Tidskrift/serie: Crop Production Science
Utgivare: SLU, Institutionen för växtodlingslära
Utgivningsår: 1990
Nr/avsnitt: 7
Författare: Keftasa D.
Titel: Effects of developmental stages at harvest, nitrogen application and moisture. Availability on the yield and nutritional value of rhodes grass (Chloris gayana Kunth)-lucerne (Medicago sativa L.) pastures. Field studies at Kulumsa, Arsi, Ethiopia
Huvudspråk: Engelska
Målgrupp: Rådgivare
Nummer (ISBN, ISSN): ISBN 91-576-4165-X, ISSN 1100-1186

OBS! Fel i texten kan ha uppkommit då dokumentet överfördes från papper.

Effects of developmental stages at harvest, nitrogen application and moisture. Availability on the yield and nutritional value of rhodes grass (Chloris gayana Kunth)-lucerne (Medicago sativa L.) pastures. Field studies at Kulumsa, Arsi, Ethiopia

Daniel Keftasa



Effects of developmental stages, nitrogen application and moisture availability on yields and nutritional values of Rhodes grass (Chloris gayana Kunth) and lucerne (Medicago sativa L.) in pure stands and mixtures were studied at Kulumsa, Ethiopia, during 1984-1988.

In the first set of experiments the effects of seed proportion, amount of nitrogen application, and harvesting system on competition, dry matter yield and nutritional value of Rhodes grass and lucerne in pure stands and mixtures were studied. The mixtures produced higher dry matter yield, crude protein content and crude protein yield than Rhodes grass, the high-yielding monoculture by 22, 61 and 104 per cent, respectively. Cutting three times a year resulted in a higher competitiveness of lucerne and higher yields of nutrients in mixed stands than the conventional two-cutting system.

In the second and third sets of experiments, effects of developmental stages on nutritional values of Rhodes grass and lucerne in pure stands were studied. With advance in development the crude protein content declined, on average, at a rate of 0.16 and 0.11 percentage units daily with and without nitrogen application, respectively, in Rhodes grass and at a rate of 0.12 percentage units per day in lucerne. In vitro organic matter digestibility declined at rates of 0.27 and 0.14 percentage units daily in Rhodes grass and lucerne, respectively.

In the fourth set of studies effects of variation of moisture availability on dry matter yields in Rhodes grass-lucerne pastures were analysed using different water balance models. All indices of moisture availability, namely rainfall, estimated actual evapotranspiration, length of growing period and moisture index, closely explained the seasonal variation in dry matter yields. From statistical relationships between moisture availability and dry matter yields in the pastures, probability distribution curves of dry matter yields were determined using the rainfall data over the last 22 years.

Utilization of the results in practice are discussed in relation to the technical and economic implicatons of legume-based pasture systems in improved dairy cattle feeding in Ethiopia.

Key words: Rhodes grass (Chloris gayana Kunth), lucerne (Medicago sativa L.), dry matter yield, competition, relative yield total, crowding coefficient, relative reproductive rate, developmental stages, crude protein, in vitro organic matter digestibility, metabolizable energy, neutral detergent fibre, acid detergent fibre, lignin, major minerals, length of growing period, WATBAL, soil water balance, evapotranspiration, water index.

List of papers included in the thesis

This thesis is based on the following papers, which are referred to in the text by capital letters as indicated below:


Field studies at Kulumsa, Arsi, Ethiopia.

Livestock husbandry plays a vital role in Ethiopian agriculture. Cattle are the main source of milk, meat, skins and hides. In addition oxen are the main source of draught power, indicating close interdependence of crop and livestock production. Approximately 60 per cent of the total land area of the country is utilized for raising livestock, which are thus the largest single users of the land resource (Taylor, 1984). About 80 per cent of the total of 27 million cattle, 88 per cent of the estimated 48 million human population, and 95 per cent of the regularly cropped land exist in the high-potential ( > 1 500 m alt. & 700 mm rainfall) agricultural areas which cover only 43 per cent of the country's land area. This indicates a high density of both human and livestock population.

