3.1 Research background
3.2 The research approach
3.3 Results of systems research at Debre Zeit
3.4 Results of component research at Debre Zeit and Debre Berman
As indicated in Chapter 1, agricultural research and extension in the Ethiopian highlands has concentrated in the past on the crop component, and particularly on improvements in crop husbandry and marketing. Although livestock production is important in Ethiopian agricultural systems, comparatively little research has been done on improving overall farming systems via the development of animal production at the farm level1.
1 ILCA has conducted some surveys on a number of small farms that had been previously supplied with crossbred cows by EPID and DDA. The surveys showed that farmers encountered substantial problems in ensuring adequate fodder supplies, that veterinary services were inadequate, and that in many cases the overall management of the anneals was poor, resulting in high calf mortalities and extended calving intervals.
In 1977 the ILCA Programme Committee authorized the initiation of an ILCA Highlands Research Programme with the objective of "bringing together available technologies in a range of farming systems that will increase animal productivity and profitability in the highlands" (ILCA, 1978). The research programme was to be guided by the following assumptions (adapted from ILCA, 1978):
- smallholder farming for the production of subsistence crops will be the prevailing system for the foreseeable future in the East African highlands;- there is an urgent need to maintain and build up soil fertility;
- animal draught will remain an integral part of the Ethiopian highland farming system;
- improved forage production and better genetic capabilities are the basis for improving livestock output;
- improved forage production on arable land is possible only when basic subsistence requirements are fulfilled; because of the extreme scarcity of land this generally means that yields per unit area of subsistence crops have to be increased in order to make room for the planting of forage crops;
- there is a growing demand for dairy products and increasing interest among smallholders in starting up a dairy enterprise;
- sheep are an important feature in the highlands of Ethiopia, and there is considerable scope for increased output; again, only if fodder production and genetic potential can be improved;
- at present there is a certain amount of underutilization of labour in most Ethiopian farming communities at certain times of the year.
The basic hypothesis of the research was that substantial gains in farm productivity could be achieved by applying technologies which were, for the most part, already proven. At current human population growth rates of 2.5% p.a, it is essential that such gains be realized by the farming population at large, or there will be a severe food crisis in the Ethiopian highlands by the year 2000.
In the initial years of the research, ILCA concentrated on acquiring knowledge about the role and potential of smallholder dairy development in the central highlands of Ethiopia. Experience elsewhere in the world, for example in India and Kenya, has shown that introducing smallholder dairy production based on crossbred cattle is a powerful means of raising net farm incomes and farmer welfare. In the longer term, a much expanded smallholder dairy sector in Ethiopia would also be a major employer of labour.
The results of ILCA's research since 1977 have indicated that the basic assumptions underlying the programme are still most relevant. Although the programme has not so far dealt with all the topics arising from these assumptions, substantial progress has been made in terms of both systems and component research. The following sections of this chapter present details of the research results from 1977 to 1980 at ILCA's Debre Zeit and Debre Berhan research stations. The former became operational in 1977, 2 years before the latter, and the details presented reflect the longer duration of research at Debre Zeit.
Prevailing prices and price ratios, favourable climatic and environmental conditions, the orientation of small farms towards mixed agriculture, and accessibility to urban markets all suggested a substantial potential for an expansion of smallholder dairy production in both the ILCA study areas. Debre Zeit and Debre Berhan are both included in the milk collection network of the government-owned dairy processing plant at Shola near Addis Ababa². Research results acquired at both stations would thus be directly applicable to other smallholders with good market access for their livestock and dairy products.
² An ILCA survey of the Addis Ababa dairy market indicated that the market is able to absorb a substantial increase in supply without a significant reduction in price.
However, an estimated 50 to 75% of Ethiopian farmers are more than half a day's walk from an all-weather road. These off-road producers will continue to have milk marketing problems until the national road network and general infrastructure are much expanded. On the other hand, if smallholder-oriented dairy processing technologies were to be made available to them they too would benefit from ILCA's research. During the early years of the Highlands Programme research was directed primarily towards solving the basic problems of establishing an improved dairy enterprise, but even at this early stage it was envisaged that research on processing technologies would become an integral part of the programme's future activities.
The approaches to research at Debre Zeit and Debre Berhan are presented in Sections 3.2.1 and 3.2.2.
The general scarcity of land in Ada Wereda has been described in Section 2.3. It has also been noted that, even with current stock numbers, the quality and quantity of forage available from natural pastures, fallow and crop stovers, are seasonally variable, well below the optimum for livestock production, and occasionally barely sufficient even for maintenance needs. Furthermore, indigenous breeds have a limited milk yield potential even with improved feeding, as reported by IAR (1976). Significant increases in the production of milk in smallholder dairy enterprises will only be achieved by using higher-yielding breeds or crossbreeds of cattle. The use of Friesian sires for crossbreeding, in accordance with the upgrading programme already long since undertaken by the Ethiopian Government, results in crossbred progeny that are substantially larger than local indigenous breeds. To realize economic milk production from locally available crossbred cows therefore requires a high plane of nutrition not generally available in the existing smallholder production setting.
Combined with the land scarcity problem, the need for this high nutritional plane indicated that a farming system with forages grown as an integral part of the cereal - pulse rotation would be necessary. However, Ethiopian farmers do not traditionally grow forage crops, so in this sense forage production would be a new enterprise. Furthermore, the primary goal of subsistence farmers is to satisfy family food requirements through grain crop production, implying that diversification into a new enterprise would only be acceptable if the latter were not put at risk.
ILCA's hypothesis was that subsistence farmers would grow forage crops only if the yields per ha of human food crops could be increased sufficiently to allow some land in the cereal phase of the rotation to be sown to them, and if the net income from the livestock enterprise using the forage were significantly greater than the value of the cereal crop that could have been grown instead. If either of these conditions could not be satisfied, then there would be no basis for expecting farmers to reallocate their limited land, labour and capital resources to the specialist dairy enterprise.
Income from the dairy enterprise is realized only after the conversion of the forage into livestock products via the animal. The delays inherent in this process and local unfamiliarity with the improved forage/dairy cow enterprise created production risks which would need to be overcome by the attractions of a substantially higher gross margin and a regular cash flow.
A multidisciplinary ILCA team discussed the feasibility of introducing a higher-yielding cow along with forage production into the smallholder system as a means of raising income levels in areas such as Ada Wereda. The "package" approach to testing innovations in the traditional system was selected as the most appropriate.
A package based on the best available knowledge and experience would be introduced to a number of farmers in the area. The functioning and performance of the package as adopted by the farmers, called by ILCA "participating" farmers, would be closely monitored. Parallel research on the package would be conducted by ILCA on research farms under ILCA management. Selected components of the production system would also be studied on-station. The results from the monitoring of the participating farms would be used as an input to the design and operation of both the research farms and the component research.
The package was designed taking into account the results of research by institutions such as the JAR, the Debre Zeit Agricultural Research Station of the University of Addis Ababa, EPID, DDE and CADU³.
³ The Arsi Rural Development Unit (ARDU) is now responsible for all the former activities of CADU. Chilalo Awraja is part of Arsi Province.
It comprised the following elements:
- the use of improved cereal seeds and chemical fertilizer based on IAR recommendations;- the use of recommended seeding rates and sowing dates on the different crops;
- the incorporation of forage crops as a regular part of the rotation;
- the provision of initial advisory services through a weekly visit by an extension officer;
- the introduction of an F1 crossbred (Boran x Friesian) heifer and rearing of its female calves to their first calving; male calves were to be sold or slaughtered as soon as practicable after birth4;
- the provision of initial credits according to government guidelines at an interest rate of 12% p.a., to allow farmers to establish their enterprise.
4 The Boran is an indigenous Zebu cattle breed. Only Friesian sires have been used for crossbreeding on any scale in Ethiopia. As a consequence only Friesian x Boran crossbred cows are available in sufficient numbers for research purposes. Management and health problems at the smallholder level favour the use of F1 crossbred cows rather than purebred Friesians or even cows with ¾ or more of Friesian blood.
For the most part the separate elements of this package were considered to be well understood. However, it was necessary to conduct research on the package as a whole, as well as on certain aspects of it, to assess its relevance and robustness in the wider context of the Ethiopian highlands. The original objectives of the research can be stated as follows:
- to test the effect of the package under smallholder conditions;- to quantify the inputs required for and the outputs resulting from the operation of the modified farming system;
- to identify other constraints on the improvement of livestock production and on the further integration of livestock production with the cropping system;
- to evaluate the approach and the results as regards their transferability to other sites and production situations.
