects for its success. The successful adoption of the technologies will teed to use of adapted high yielding forages. This will lead to more feed resources of high quality, better utilisation of feed resources resulting into:
- improvement of livestock
- reduced calving intervals
- increased milk production
- increased rate of replacement heifers
- increased reproductive efficiency of animals
- enhanced soil fertility maintenance with the final attainment of food security and improved farm income
Anonymous, 1982. Pasture Research in Mzuzu Agricultural Development Division - 1982/83. Department of Agricultural Research, Ministry of Agriculture.
Anonymous, 1985. Kenya Agricultural Research Institute Technical Note No. 1. On Animal Nutrition for the Animal Production Research Department.
Anonymous, 1987. Project Outline for a Dairy Nutrition Study in Mzuzu Milkshed Area - 1987/88. Department of Agricultural Research, Malawi.
Anonymous, 1988. Malawi National Census Report. National Statistical Office, Government Printer, Zomba.
Cusack, T.J. 1988. Report of the a Reconnaissance Survey in Mzuzu Milkshed Area. Unpublished.
Dzowela, B.H. 1987. Maize stover improvement with legume forages: In: (ed). Kategile, J.A., A.N. Said and B.H. Dzowela. Animal Feed Resources for Small-scale Livestock Producers. Proceedings of the Second PANESA workshop, held in Nairobi, Kenya, 11-15 November, 1985.
Katuma, K., and Stoz, D. 1987. Malawi German Livestock Development Programme Working paper No. 5. Ministry of Agriculture, Malawi.
Kumwenda, M.S.L. and Munthali, J.T. 1988. Utilisation of Maize Bran and Cottonseed Cake by lactating dairy cows in Malawi. African Research Network for Agricultural By-products (ARNAB) newsletter (in press).
Ministry of Agriculture. 1983. Pasture Handbook for Malawi, Lilongwe, Malawi.
Topps J.H. 1961. Animal Feeds of Central Africa. Occasional paper No.1. Department of Agriculture, University of Rhodesia, Salisbury.