Land which was previously used for grazing is gradually brought into cultivation to satisfy the food needs of the increasing human population. Livestock are forced to graze on marginal land and use crop residues which are often low in both quantity and quality, which results in poor livestock performance. To improve livestock productivity, research on crossbreeding, feeds and nutrition has been conducted over the last 20 years. In Arsi, mean lactation yields of local Zebu cattle (Bos indicus) range from 220 kg at first lactation, in unselected population to about 500 kg at third lactation, in selected population, (Schaar et al., 1981). Under fairly good management, cross-bred dairy cows (50% local Arsi x exotic dairy breeds, Bos taurus, mainly Friesian) produce about 2 000 kg milk per lactation (300 to 800% increase over local cows) at farm level (Schaar et al., 1981; Olsson, 1986). Besides upgrading the genetic potential of the cattle, introduction of high yielding nutritious forage crops, mainly forage legumes, has become necessary. It has long been known that forage legumes are more nutritious as livestock feed than grasses and crop residues but this potential is often underutilized. In addition to being high quality feed for livestock, forage legumes contribute to raising the overall productivity of farming through increased soil fertility and eventually increase in the yield of subsequent crops when used in rotation with cash crops. Reviews made by Haque & Jutzi (1984) and Tothill (1986) indicate that legumes have enormous potential roles to improve the present farming in Sub-Saharan Africa. Integration of forage legumes in rotation with the existing cereal-dominated farming systems, such as in the Ethiopian highlands appears to be the most important package as cereal-legume link and thus crop-livestock interdependence. On the other hand, experience shows that pure stands of forage legumes do not persist for a long period, being more quickly invaded by weeds and also attacked by insects and diseases as compared to grasses.

Grass-legume pastures are usually superior to either pure stands of grass or legume, both in herbage yield, quality and ease of utilization. Grass-legume mixtures have extra advantages of extending the period of pasture growth and the overall quality of the herbage does not drop as fast as in pure grass stands.

In Ethiopia, Rhodes grass (Chloris gayana Kunth) and lucerne (Medicago sativa L.) are among the most suitable pasture species for low to medium (1500-2300 m above sea level) altitude zones and are utilized to a relatively larger extent than any other species so far introduced. Rhodes grass is a popular grass in the tropics and sub-tropics of eastern and southern Africa, Australia and central America. Originating in eastern and southern Africa, it is valued for its ability to set seed, the relative ease of its establishment and ability to cover the ground, its tolerance to drought, to light frost and to soil salinity as well as its suitability to be grown in association with many tropical legumes, clover and lucerne. A detailed review of botanical and agronomic characteristics and utilization of Rhodes grass has been published by Bogdan (1969). Agronomic characteristics and cultivation techniques of Rhodes grass with reference to seed production has been published by Boonman (1973) in Kenya and by Keftasa (1985) in Ethiopia.

Lucerne is probably the oldest cultivated forage crop in t0he world (Hanson, 1988). Although it originated in the Mediteranean climate, it is grown in almost all parts of the world and found valuable in tropical highlands due to its continued growth during cool seasons. A detailed review by Leach & Clements (1984) on adaptation of lucerne and management of lucerne-based pastures in the sub-tropics concludes that research on lucerne in pure stands and in mixtures with tropical grasses is limited. There are few reports published on the influence of seed proportion at planting, level of nitrogen application, and of cutting frequency on yield, species composition and nutritional value of the mixtures of lucerne with grasses under tropical environments.

The studies reported in this thesis were carried out to investigate effects of management practices on yield, competition and quality of Rhodes grass mixtures (Paper A) and effects of developmental stages on the nutritional value of Rhodes grass (Paper B) and lucerne (Paper C) separately. In Paper D, a detailed description of weather parameters is presented and effects of variability in moisture availability on dry matter yields of Rhodes grass-lucerne pastures are studied.

All trials were conducted in the field at Kulumsa, in Ethiopia (details of soils and climatic conditions are shown in Paper D).

The crude protein content was determined by the Kjeldahl nitrogen procedure (N x 6.25).

In vitro organic matter digestibility (IVOMD) was determined by the modified one-stage in vitro technique (Lindgren, 1979).