Thirty-four families were originally resident on the land granted to ILCA at Debre Zeit. Some of the household heads were part-time farmers with outside employment. Eighteen of the full-time farmers subsequently became participating farmers who adopted the ILCA package. Originally it was planned to have over 30 participating farmers, but this target was not achieved because insufficient crossbred cows were available to establish the enterprises. In return for agreeing to participate in the research programme, 18 farmers received long-term credits for the purchase each of an in-calf crossbred heifer and a pair of local oxen, and the construction of a cow shed. Short-term credits were given for the annual cropping programme5. The maximum credit provided to any farmer was US$ 750. As the farmers had no prior experience with forage crops, ILCA agreed to compensate them for any losses in the first year arising directly from ILCA's advice6. This guarantee was offered because in this part of Ethiopia oats - one of the forage crops to be used - are usually regarded as a weed.
5 Long-term credits have to be repaid within 5 years, while short-term credits cover a single cropping season.6 In the first year only one farmer was compensated for losses.
The land holding of each participating farmer was fixed according to the rules of the local PA, the available land and the family size. Bachelors were allocated 1 ha of cropland, married couples 2 ha, and families with children 2.5 ha. Each farmer had separate plots on the different soil types. The typical rotation in the Debre Zeit area is a 3-year cereal - cereal - pulse rotation. The crop rotation under the improved system was planned in accordance with soil type, but the basic rotation recommended was cereal forage pulse. In addition to their individual land holdings and garden plots, the 18 participating farmers also had access to a 30-ha communal pasture area.
In order to evaluate the package, a comprehensive data collection system was set up covering all aspects of the production and household system of the participating farmers. The information collected closely paralleled the data collected from "outside" farmers (i.e. those in the traditional system), the details of which were provided in Section 2.1.
Research parallel to that undertaken with participating farmers was carried out on research farms at the station which were managed and operated by ILCA employees on a fixed salary plus an annual bonus as part of an incentive scheme. In addition to providing a direct means of assessing the technical and biological feasibility of a range of alternative enterprise combinations, the research farms served as a clearing house for more innovative enterprises which might later be introduced to the farming community. If the best available technologies are tested on research farms, these indicate the upper limits of the productivity increases feasible off the station.
A range of different systems is being assessed on the six research farms which have been in operation at Debre Zeit since 1978. Two farms with 2.5 ha of cropland are variants of the traditional farming system, testing the same package as participating farms but with a modified traction component. Three operate on a somewhat larger scale, holding 4 ha of cropland and 1.5 ha of natural grazing, and examine the potential of more intensive dairy enterprises. The last farm is a specialist goat enterprise on 1.5 ha of light soil, of which two thirds is arable.
The basis for examining this last system is that a larger number of goats than cows can be kept on a given area, while lactation yields per doe can be as high as in local cows, implying less production risk. Goat farming appears, prima facie, appropriate in those areas where demand for milk is high but land is scarce and not well suited to permanent cultivation.
The station at Debre Zeit is also used for component research. The major activities are screening and production trials on a wide range of annual and perennial forage crops; a comparative study on different, low-cost surface-water drainage methods; cultivation trials with different breeds of oxen and, finally, a small-scale project on biogas production using cattle dung.
Forage agronomy studies have accounted for the major part of the overall component research. They have been particularly important as a technical complement both to the package in use by participating farmers and to the research farms managed by ILCA. The forage research programme at Debre Zeit and Debre Berhan is described in Section 3.4.1, together with selected results.
ILCA's drainage research addresses the problems and potentials of making routine use of the fertile but seasonally waterlogged bottomlands by the construction and maintenance of low-cost surface-water drainage works. The use of animal rather than tractor power for the engineering works indicates the importance ILCA attaches to developing the renewable farm-based energy resources on which smallholder agricultural systems will depend in the long term. Key results on this topic are also reported in Section 3.4.1.
ILCA is conducting research at both Debre Zeit and Debre Berhan on the efficiency of various cultivation schemes using different breeds of oxen. Innovative research has also been initiated on the use of cows for both milk production and traction. Cows are not used for traction in the traditional agricultural system in Ethiopia, nor, with rare exceptions, anywhere else in sub-Saharan Africa. The initial results of this work at both stations are given in Section 3.4.2.
The stress ILCA places on the role of renewable energy resources in the agricultural small-holding has already been noted. ILCA's experience with methane digesters has been limited, but although modest in scale this activity has attracted substantial interest on the part of farmers in the area. Details of this aspect of ILCA's research are presented in Section 3.4.3.
Research at the ILCA Debre Berhan station has followed a more phased approach than at Debre Zeit. Little information was available on the nature of the technologies which would be applicable under the prevailing conditions of the traditional farming system. It would therefore have been premature to formulate and introduce packages with traditional farmers for testing at their own risk. Accordingly, it was proposed to evaluate a number of farming systems under ILCA's own control on the research station. Once various enterprise combinations had been appraised, innovation packages would then be designed and specific recommendations made to farmers in the area. Research on specific components of the agricultural system, such as forages and pastures, early maturing crop varieties, animal traction and low-cost surface-water drainage methods, would also be undertaken.
The Debre Berhan station is located on yemeda class land (see Section 2.3). The agricultural hazards - frosts and waterlogging - to which such land is subject added significantly to the research task. However, if the research was successful it would have wide applicability over large parts of the highlands where cultivation is limited by these constraints. In particular it would allow the arable area of the highlands to be expanded significantly, providing for greater national food security.
During the first 2 years of operations at Debre Berhan the emphasis was on developing the station infrastructure. Research was conducted both on and off the station as the site was being built. As a follow-up to the baseline survey conducted in the first year, data collections continued in the traditional agricultural system, to provide information on particular production constraints and to allow year-to-year dynamics in the system to be quantified. Together these data would allow formulation of improvement packages that could be tested with farmers in the area. The routine data collection system has been discussed in Section 2.1.
The on-station systems research was to focus on the evaluation of different farming systems with an improved livestock component. Studies on the station with Friesian x Boran crossbred cows complemented the work at Debre Zeit. Sheep were also to be included in the research. The systems were to be evaluated by using ILCA-managed research farms, as at Debre Zeit. Initially 12 farms, each of 5 ha, were established with an additional 4 farms being added in 1980. Eight different combinations of livestock with forage or subsistence cropping enterprises were tried, each replicated twice.
Crops were extensively damaged by frost in both 1979 and 1980. Early growth of crops was very favourable, but the frosts occurred at the maturing stage and seed yields were severely reduced. Barley yields on the research farms in 1979 averaged 1010 kg per ha. Average wheat yields were 488 and 810 kg per ha in 1979 and 1980 respectively. Forage crop yields based on a combination of oats and vetch ranged between 2 t and 4 t of DM per ha. Cows on the research farms calved first at the end of 1979. Average annual milk yield, adjusted for calving interval, was over 1700 kg ± 510 in the first lactation, with a range of 900 to 3000 kg.
In early 1981, the eight different systems were consolidated into four replicates of four innovative enterprise combinations considered applicable in the larger agricultural zone of which the Debre Berhan station is a part. The main distinguishing features of the new systems are summarized below.
The first system is an intensive dairying option in which cows are used for both traction and milk production. Four Friesian x Boran crossbred cows are kept on each replicate. The system is suitable for farms where milk marketing is not a problem, such as those near towns and along major roads where collection is feasible.
The second system is a combination relevant to small farms away from major roads with no ready access to crop inputs or market outlets. It is a farm heavily dependent on sheep production. Besides approximately 30 ewes plus followers, one crossbred cow plus female followers is kept. The animal draught power is provided from off the farm. According to the design a Jersey x Boran cow is required, but to date these have not been available. A Friesian x Boran cow is used until the former become available7.
7 Although a Jersey type cow yields less than a Friesian type its milk has a higher butterfat content. Furthermore, its total feed requirements for maintenance and production are lower than those of the larger Friesian type cows.
The third system is suitable for both on-road and off-road smallholders, and is a mixture of the above systems. It has one Friesian x Boran crossbred cow plus followers and one Friesian x Boran crossbred ox. Together these two adult animals provide the traction power for farm cultivation needs8. Approximately 15 sheep are also being kept during the first year of implementation, with more or less to be kept in later years according to the capacity of the farm to carry them.
8 Preliminary work at ILCA has shown that the crossbred ox is capable of cultivating as a single animal, so that pairing for cultivation is not an absolute necessity. This line of research is being pursued further.
The fourth system is expected to produce the highest human food surplus of all four systems under test. It is a combination of forage crops for use by crossbred cows and potato growing for human consumption. It is well suited to producers with good road and market access so that farm products can be sold readily. Each replicate of this set of enterprises has three Friesian x Boran crossbred cows providing both milk and traction power. No oxen are used. Approximately 1 ha of potatoes are grown each year. Indications are that potato yields of over 15 t per ha are feasible under smallholder management.
All 16 farms grow 1.5 ha of subsistence crops as well as 1 ha of annual forage crops out of their total allocation of 5 ha. Feed concentrates are used only when difficulties are encountered in animal feed supplies during the establishment phase. Modifications to these systems will be made in response to results obtained and as further possibilities for improvement are identified.
3.3.1 Participating farmers
3.3.2 Research farms
3.3.3 Cooperative farms
This section summarizes results and experiences from 1977 to 1980 with the package of innovations adopted by the participating farmers at Debre Zeit. Generally, this part of the programme has been successful in showing that the proposed improved system is feasible at farm level, and that crop and livestock enterprises can be profitably integrated, giving rapid and substantial increases in farmers' incomes.
Throughout this section the term "outside" farmers is used to refer to those farmers outside the research station who have been studied in ILCA's surveys of the traditional agricultural system (see Chapter 2). These outside farmers have a similar resource base to that of participating farmers, with about 2.5 ha of cropland each, and at the start of the programme also had similar inventories and asset positions.
An underlying assumption of the ILCA package was that yields per ha of traditional crops could be increased significantly through the use of improved seeds and cultivation practices, and the strategic application of chemical fertilizer9. Table 3.1 gives the average yields of participating and outside farmers for subsistence crops traditionally grown in Ada Wereda. Yields on the plots of participating farmers were 45% and 43% higher than on outside farmers' plots for teff and wheat respectively. Pulse yields, on the other hand, were comparable: yield improvements in pulse crops were not included in the ILCA package.
9 Fifty kg per ha of DAP on teff and bread wheat.
The higher yields shown for participating farmers in Table 3.1 are due to the fertilizer effect and to improved crop husbandry, including better seedbed preparation and weed control. Participating farmers had a higher proportion of seasonally waterlogged vertisols, in which teff grows particularly well, and this may also have contributed to their higher average yields. Wheat yields were below expectations and, with the exception of 1978, have been disappointing. Average yields were highest in 1978 due to favourable climatic conditions: total rainfall in 1978 was 980 mm, compared with 673 mm in 1979 and 766 mm in 1980. In 1979, approximately half the participating farmers did not use fertilizers, but their yields were still substantially higher than those achieved by outside farmers, who also did not use fertilizers. These higher yields reflected previous fertilizer use and more careful weeding of crops.
Although the package showed that average cereal yields could be increased significantly, farmers did not reduce the areas sown to cereals as anticipated. Prices of cereals more than doubled over a 2-year period, and farmers responded to this price incentive by allocating larger areas to teff. Among participating farmers, the average proportion of land under teff went up from 32% in 1978 to 49% in 1979. Among outside farmers, it increased from 49% in 1978 to 56% in 1979. Teff prices stabilized in 1980, and the proportion of land under teff in that year was 48% for participating farmers and 50% for outside farmers. This finding indicates that Ethiopian smallholders are price-responsive and willing to modify their production mix to increase total farm gross margins. Average cropping patterns from 1977 to 1980 of participating and outside farmers are summarized in Table 3.2.
Table 3.1. Average crop yields for participating and outside farmers at Debre Zeit, 1977-1980 (kg/ha).
|
Crop |
Participating farmersa |
Outside farmers |
||||||||
|
1977 |
1978 |
1979 |
1980 |
Average |
1977 |
1978 |
1979 |
1980 |
Average |
|
|
Teffb |
1 154 |
1 524 |
1 289 |
1 191 |
1 290 |
722 |
984 |
932 |
879 |
892 |
|
Wheat |
627 |
1 209 |
766 |
1 011 |
903 |
470 |
850 |
407 |
804 |
633 |
|
Horse beans |
970 |
874 |
938 |
817 |
900 |
790 |
878 |
769 |
597 |
758 |
|
Chick peas |
650 |
304 |
182 |
n.ap. |
379 |
620 |
585 |
420 |
317 |
486 |
a Using 50 kg per ha of DAP on teff and wheat.
b Yield differences of teff crops between participating and outside farmers for corresponding years were significant at the 1% level.
Table 3.2. Average cropping patterns of participating and outside farmers at Debre Zeit, 1977-1980 (percentage of total area under crops).
|
Year and farmer type |
Cereals |
Pulses |
Forage |
Fallow |
|
|
1977
|
Participating |
40 |
30 |
30 |
- |
|
Outsidea |
n.av |
n.av. |
n.av. |
n.av. |
|
|
1978
|
Participating |
43 |
30 |
27 |
- |
|
Outside |
60 |
37 |
- |
3 |
|
|
1979
|
Participating |
60 |
10 |
30 |
- |
|
Outside |
68 |
31 |
- |
1 |
|
|
1980
|
Participating |
53 |
9 |
37 |
1 |
|
Outside |
61 |
36 |
- |
3 |
|
a The ILCA survey of outside farmers began in the 1978 crop year.
Average gross margins per ha (value of production minus cash costs) are summarized in Table 3.3. Teff is a profitable crop principally because of its grain price, but the amount of teff that can be cultivated on a farm is limited by seasonal labour constraints, the timely availability of draught oxen power, and rotation restrictions. As indicated in Table 2.8, teff requires more labour per ha than any other crop. Pulses are much less profitable than teff because of their low yields, low output prices and high seeding rates. They are rich in protein, however, which makes them an essential part of the diet, especially during the long fasting periods. In the long run pulses must be included in the crop rotation to maintain soil fertility, especially as inorganic fertilizers are expensive for farmers to use.
Participating farmers had consistently higher gross margins per ha than outside farmers. Taking the average gross margin per ha for all food crops (as in Table 3.3), the increase was 31% in 1977, 27% in 1978, 43% in 1979 and 22% in 1980. Over the 4-year period, participating farmers had an average gross margin 31% higher than outside farmers. The higher gross margins for cereal food crops are due to the effects of fertilizer, the use of genetically superior seed, and higher labour inputs for land preparation and weeding. Much of this extra labour was hired. Participating farmers also followed ILCA's advice on early seedbed preparation, and used recommended sowing dates.
In 1977, fodder oats and maize were grown, but harvests were poor because of late sowing, a low seeding rate and the reluctance of farmers to accept and manage the forage crops as a necessary part of a farming system to support a high-yielding cow. Maize produced only 2 t of DM per ha in that year. The maize and some forage oats were ensiled, but labour for this operation was required while teff was being weeded, and a substantial capital investment was also necessary for transport and machinery. Unless silage making technologies suited to the ressources of the Ethiopian smallholder become available, the prospects for using this alternative for producing high-quality forage appear bleak.
Only pure stands of oats were planted by participating farmers in 1978. They were used as green feed in cut-and-carry systems, and for hay making. Oats were grown on three plots on each farm, one each for green feed, hay making and seed production. Hay yields averaged 5.9 t of DM per ha with a 10% CP content. Average seed grain yields were 550 kg per ha.
Table 3.3. Average gross margins on food crops for participating and outside farmers at Debre Zeit, 1977-1980 (US$/ha).
|
Crop
|
Participating farmers |
Outside farmers |
||||||||
|
1977 |
1978 |
1979 |
1980 |
Average |
1977 |
1978 |
1979 |
1980 |
Average |
|
|
Teff |
435 |
657 |
511 |
346 |
487 |
307 |
443 |
363 |
362 |
369 |
|
Wheat |
66 |
549 |
189 |
246 |
263 |
75 |
379 |
113 |
299 |
217 |
|
Horse beans |
75 |
146 |
181 |
116 |
130 |
44 |
107 |
106 |
84 |
85 |
|
Chick peas |
91 |
61 |
23 |
n.ap. |
58 |
80 |
144 |
81 |
75 |
95 |
|
Weighted average |
216 |
404 |
371 |
305 |
324 |
165 |
317 |
259 |
250 |
248 |
Beginning in 1979 a mixture of oats and vetch (Vicia dasycarpa) was recommended to participating farmers, on account of its favourable yield and its high protein content10. Some practical problems arose because farmers had limited experience with forage crops and particularly with crop mixtures. Sample yields in 1980 were promising, but late rains delayed the harvest of the hay crop and farmers left their forages in the fields until after the food crops had been harvested. As a result field losses were substantial, as much of the crop was eaten by grazing animals. Usually, the harvesting of forage crops does not compete with teff harvesting, hay making taking place a few weeks earlier. Farmers in 1979 and 1980 harvested on average only 3 t of DM per ha from their forages.
10 Difficulties with the supply of vetch seed have been encountered since the start of the programme. To alleviate the problem ILCA contracted with a government farm to grow vetch seed beginning in 1980.
When forage crops fail, the farmers must supplement their animals with purchased concentrates, usually middlings, wheat bran and oilseed cake, to maintain milk yields. These concentrates are not expensive, at US$ 0.08 per kg, but the supply is irregular and transporting them from the town to the farms has proved to be a practical problem.