Metabolizable energy content (ME) was calculated from IVOMD values using the following equations suggested by Lindgren. For Rhodes grass and mixtures of > 50 per cent Rhodes grass

Y = 0.16x - 1.91

For lucerne and mixtures of > 50 per cent lucerne

Y = 0.106x + 2.93

where x = IVOMD and Y = ME MJ (KG OM)-1

Neutral detergent fibre (NDF), acid detergent fibre (ADF), ADF ash and lignin were analysed on selected samples of experiments reported in Papers B and C according to Goering & van Soest (1970).

Contents of phosphorus, calcium, potassium, magnesium and sodium were analysed according to A.O.A.C. (1975).

Summary of papers presented in the thesis

Paper A. Effects of harvesting systems and nitrogen application on competition, yield and nutritional value of Rhodes grass and lucerne in pure stands and mixtures

The subject of competitive interference on mixed stands of plant communities has been widely studied (de Wit, 1960; Donald, 1963; Harper, 1964; de Wit & van den Bergh, 1965; Rhodes, 1970; Håkansson, 1988). Use of mixed stands of grass-legume pastures is often preferred due to higher yield, better seasonal distribution of the production and higher quality of mixtures than in pure stands of grasses. It has been well documented that productivity of mixtures and their balance are highly influenced both by environmental and management factors. Experimental studies and literature reviews of mixtures of lucerne and grasses (mainly in temperate environments) show that nitrogen application (Vallis, 1978; Frankow-Lindberg, 1987) and cutting frequency (Leach 1978; Leach & Clements, 1984) influence the competitiveness of lucerne.

The study reported in Paper A presents effects of nitrogen application and cutting frequency on the overall productivity, the balance of mixtures and the major nutritional quality of Rhodes grass and lucerne pastures.

The trial included both Rhodes grass and lucerne alone and three different seed combinations of the two species. Four rates (0, 46, 92 or 138 kg ha-1 yr-1) of nitrogen were applied, and the herbage was cut two or three times a year. The data taken included dry matter yield, botanical composition, in vitro organic matter digestibility and contents of crude protein and major mineral elements. Competition was quantified by crowding coefficients (van den Bergh, 1968) and the change in proportion of lucerne with time was quantified by the relative reproductive rate (de Wit, 1960).

The main results can be summarized as follows:

Paper B. Effects of developmental stages and nitrogen application on yield and nutritional value of Rhodes grass (Chloris gayana Kunth)

In many situations, the nutritional value of tropical grasses is low (French 1957; Minson & McLeod, 1970; Minson, 1971; Moore & Mott, 1973). Among other factors, the advanced stage of development at grazing or cutting has been recognized as a major factor responsible for a low nutritional value of pasture grasses. Milford & Minson (1968), Minson (1971), Soneji et al. (1972) and Said (1974) reported rapid decline in digestibility and protein content of Rhodes grass with advance in development. The effect of nitrogen application on the nutritional value of grass herbage has, in general, been variable and usually interacts with the developmental stage (Gomide et al., 1969; Ford & Williams, 1973; Minson, 1973; Wilson, 1973).

The experiment reported in Paper B was designed to study the effect of the stage of development on the nutritional value of Rhodes grass fertilized with different levels of nitrogen.

Pure stands of Rhodes grass were established in 1984 and four levels of nitrogen fertilizer (0, 46, 92 or 138 kg N ha-1 yr-1) were applied in the beginning of each growing season. Three sub-plots of 0.25 m2 were cut from each fertilizer treatment at an interval of about 10 days during each of four growing seasons in 1985-87. Developmental stage, herbage yield, in vitro organic matter digestibility (IVOMD), neutral detergent fibre (NDF), acid detergent fibre (ADF), ADF ash, lignin and major mineral elements were determined.

The main results can be summarized as follows:

Paper C. The nutritional value of lucerne (Medicago sativa L.) in different developmental stages

Lucerne is widely grown as a forage crop in temperate and sub-tropical climatic zones as a source of high quality forage for livestock (Hanson, 1988). In Ethiopia there is an increasing interest in lucerne due to its fast growth and persistence in cool temperatures and heavy soils of high altitudes.