In December 1977, each farmer purchased one in-calf Friesian x Boran crossbred heifer for US$ 190. The animals were bred on the government ranch at Abernossa. They calved between 1 and 3 months after distribution. Purchase was on standard Ethiopian Government credit terms of 12% with a 5-year repayment period, including insurance for 1 year which guaranteed replacement of the animal in the event of death or a milk production of less than 800 kg in the first lactation. Loan repayment was through a levy of US$ 0.05 per litre on milk sold. The price paid to farmers for milk sold to the DDE is US$ 0.22 per litre.
For breeding purposes ILCA relied initially on the AI service of the government, which proved unsatisfactory. Delays in servicing cows resulted in extended calving intervals and dry periods between first and second carvings. The average calving interval between first and second calving was around 18 months. In May 1978 a purebred Friesian bull was purchased for the station, and since then all matings have been by natural service.
The overall mean birth weight of the crossbred (¾ Friesian) calves was 29.7 kg. Male calves' weight was on average 30.3 kg at birth, 8.6% higher than the birth weight of female calves. However, the growth rates of the females were faster, and average weight after 3 months was 56 kg for male calves as against 63.7 kg for females. After 8 months, male calves weighed 101 kg, compared with 112 kg for females. The farmers fed and managed female animals better than males. Rearing male calves is unprofitable, they are often sold at an early age.
For the first 60 days after birth an average of 180 kg of whole milk was fed to all calves. This increased to 246.8 kg over the first 90 days after birth. The average milk intake of calves over 120 days amounted to 273 kg per head.
Adjusted annual milk yields from the crossbred cows averaged 1769 litres per cow during the first lactation, ranging from 1197 to 2809 litres with a standard deviation of 412 litres11. Milk production increased substantially in the second lactation, with an adjusted annual yield of 2347 litres (range 1141 to 3929 litres, standard deviation of 581 litres). The interval between second and third carvings averaged 12 months. These cows had liveweights between 400 and 500 kg.
The average length of the first lactation was 439 days, falling to 304 days for the second lactation. Table 3.4 summarizes the production and disposal of milk from the cows of participating farmers. Cows first calved in 1978 and the resulting lactation extended into 1979. Second lactations began in 1979 and extended into 1980.
As stressed previously, acceptability of the composite forage/dairy cow enterprise will depend on its being substantially more profitable than the grain cropping alternatives. Table 3.5 presents the costs and returns of an improved cow enterprise at Debre Zeit. The values tabulated are based directly on the results of ILCA's own data collection, but are modified to give the budget expected if the enterprises mature and steady-state production is achieved. This modification was necessary to overcome the procedural and interpretational difficulties of determining profitability in developing enterprises. Full details of the assumptions made in preparing Table 3.5 are given as notes to the table. The impact on profitability of alternative production assumptions can be readily determined. The budget is presented on a per cow basis, and assumes that all milk sold is in liquid form.
Table 3.4. Average annual milk yields and uses for participating farmers at Debre Zeit, 1978-1980 (litres/farm)a.
|
Parameter |
1978 |
1979 |
1980 |
Average 1978-1980 |
|
Total production |
2 185 |
1 805 |
2 139 |
2 043 |
|
Sales |
1 821 |
1 375 |
1 650 |
1 615 |
|
Home consumption |
133 |
172 |
322 |
209 |
|
Calf consumption |
231 |
258 |
167 |
219 |
a Each farmer has one cow; values tabulated are the yearly averages for the 18 participating farmers.
The budget indicates an expected return, from an enterprise consisting of a single improved cow plus followers utilizing 1 ha of forage land, of US$ 437 per year. However, the enterprise must be evaluated in the context of the whole farm, because arable land is used to grow forage. Table 3.6 presents average returns per ha for food and forage crops for the period 1978 to 1980 for the participating farmers at Debre Zeit. In this way the profitability of the dairy enterprise can be directly linked to the land resource. Table 3.6 is based on the actual costs and returns, with the values adjusted to a per ha basis, while average gross margins actually recorded per farm from the two enterprises are given in Table 3.7.
Table 3.6 indicates that the average gross margins for all food crops combined are slightly below those for forage crops. Gross margins for teff (see Table 3.3) are higher than for forage, but rotational restrictions limit continuous teff cropping as a means of increasing incomes. These results satisfy the precondition for adoption of improved dairy production in the area - namely that the per ha profitability must be substantially higher than the traditional arable land alternatives available to the smallholder. This conclusion is reached despite the fact that the market price for teff has increased markedly over the study period, while prices paid for milk have not increased. Any increase in milk prices will directly favour the relative and absolute profitability of the specialist dairy enterprise. Profitability of the dairy enterprise would also increase if the milk were converted into products such as butter or cheese, using surplus family labour and traditional processing technologies.
Table 3.5. Expected annual costs and returns of a mature crossbred dairy enterprise on 1 ha in the Debre Zeit area (US$/cow)a.
|
Costsb |
Returnsc | ||
|
Item |
Amount |
Item |
Amount |
|
1. Forage |
120 |
1. Milk production |
471 |
|
2. Feed concentrates |
50 |
2. Dung cakes |
60 |
|
3. Veterinary/health care |
10 |
3. Male calves |
6 |
|
4. Housing and tools |
20 |
4. Surplus females |
85 |
|
5. Breeding |
10 |
5. Residual value of cow |
50 |
|
6. Hired labour |
25 |
Total returns |
672 |
|
Total costs |
235 |
|
|
a The enterprise consists of one cow plus followers to time of sale.b Assumptions as follows:
1. Cash costs for the forage are seed (100 kg of oats at US$ 15/100 kg; 20 kg of vetch at US$ 250/100 kg) and fertilizer (100 kg DAP at US$ 55/100 kg). Ploughing is assumed to be done by the farmers' own oxen at no cash cost. 2. 1000 kg of feed concentrates at US$ 0.05/kg. 3. Indicative cost for veterinary/health care. 4. Housing is fully depreciated over 5 years with no residual value. 5. Breeding cost per effective insemination based on estimated AI cost in Addis Ababa. 6. A total of 25 worker-days of hired labour used per year at US$ 1 per day.c Assumptions as follows:
1. Milk production is 2500 litres/lactation or 2143 litres/year, assuming an average calving interval of 14 months. This yield is the amount available for sale and home consumption after allowing for milk consumed by the calf. Milk is valued at the official purchase price of US$ 0.22/litre. 2. Dung cakes are valued at US$ 0.05/kg. Annual production of dung cakes from a cow plus followers is 1200 kg. 3. Male calves are sold at US$ 15 per head and do not consume milk. Mortality rate to time of sale is 10%. 4. Cows rear their own replacements. Each cow produces an average of five calves over a 6-year productive life. One of the female calves is kept as a replacement, allowing for the sale of 1.5 heifers in 6 years, less allowance of 15% for mortality to time of sale. Surplus heifers are sold at 15 months of age at US$ 400 per head. 5. No data available; rough estimate only.
Table 3.6. Average gross margins from food and forage crops for participating farmers at Debre Zeit, 1978-1980 (US$/ha).
|
Crop group
|
Year |
Average |
||
|
1978 |
1979 |
1980 |
1978-1980 |
|
|
Food cropsa |
404 |
371 |
305 |
360 |
|
Forage cropsb |
496 |
411 |
319 |
409 |
a Cereals plus pulses.
b Assuming all returns on the dairy enterprise to be directly attributable to the forage crop.
As already indicated, participating and outside farmers have similar resource bases. Nonetheless, individual holdings of participating farmers, at 2.5 ha, were somewhat smaller than the average holding of an outside farmer, which was almost 2.8 ha in 1978 but declined to an average of 2.5 ha in 1980. This decline is due to increasing population pressures and resettlement programmes. Both participating and outside farmer families have an average of 5.5 members, consisting of two adults, and two or three children under 15 years old. The families of outside farmers were thus working slightly larger areas than those of participating farmers, a factor which partially explains their higher labour inputs. Overall labour inputs (family, hired and exchange) for crop-related activities by both groups of farmers at Debre Zeit were as shown in Table 3.8.
Table 3.7. Average gross margins per farm of participating farmers at Debre Zeit, 1978-1980 (US$).
|
Enterprise
|
Year |
||
|
1978 |
1979 |
1980 |
|
|
Dairying (including forage)a |
372 |
308 |
293 |
|
Food crops |
721 |
631 |
454 |
|
Total |
1 093 |
939 |
747 |
|
Average farm gross margin/hab |
434 |
403 |
299 |
a Including interest charges, but excluding cow replacement charge.
b The area cultivated varied slightly from year to year.
Table 3.8. Average labour inputs per year for crop-related activities by participating and outside farmers at Debre Zeit, 1978-1980 (hours/farm)a.
|
Farmer type
|
Year |
||
|
1978 |
1979 |
1980 |
|
|
Participatingb |
1 377 |
1 193 |
1 082 |
|
Outside |
1 513 |
1 124 |
1 183 |
a Includes land cultivation, weeding, harvesting, transport, threshing and storing of the produce.
b Includes forage crops.
The decline in labour inputs reflects the lower crop yields of 1979 and 1980 and, for outside farmers, the smaller land area cultivated. Most of these labour inputs are provided by the family, although participating farmers used substantially more hired labour than outside farmers, as illustrated in Table 3.9.
Table 3.9. Average annual cash expenditure on hired labour by participating and outside farmers at Debre Zeit, 1978-1980 (US$/farm)a.
|
Farmer type
|
Year |
||
|
1978 |
1979 |
1980 |
|
|
Participating |
30 |
48 |
54 |
|
Outside |
25 |
11 |
13 |
a Average wage rates are around US$ 1 per day. Expenditure occasionally also includes rent of one or two oxen.
The inputs of family, exchange and hired labour in 1980 were as summarized in Table 3.10. The table indicates that participating farmers substituted a substantial part of their exchange labour for hired labour, and that their own family labour input for field activities was somewhat reduced, allowing labour resources to be allocated to the dairy enterprise.
The improved dairy enterprise adds around 1 hour daily to labour requirements, for milking, calf care, barn cleaning, marketing of the milk and extra hand feeding of the cow and her followers. Herding is done jointly with other traditional livestock, and is therefore not taken into account. Although herding activities take an average of 8 hours per day, they are performed by the younger children who would otherwise have little input into farm operations. Thus the opportunity cost of the labour input for herding is close to zero. If the additional 1 hour per day or 365 hours per year is added to the existing labour inputs of participating farmers (Table 3.8), then their overall labour input into the combined crop and dairy enterprises was 1742 hours in 1978, 1558 in 1979 and 1447 in 1980.
Table 3.10. Distribution of family, exchange and hired labour inputs by participating and outside farmers at Debre Zeit, 1980 (hours/farm)a.
|
Farmer type |
Family labour |
Exchange labour |
Hired labour |
|
Participating |
653 |
119 |
310 |
|
Outside |
763 |
330 |
90 |
a Assuming no quality differences between labour inputs of different age and sex groups.
A comparison of average gross margins per ha of participating farmers with those realized by outside farmers is given in Table 3.11, which shows an increase in the average value of production per ha attributable to the innovation package of 37% in 1978, 56% in 1979 and 20% in 1980. Statistical tests showed this difference to be significant at the 5% level. This comparison only takes into account returns from cropland, whether it is used for food crops or forage crops. It excludes the returns from traditional livestock production, which are approximately the same for both participating and outside farmers, estimated at around US$ 50 per annum.
Table 3.11. Average overall gross farm margins of participating and outside farmers at Debre Zeit, 1978-1980 (US$/ha).
|
Farmer type
|
Year |
||
|
1978 |
1979 |
1980 |
|
|
Participating |
434 |
403 |
299 |
|
Outside |
317 |
259 |
250 |
Average returns to labour are summarized in Table 3.12. Values tabulated are calculated by dividing overall farm gross margin by the sum of family plus exchange labour. Participating farmers had a consistently higher return to labour than outside farmers.
Table 3.12. Average annual returns to labour for participating and outside farmers at Debre Zeit, 1978-1980 (US cents/hour)a.
|
Farmer type
|
Year |
||
|
1978 |
1979 |
1980 |
|
|
Participating |
71 |
77 |
66 |
|
Outside |
66 |
62 |
57 |
a Calculated as the overall gross farm margin divided by the sum of family and exchange labour.
Including specialist dairy enterprises in farm operations thus led to a rapid increase in cash incomes. As shown in Table 3.13, the gross cash incomes of participating farmers were over two and a half times those of outside farmers in 1979. This ratio rose to nearly 4 to 1 in 1980.
Table 3.13. Average gross cash incomes of participating and outside farmers at Debre Zeit, 1979-1980 (US$/farm).
|
Farmer type
|
Year |
|
|
1979 |
1980 |
|
|
Participating |
575 |
830 |
|
Outside |
212 |
218 |
This extra income enters the household as a regular cash flow, as the DDE pays for milk deliveries bimonthly. A part of this higher cash income of participating farmers is used to repay the loans for establishment of the dairy enterprise, and also to pay for the extra inputs (seeds, fertilizer, hired labour) for the crop enterprise.
Initially, participating farmers used the higher income to improve their own family welfare. A substantial amount was spent on better clothes for the family, children-were kept at school to complete more advanced grades than usual, and most participating farmers constructed a new and better house, or started buying the necessary construction materials to do so.
It is anticipated that once basic social needs have been met a substantial part of the extra cash generated through the dairy enterprise will flow back into the farming system to increase the productivity of the crop component still further. This trend will illustrate not only the close interaction of crop and livestock production in mixed agricultural systems, but also the advantages of using the livestock component as a mechanism for benefitting the farming system as a whole.
The initial results of the ILCA package at Debre Zeit confirm that a farmer can satisfy his subsistence cereal requirements from a substantially reduced area: average teff yields per ha of participating farmers were 45% higher than those achieved by outside farmers. Although it was initially assumed that the land made available by growing subsistence grain crops on a smaller area would be used for forage production, in order to grow their fodder crops the farmers reduced instead the area sown to pulses. In this regard farmers were more responsive to prices than initially expected.
Average annual milk yields more than 10 times those from local breeds are obtainable from improved crossbred dairy cows under smallholder conditions. Dairy production with improved cows is a profitable enterprise even at the milk price now paid to farmers. It is a most efficient means for generating rapid productivity increases in the smallholder farming system, and leads in the short run to a substantial increase of both the gross cash and the net farm income.
Rearing and selling female heifers is also a profitable aspect of the improved dairy enterprise as now operated. So far however, only a few of the participating farmers have reached the stage of selling offspring.
Finally, the increased home consumption of dairy products is making a valuable contribution to overall nutritional levels in participating farmers' families.
The ILCA research experience at Debre Zeit differed from that of other institutions in that the expected beneficiaries of the results, the traditional farmers, were represented from the outset in the research process. This approach gave rise to a greater degree of insight into the problems and opportunities associated with the introduction of a markedly different kind of livestock enterprise into a traditional agricultural system.
The impact of the innovations on the social and economic circumstances of the participating farmers cannot yet be fully assessed. Nevertheless, there are strong indications that the basic package would be adopted and maintained by a wider population of farmers if it were made available to them.
At present supplies of suitable cattle to provide the foundation stock for the enterprise are generally limited. The supply of forage seeds is also a constraint on the wider application of the package. For both these reasons, the ILCA research experience with participating farmers can be considered only as a preliminary step towards development. The research results provide the Ethiopian Government with sound information to guide the future development of smallholder dairy production in the central highlands, assuming that the basis for planning such development is that solutions are first found to these two problems.
Of the two research farms at Debre Zeit which are variants of the existing system, one is on light soil and the other on heavy vertisols. In addition to the 2.5 ha of arable land available for food and forage crop cultivation, each farm has 1.5 ha of non-arable land for grazing. Both have one Friesian x Boran crossbred dairy cow rearing its own replacements. Some improved crop varieties with different fertilizer treatments are also being evaluated on each farm.
The adjusted annual milk yields over the first two lactations were 1591 and 1998 litres on the two farms. Both farms have 0.9 ha of forage, of which approximately half is used for green feeding and half for hay making. A mixture of Sorghum almum and Dolichos lablab is grown for green feeding, while hay is made from a mixture of oats and vetch. DM yields over the 3-year period 1978 to 1980 averaged approximately 6 t per ha. Since 1980 the farm on light soil has been allocated two additional crossbred cows for ploughing, thereby taking advantage of the ability of cows to cultivate light soils. The yields of crops grown on these two research farms are summarized in Table 3.14.
The third research farm produces teff and forages and uses three crossbred cows for both dairy production and ploughing. It has 5.5 ha of heavy soils, of which 4 ha is arable land sown to teff and forage crops and 1.5 ha is non-arable swampy bottomland used for grazing and hay making. With fertilizer applications of 100 kg of DAP and 100 kg of urea per year, teff yields for 1979-1980 averaged 1665 kg per ha. Adjusted annual milk yields for the cows on this farm averaged 1388 kg for the first lactation and 1629 kg for the second. A mixture of Rhodes grass (Chloris gayana) and lucerne (Medicago sativa) was sown for cut-and-carry forage production. This mixture has produced an annual average of over 3 t of DM per ha over the 3 years it has been established.