Among many others, Kilcher & Heinriches (1974), Kalu & Fick (1983), Wilman & Altimini (1984), Buxton et al. (1985) and Nordkvist & Åman (1986), have reported that the deterioration of the lucerne quality with advancing development is mainly associated with the proportional decrease in leaf mass and rapid increase in cell wall concentration and lignification of stems with growth. Although the trends of changes in the nutritional value are similar in different studies, the extent of changes is influenced by environmental conditions.

The study reported in Paper C describes the rates of changes in crude protein, in vitro organic matter digestibility (metabolizable energy content), fibre, lignin and major mineral elements with advancing development under the prevailing conditions at Kulumsa in Ethiopia. The experiment was carried out in the field during four growing seasons in 1985-87. Sub-plot samples were taken at intervals of about 10 days in each season, and the nutritional value was determined.

The main results are summarized as follows:

Paper D. Effects of moisture availability on the yield in Rhodes grass-lucerne pastures. Evaluation using water balance models

Pasture growth is closely related to the amount and distribution of rainfall. In most parts of Ethiopia rainfall is the most extremely varying climatic parameter.

Although a close linear relationship between dry matter yields and rainfall often exists in seasonally dry tropical environments, a considerable amount of water is lost through surface run-off and/or deep percolation. Currently, a number of models are available which partition rainfall into that effectively used by plants and that lost in different ways.

Calculations of soil water balance are being widely used to relate actual water supply to pasture production. Some studies (e.g. McCown et al., 1974) reported a closer relationship of pasture dry matter production to actual evapotranspiration (AET) than between production and rainfall alone as a measure of moisture availability. Besides rainfall as a source of moisture supply, differences in soil types and capacity of available water storage influence the moisture supply and pasture yields (McCown, 1973).

The study reported in Paper D contains analyses of relationships between yield and moisture availability for Rhodes grass-lucerne pastures at Kulumsa during 1985-1988.

The length of the growing period (LGP) was calculated from weather parameters for the years 1984-88 according to the F.A.O. (1978) model. The weekly soil water balances were simulated using a water balance modelling program commonly known as WATBAL (Keig & McAlpine, 1974), from which actual evapotranspiration (AET), soil moisture store, moisture index (ratio of actual to potential evapotranspiration) were estimated. Seasonal and annual pasture yields were related to rainfall, AET, moisture index and LGP separately, using linear regressions. From the calculated relationships between AET and dry matter yields, probability curves of expected yields of different pasture systems were constructed for each of the 22 years on the basis of rainfall data.

The main results are summarized as follows:

General discussion

A large volume of literature reviews suggest that the main constraint in livestock production in the tropics is the low nutritional value of tropical grasses. Tropical climate results in rapid growth and development and a rapid decline in nutritional value of pastures. The studies reported in this thesis demonstrate that the nutritional value of mixtures (Paper A) and pure stands of Rhodes grass (Paper B) and lucerne (Paper C) were highly influenced by the developmental stage of the plants at cutting. In Rhodes grass the decline in nutritional value with advancing development is associated with the stem development and the leaf scenescence and accumulation of dead material which is high in fibre and lignin. In lucerne, the main cause of decline in nutritional value is shedding of leaves due to ageing, which is often accompanied by leaf and stem diseases. In mixtures, delayed cutting resulted in decline in overall quality, largely as a result of reduced proportion of lucerne due to vigorous growth of the grass. In the studies reported in Papers B and C, declines in the nutritive value are expressed in terms of growth days starting from the beginning of the rainy season. An alternative system, based on developmental stages was also included when appropriate. Cutting based on growth days was favoured mainly to avoid the problems of inconsistencies in terms used to identify and classify the developmental stages. Estimates of rates of the changes with growth days reported in Papers B and C for Rhodes grass and lucerne, respectively, give a guideline when to cut for desired level of nutritional value under comparable environments. Extending such studies to different pasture systems under different environments would improve our ability to predict nutritional values of pasture plants.