The fourth and fifth research farms are specialist dairy operations on 5.5 ha each of heavy vertisols. As for the third research farm, 4 ha is arable land and 1.5 ha is bottomland used for grazing and hay making. Both farms produce perennial and annual forages and have three crossbred cows. One farm rears female calves as cow replacements; the other does not. The cows on these farms had adjusted annual milk yields averaging 1652 kg in the first lactation and 1889 kg in the second. Both farms initially used perennial forages such as Rhodes grass/lucerne mixtures and Napier grass (Pennisetum purpureum)/Desmodium uncinatum for cut-and-carry feeding. In 1980, a mixture of oats and vetch was added for green feeding and hay making.
Table 3.14. Crop yields on two research farms at Debre Zeit, 1978-1980 (kg/ha)a.
|
Crop
|
Light soil farm |
Heavy soil farm |
|||||
|
1978 |
1979 |
1980 |
1978 |
1979 |
1980 |
||
|
Cerealsb |
|||||||
|
|
White teff |
1 550 |
1 467 |
1 569 |
1 555 |
1 324 |
1 469 |
|
|
Bread wheat |
1 739 |
1 672 |
2 164 |
806 |
1 232 |
796 |
|
|
Horse beans |
932 |
1 475 |
1 661 |
416 |
1 105 |
897 |
|
|
Field peas |
1 056 |
1 235 |
1 010 |
n.ap. |
1 617 |
184 |
|
Foragesc |
|||||||
|
|
Oats/vetch |
6 900 |
6 100 |
5 800 |
6 300 |
3 500 |
4 700 |
|
|
Sorghum almum/Dolichos lablab |
4 500 |
3 400d |
5 500 |
2 600 |
2 000d |
4 200 |
a The farms had enterprises very similar to the package used by participating farmers.
b Seed yield (kg/ha).
c DM yield (kg/ha).
d Sorghum almum and field peas mixture.
On the sixth research farm - the specialist goat dairy enterprise - Toggenburg she-goats were first used, but were of poor quality and subsequently replaced by six Saanens. The latter were delivered in late 1979. Milk yields to date have been discouraging. A contributing factor to the poor performance of this system has been the limited experience of both the staff and the farm operator in handling goats. There is no tradition of specialist goat production or of goat milk consumption in the Ethiopian highlands. Individual lactation yields have not exceeded 100 litres over a 5-month lactation. The full potential of this enterprise has thus not yet been realized. Perennial forages, including lucerne, Rhodes grass and Panicum coloratum, have been sown on the farm's 1 ha of arable land. Shrubs, including Leucaena leucocephala, have been established on the non-arable 0.5 ha of hilly land.
Although the research farms at Debre Zeit are not replicated, results have been obtained and constraints identified which would not have been discovered so readily in a different research setting. Teff yields have on average been 23% higher than those achieved by participating farmers and 67% higher than on outside farms. Wheat yields on the research farms were 55% and 221% higher than on the plots of participating and outside farmers respectively. For pulses the yield margin of research farms over participating and outside farms was around 66%. Forage yields were twice those of participating farmers, and perennial species performed well under farmer conditions. More testing is needed in the smallholder context but it seems that there is a role for a cut-and-carry feeding system in the central highlands. The use of crossbred oxen and cows for traction has also given favourable results, offering the prospect of wider use of such innovative traction systems in a practical smallholder setting.
In 1979, ILCA established a 20-ha cooperative farm adjoining the station. It was managed jointly by 10 farmers nominated by the nearby Dembi PA. This extension of ILCA's activities was a response to government policy encouraging cooperative production. In the research context it may provide indications of the strength of any economies of scale.
A production package similar to the one for participating farmers was introduced. It included a dairy enterprise based on 12 Arsi x Friesian crossbred dairy cows, fertilizers and improved seeds. Cultivation implements and four each of local and crossbred oxen together with an animal shed were provided on credit. ILCA has made routine data collections to allow comparison of the performance of this cooperative with that of three outside ones. The three outside cooperatives belong to the PAs of Dembi, Godetti and Babogaya, three villages which surround the ILCA station. Individual outside farmers also belong to one of these three PAs, so that part of their labour is necessarily devoted to the three cooperatives. In 1979, the average amount of land cultivated cooperatively was 63 ha for each of the three PAs, but this fell to 14 ha in 1980.
These outside cooperatives are using fertilizer supplied to them more cheaply than to individual farmers (see Table 2.15). For the participating cooperative farm, the average labour input per ha in 1980 for field activities was 352 hours, compared with 740 hours for outside cooperatives and 440 hours for individual participating farms. Crop yields in 1979 were encouraging, but damage by floods and insects was substantial in 1980. Table 3.15 summarizes the average crop yields of participating and outside cooperatives for 1979 and 1980.
Table 3.15. Crop yields on participating and outside cooperative farms at Debre Zeit, 1979-1980 (kg/ha).
|
Crop
|
Participating cooperative |
Outside cooperativesa |
||
|
1979 |
1980 |
1979 |
1980 |
|
|
Teff |
1 170 |
660 |
762 |
903 |
|
Wheat |
1 170 |
711 |
184 |
n.ap.b |
|
Horse beans |
1 410 |
572 |
733 |
581 |
a Average for three cooperatives.
b Not grown in 1980.
The production of the participating cooperative is distributed to its members according to their individual labour inputs. These inputs included labour for both the crop and the livestock enterprise, and averaged 1406 ±121 hours per member in 1979, and 1895 ± 142 hours in 1980.
The participating cooperative farm is on black vertisols. Nineteen hectares out of the 20 are cultivated, half for food crops and half for forage crops. Sorghum almum with field peas is grown for green feed, while a mixture of oats and vetch is grown for hay making. In both 1979 and 1980 the forage yields were less than 3 t of DM per ha, due mainly to farmers' lack of experience in growing and managing fodder crops. The Arsi x Friesian cows originally supplied to the cooperative were poor milk producers, and in 1980 were replaced by eight Boran x Friesian crossbreds. To date the cows have produced an average of 4 litres of milk per head per day. The lack of previous dairying experience of these farmers is a serious handicap. It may take time before this venture will show convincing results, and as yet little can be said on the potential of cooperative dairy production.
3.4.1 Pasture and agronomy research
3.4.2 Animal traction research
3.4.3 Alternative energy sources for smallholders
The design of the pasture and agronomy research begun in 1978 has taken into account the various constraints of the existing agricultural system. Species selection and agronomic work were to be guided by the following considerations:
- land is generally scarce in the Ethiopian highlands, and forage species must be sufficiently high yielding to provide returns per ha comparable with food crops. On the basis of initial projections (ILCA, 1978) the target production level for the lower elevations in the highlands was set at 5 t of DM per ha, with 12% protein and using only moderate applications of phosphorus (P) (50 kg P2O5 per ha per year). The target production level at higher altitudes was 4 t of DM per ha, with the same protein content and fertilizer regime;- the medium and high elevations of the Ethiopian highlands would be the target areas for the forage research. Separate work should be conducted on the three main soil types found, namely the black vertisols of the bottomlands, the red soils of the lower slopes and the thin, poorer soils of the upper slopes;
- special-purpose forage production must be tailored to a subsistence-oriented farming system. Forage crops must combine with existing crop rotations. Annual forages are therefore more likely to be accepted by farmers than perennial forage crops. The initial research effort would therefore focus on annual forages;
- declining soil fertility, caused by continuous cropping and the export of nutrients, makes it necessary to concentrate on nitrogen-fixing legumes;
- peak labour demands at harvesting time and inadequate small-scale technologies limit the opportunities for silage making. Species should therefore be screened and evaluated for their use in cut-and-carry and hay making systems.
Table 3.16. Number of forage grass and legume accessions screened at Debre Zeit, 1978-1980.
|
Year |
Total accessions |
Accessions selected for further study |
|
1978 |
137 |
60 |
|
1979 |
235 |
155 |
|
1980 |
131 |
97 |
Screening and evaluation of species takes place as follows. Accessions are planted in rows 8 m long for screening. Grasses receive a basal dressing of 50 kg of P2O5 per ha, and are split into low-level nitrogen (N) (50 kg of N per ha) and high-level N (200 kg of N per ha) treatments. N is not applied to the legumes, which receive 25 kg of P2O5 per ha and 100 kg of P2O5 per ha respectively as low- and high-level fertilizer treatments. Promising species from this screening work are then evaluated in terms of their yield and nutritive value on production plots of 60 m² each in four replicates. At the evaluation stage the legumes receive standard phosphate fertilization of 50 kg of P2O5 per ha, while the annual grasses are tested at zero and 100 kg of N per ha. The most promising species or species combinations are then introduced on the research farms. Table 3.16 gives the number of accessions screened at Debre Zeit station in the period 1978-1980.
Table 3.17. Average annual performance of the most promising forages and mixtures for hay making at Debre Zeit, 1978-1980a.
|
Species |
DM yield (kg/ha) |
CP (%) |
|
Avena sativa cv Lampton |
4 550 |
7.8 |
|
Avena sativa cv Grey Algier |
4 850 |
8.2 |
|
Sorghum almum |
6 500 |
6.0 |
|
Avena sativa/Vicia dasycarpa |
3 950 |
14.4 |
|
Sorghum almum/Lathyrus purpureus |
6 700 |
10.2 |
The most promising grass species belong to the Avena, Sorghum and Lolium genera. Among the legumes, Vicia, Dolichos, Lathyrus, Trifolium and Medicago spp. are the most promising. Productivity results over the 3-year period indicate yield levels of between 4 and 6 t of DM per ha when the forage is used for hay making. Table 3.17 summarizes the results. The beneficial effect of legumes on the protein content of hay is amply shown. Although Sorghum almum alone or in mixture outyields the Avena cultivars, it is unsuitable for hay making.
Regrowth characteristics are important when forages are grown for green feeding. Increased protein yields per ha and higher average digestibility levels can be expected through a combination of early and late cutting for green feed. Table 3.18 gives the performance of different forages and mixtures for the purposes of green feeding. The table shows the higher persistency of Sorghum almum, indicated by its higher yield on the second cut, and hence its usefulness for green feeding. However, because of its low average feeding value, it should normally be used in a mixture with a legume.
Within species there were consistent differences in productivity on different soil types, as can be seen in Table 3.19. Yields were consistently lower on the heavy bottomland soils, although these are higher in soil fertility. They were depressed by the effects of excess surface water, persistent for most of the wet season, highlighting the potential contribution to productivity of improved drainage in these areas.
Table 3.18. Average annual performance of different forages and mixtures for green feeding at Debre Zeit, 1978-1980a
|
Species
|
Early cut |
Late cut |
Total |
|||
|
DM yield (kg/ha) |
CP (%) |
DM yield (kg/ha) |
CP (%) |
DM yield (kg/ha) |
CP (%) |
|
|
Avena sativa cv Lampton |
2 100 |
12.1 |
950 |
11.1 |
3 050 |
11.8 |
|
Avena sativa cv Grey Algier |
2 100 |
11.7 |
1 200 |
10.8 |
3 300 |
10.1 |
|
Sorghum almum |
1 350 |
9.0 |
1 950 |
6.3 |
3 300 |
7.4 |
|
Avena sativa/Vicia dasycarpa |
1 600 |
15.9 |
1 550 |
12.0 |
3 150 |
14.0 |
|
Sorghum almum/Lathyrus purpureus |
1 650 |
14.4 |
2 350 |
9.5 |
4 000 |
11.5 |
a Unfertilized.
Table 3.19. Average annual yields of the main forage crops on light and heavy soils at Debre Zeit, 1978-1980 (kg/DM/ha)a.
|
Species
|
Soil type |
|
|
Light |
Heavy |
|
|
Avena sativa cv Grey Algier |
7 050 |
3 410 |
|
Sorghum almum |
8 500 |
4 650 |
|
Avena sativa cv Grey Algier/Vicia dasycarpa |
8 600 |
5 600 |
a Unfertilized, single cut.
Application of N increased forage yields in most cases. The average yield increment was higher on the heavy bottomland soils, at 20 to 30 kg of DM per kg of N, than from the light soils on the arable slopes, where recorded returns were zero to 20 kg of DM per kg of N. The shortage of other nutrients in the light soils may have inhibited full response to N.
Work has also been carried out on improving the natural vegetation of eroded slopes through N and P fertilization of pastures oversown with legumes. The results of 3 years' experimentation indicate a poor response to N - less than 10 kg of DM per kg of N and no apparent effect of the legume on total productivity and quality. Cultivation of fodder shrubs and trees is an alternative use of these slopes which may allow these areas to contribute to the productivity of the farming system. This avenue is being researched and species including Leucaena leucocephala have been planted and are being evaluated.
Table 3.20. Average annual performance of perennial forage species at Debre Zeit, 1978-1980a.
|
Species |
DM yields (kg/ha) |
CP (%) |
|
Pennisetum purpureum |
4 800 |
8.3 |
|
Panicum coloratum |
4 000 |
7.1 |
|
Chloris gayana |
3 550 |
5.4 |
|
Pennisetum purpureum x P. typhoides |
3 100 |
5.0 |
|
Pennisetum purpureum/Desmodium intortum |
4 250 |
7.9 |
a Annual fertilizer application of 100 kg P2O5/ha.
Perennial grasses which have shown some promise are Chloris gayana, Panicum coloratum, and Pennisetum purpureum x P. typhoides (Banagrass). Average performances over the period 1978 to 1980 are provided in Table 3.20.
Establishing adequate stands of both perennial and annual forages on heavy bottomland soil proved difficult owing to the lack of surface drainage. For this reason different drainage treatments have been evaluated for these areas. The effects of the different treatments on various crops are summarized in Table 3.21. Natural vegetation without drainage served as the control for these trials, and yielded an average of 2 t of DM per ha over the period 1979-1980. As the table shows, crop yields on drained bottomland were considerably higher than on regularly cultivated arable plots on the lighter soils. If low-cost techniques can be used to develop and crop these more fertile bottomlands, they could contribute significantly to farm production. Both drainage works and cultivation in these areas require more draught power than can be provided by local oxen, so the use of larger, more powerful crossbred oxen is being evaluated as a farm-level means of allowing these areas to be brought into production.
At higher altitudes such as Debre Berhan, agricultural conditions are considerably harsher than at Debre Zeit. Little prior information was available regarding suitable forage species for these higher altitudes, with the result that ILCA strongly emphasized the screening work at Debre Berhan. At the same time, collection of the numerous ecotypes of local legumes found at these altitudes was initiated. These accessions are being screened at Debre Berhan. With regard to introduced species, the emphasis was on temperate varieties, with screening for productivity in both the meher and belg seasons. In addition to forage species, accessions evaluated in 1979 and 1980 included potato varieties from the International Potato Centre (CIP), and triticale and wheat varieties from the International Maize and Wheat Improvement Centre (CIMMYT). The total numbers of forage accessions tested are summarized in Table 3.22.
The research period has been too short to allow definitive selection of species suitable for the higher altitudes. The promising genera seem to be Avena and Lolium among the grasses and Vicia, Pisum and Lathyrus among the legumes. To date, DM yields of 2000 to 6000 kg per ha have been achieved, with Avena sativa var. 8237 as the highest yielding forage crop.
Table 3.21. Average annual yields of food and forage crops on bottomland under different drainage treatments at Debre Zeit, 1979-1980 (kg/ha).
|
Crop
|
Drainage treatment |
||
|
Ridge and furrow |
Ditch and dyke |
Ridge and furrow plus ditch and dyke |
|
|
Wheata |
1 150 |
1 200 |
1 750 |
|
Teffa |
1 150 |
1 500 |
1 350 |
|
Chick peas |
400 |
500 |
500 |
|
Horse beansb |
1 000 |
2 150 |
2 650 |
|
Avena sativa cv Lamptonc |
2 900 |
6 300 |
5 500 |
|
Sorghum almumc |
2 250 |
7 800 |
7 800 |
a Grain yield (kg/ha). Fertilizer application of 50 kg/ha DAP applied to all plots in both years.
b 1980 Only.
c DM yield (kg/ha).
The traditional agricultural system in the Ethiopian highlands depends heavily on animals for draught power to cultivate and transport agricultural produce. ILCA surveys in 1980 around Debre Zeit and Debre Berhan estimated that animal power used for crop activities on outside farms averaged 1050 animal hours per farm. Most of this power was supplied by oxen, but other classes of cattle were sometimes used for threshing. Some 50 to 80 hours were also contributed by donkeys transporting the produce from the field to the homestead12. Some 60 to 70% of the total animal power input was for seedbed preparation and planting: approximately 350 oxen-pair hours were used for these purposes.
12 Donkeys are again used to transport produce from the homestead to the market, but no data are yet available on this activity.
Research on various aspects of animal traction has been an important component of the ILCA Highlands Programme. Cultivation throughout most of the Ethiopian highlands is by the maresha, drawn by a pair of indigenous oxen each weighing from 250 to 350 kg (see Chapter 2). The number of animal-drawn metal implements in Ethiopia is insignificant. Few such implements have been introduced and tested in Ethiopia. Furthermore, the draught power developed by oxen of indigenous breeds is insufficient to operate heavy metal implements efficiently.
Table 3.22. Number of forage grass and legume accessions screened at Debre Berhan, 1979-1980.
|
Year |
Season |
Total accessions |
Accessions selected for further study |
|
1979
|
Belg |
101 |
60 |
|
Meher |
337 |
139 |
|
|
1980
|
Belg |
111 |
76 |
|
Meher |
186 |
96 |
Oxen are commonly worked for 4 to 9 hours per day. Oxen in the Debre Zeit area are worked throughout the day with few breaks. By contrast, farmers in the Debre Berhan area normally work oxen in two periods each day, and the animals are watered and fed in between these two sessions. The availability of draught power has important effects on area cultivated and crops sown. These effects are being investigated.
Oxen drawing the maresha have been observed working at an average speed of 2 km per hour and ploughing at an average depth of 13 cm in light soils with an average moisture content of 22% (by weight). In such conditions a pair of oxen produce an average power output of 0.68 kW per hour. The gross energetic efficiency of oxen, defined as the energy in the form of work accomplished as a percentage of energy consumed for maintenance and work on working days, ranges between 11% and 15%. These values highlight the relevance of making efficiency gains in the draught animal component of farming systems where forages are in short supply.
It was against this background that in June 1979 ILCA began comparative investigations into the use of crossbred oxen, which weigh more than the local breeds and produce more draught power. The use of crossbred oxen is linked with the prospect of the more widespread introduction of dairy enterprises using crossbred cows. Currently, the few thousand crossbred males produced annually in Ethiopia, mostly on state-owned dairy farms and ranches, are slaughtered soon after birth rather than reared and used for traction purposes. The routine use of crossbred males as draught oxen should allow cultivation of lands not tillable by oxen of indigenous breeds - and particularly the bottomlands with their heavier soils. Additionally, crossbred oxen can contribute to agricultural production on lands now tilled, because they cultivate faster and thereby allow more timely planting, and because they can draw a range of implements giving improved seedbed preparation and weed control, resulting in higher average crop yields.
The paragraphs which follow summarize the most important results of the animal traction research from June 1979 to the end of 1980. Work in progress includes the evaluation of different cultivation systems using oxen of local and crossbreeds, the technical and economic efficiency of using crossbred cows for draught purposes as well as milk production, and the use of oxen worked as singles rather than as the traditional pair. Different harnesses and yokes are also being developed.
Trials are being conducted over a 3-year period, 1980-1982, to compare cultivation systems based on improved implements with the traditional cultivation system. The systems will be compared in terms of cultivation and weeding times, power requirements of the different implements, and crop yields13. The cultivation trials are conducted at three sites: one each on the two soil types at the Debre Zeit station - clay loams - and the third at Debre Berhan, on the clay loams there.
13 Longer-term effects on soil structure are expected to arise from the use of deeper ploughing implements, but it will be some years before these effects (if any) become evident.
For the traditional system which acts as the experimental control, all seedbed preparation and seed covering is done with the maresha. For the first of the two new systems being evaluated, a mouldboard plough is used in the first pass, while secondary cultivation is done by zig-zag or chain harrow, according to the crop. In the second new system primary cultivation is by a spring tine cultivator, again followed by the use of zigzag or chain harrow according to the crop. A pair of crossbred oxen is used for drawing all improved implements, while a pair of local Zebu oxen pull the maresha. Details of the 1980 results are available in Abiye Astatke and Matthews (1981). Table 3.23 summarizes their results in terms of cultivation rates and energy costs of cultivation, while Table 3.24 gives the crop yields per ha for the various treatments on each soil type.
Table 3.23. Cultivation rates and total energy cost of three systems of cultivation on three soil types at Debre Zeit and Debre Berhan, 1980.
|
Soil type - Location/crop
|
Cultivation systema |
||||||
|
1 |
2 |
3 |
|||||
|
hours/ha |
(kWh/ha) |
hours/ha |
(kWh/ha) |
hours/ha |
(kWh/ha) |
||
|
Clay - Debre Zeit: |
|||||||
|
|
Teff |
157 |
(80) |
58 |
(50) |
49 |
(48) |
|
|
Wheat |
141 |
(80) |
55 |
(47) |
49 |
(38) |
|
|
Horse beans |
125 |
(62) |
49 |
(39) |
41 |
(30) |
|
Clay loam - Debre Zeit: |
|||||||
|
|
Teff |
158 |
(108) |
73 |
(70) |
76 |
(76) |
|
|
Wheat |
161 |
(120) |
66 |
(63) |
79 |
(80) |
|
|
Field peas |
138 |
(104) |
54 |
(50) |
64 |
(63) |
|
Clay loam - Debre Berhan: |
|||||||
|
|
Wheat |
95 |
(74) |
40 |
(34) |
31 |
(34) |
|
|
Barley |
90 |
(73) |
40 |
(34) |
30 |
(33) |
|
|
Horse beans |
92 |
(60) |
43 |
(39) |
32 |
(42) |
a System 1: all passes with the maresha.
System 2: primary pass with mouldboard plough, followed by zig-zag harrow or chain harrow.
System 3: primary pass with spring tine cultivator, followed by zig-zag harrow or chain harrow.
Table 3.24. Average crop yields for three systems of cultivation on three soil types at Debre Zeit and Debre Berhan, 1980 (kg/ha).
|
Soil type-Location/crop |
Cultivation systema |
|||
|
1 |
2 |
3 |
||
|
Clay - Debre Zeit: |
||||
|
|
Teff |
915 |
950 |
1 047 |
|
|
Wheat |
774 |
932 |
839 |
|
|
Horse beans |
1 992 |
1 764 |
1 147 |
|
Clay loam - Debre Zeit: |
||||
|
|
Teff |
1 707 |
1 680 |
1 553 |
|
|
Wheat |
2 254 |
2 250 |
2 138 |
|
|
Field peas |
843 |
905 |
802 |
|
Clay loam - Debre Berhan: |
||||
|
|
Wheat |
687 |
941 |
829 |
|
|
Barley |
1 022 |
774 |
842 |
|
|
Horse beansb |
n.av. |
n.av. |
n.av. |
a See Table 3.23 for details.
b Horse bean crops failed due to frost.
Cultivation hours per ha needed on the clay loam soils at Debre Berhan were significantly less than on either soil type at Debre Zeit. Overall, the system using the spring tine cultivator appears the most efficient as regards total hours spent per ha. Statistical tests showed no significant crop yield differences between systems. The results for 1980 only, as reported here, indicate that the total draught power required on the clay soils at Debre Zeit is 25 to 50% lower than that required on nearby clay loam soils. The clay loam soils at Debre Berhan require approximately the same draught power as the clay soils at Debre Zeit.
In addition to these trials, ILCA has initiated research on the use of cows as both milk producers and draught animals. A total of 32 cows are involved in the experimental research, with 16 each in a working and a non-working group. The cows were installed in late 1980. The trial is scheduled to continue until the end of the third lactations of the cows tested, when the effects on lifetime productivity can be estimated. If results show that cows can work without significant penalty to either reproduction rates or lactation yields, then peasant smallholders in the Ethiopian highlands will have available a development option which does away with the need to keep breeding stock in order to produce oxen which are worked only some 300 to 500 pair hours per year14. The testing of avenues such as these in order to achieve significant improvements in the overall efficiency of farming systems is a principal goal of ILCA's systems research in the Highlands Programme.
14 Preliminary results indicate that crossbred cows can be worked 4 or 5 hours a day with virtually no reduction in milk yield if they are fed a supplement to allow for the energy cost of field work.
Alternative sources of energy for household use are being explored by ILCA to permit cattle manure, which is now burned as domestic fuel, to be returned to the soil to maintain fertility. A range of tree species has been planted to identify high wood-yielding species. A manure-fuelled methane digester has been installed at Debre Zeit and is being used on an experimental basis. The digester is an Indian type with a metal gas collector of 2.83 m³ capacity. Using the manure from three stalled adult cattle, it produces enough methane for the cooking and lighting needs of a family of five. The sludge residue of the digestion process can be returned to the fields as fertilizer. Trials are being conducted to assess the value of sludge as fertilizer and to establish a practical modus operandi for the entire process.
A major constraint on the use of methane digesters is their capital cost. The plant at Debre Zeit costs approximately US$ 700, and a substantially lower capital cost will be necessary if the technology is to be applicable to smallholders with limited cash incomes. A secondary constraint on the use of digesters is the need to add approximately 1 litre of water with each kg of manure used. Domestic water supplies over much of the highlands are already limited, and this additional requirement may prove impossible to meet unless alternative water sources are developped.
To a large extent the capital cost problem may be overcome by constructing the below-ground Chinese type of digester instead of the Indian model. The Chinese digester uses a minimum of purchased materials and is also less prone to variable rates of methane production due to changes in ambient temperatures - a considerable advantage when average daily temperatures fall below 20°C. Tests with a Chinese type methane digester will be initiated during 1982.