The major effect of nitrogen application was improved dry matter yields of Rhodes grass and also of the mixtures. The effect on the nutritional value of the pastures was variable and mainly influenced by the stage of development. Generally, nitrogen application improved crude protein content and to some extent also the in vitro organic matter digestibility in pure Rhodes grass, particularly at early stages of development (Paper B). In mixtures nitrogen application depressed the nutritional value of the pasture as a total by depressing the proportion of lucerne, particularly under less frequent cutting systems. Although a number of literature reviews conclude that the influence of nitrogen application on the contents of cell wall constituents and digestibility is generally small, the results reported in Papers A and B suggest that N application improves the digestibility of the grass to a considerable level if the grass is cut at early stages but depresses it if the grass is cut at advanced stages of development. The explanation is that, at early developmental stages, nitrogen application stimulates new growth of tillers of high nutritional value but as a result of rapid growth the rate of stem development, flowering, leaf senescence and accumulation of dead material are rapid and nitrogen fertilized grasses may therefore eventually become inferior in quality in comparison with non-fertilized grass.

Lucerne by itself produced low yields, mainly due to a high invasion by weeds and attack by leaf diseases. The main contribution of lucerne was an improvement of the nutritional value of the pasture (Paper A). The study reported in Paper C shows that, at advanced stages of development, the nutritional value of lucerne did not drop to the same extent as that of pure stands of Rhodes grass (Paper C). The rapid recovery of lucerne after cutting therefore creats an opportunity to cut earlier and increase the number of cuts, keeping a high nutritional value, and to extend the pasture growth period towards the early part of the dry season using residual moisture in the soil from the main rainy season.

The availability of moisture during growth was a determinant factor for the seasonal variation in production (Paper D). The pasture responses to nitrogen application and frequent cutting were improved by increased amounts of available moisture during growth. On the analogy of the improvement of the responses to N application, the efficiency of water use in the pasture was also improved by nitrogen application, mixing grass with lucerne and more frequent cutting. Although not shown in the study reported in Paper D, reviews of literature (Wilson, 1982) suggest that the nutritional value of pastures can possibly be influenced by the availability of moisture. In many situations, moisture stress retards development in terms of plant size, stem elongation and flowering resulting in leafy swards of low dry matter yields but high nutritive value. Other climatic factors that can limit pasture productivity in Ethiopia are low temperatures and low insolations due to overcast conditions, particularly at altitudes above 2 500 m.

For practical application of the results from the present experiments and similar findings, close analysis of the existing livestock husbandry systems of the country are required. In Ethiopia, about 98 per cent of the total livestock population is indigenous stock owned by small-scale farmers. The main feeding system is open communal grazing on native pastures and aftermath crop stubbles. In such a system, utilization of inorganic fertilizers or sowing of pastures is unlikely in the present economic situation. The conditions of crop production at subsistence level create a severe competition with pasture improvement for land, labour and the limited capital. The other livestock production sector involves improved dairy and beef cattle owned by the state, by farmer's cooperatives or by private farmers with more intensive management practices. The main feeds are agro-industrial by-products, hay from native pastures and some supplementary grazing. Since cereals and pulses are used for human food there is an acute shortage of by-products such as bran and oil-seed cakes for livestock feeding. In addition, because of shortages of distribution and transportation facilities, the use of such by-products is limited to state farms close to the main towns.

Although research results clearly indicate that pasture productivity can be attained at a high level by application of nitrogen fertilizer, the use of fertilizers will be limited by inavailability and high costs. At present, only about 10 per cent of the Ethiopian farmers use fertilizers for the production of food crops and the use of fertilizers on pastures therefore seems unrealistic in the near future.

The future development of an intensively managed livestock husbandry appears to be dependent on legume-based pastures conserved as hay or silage. This might not be the best alternative for small-scale private farmers due to shortage of land, but it can be feasible in conditions where farmers are dependent on dairy and have a good market for their products. In commercial livestock production where land and capital are not constraint, intensively managed grass-legume pastures can be a dependable source of high quality feeds.

Based on the results of the experiments reported in this thesis a Rhodes grass-lucerne pasture grown under fairly favourable rainfall conditions fertilized with 92 kg ha-1 yr-1, and cut three times a year, can produce about 8 000 kg dry matter ha-1 yr-1 containing 30 per cent lucerne. Such pastures can be sufficient to satisfy the nutrient requirements of two cross-bred dairy cows and their calves and produce about 4 000 kg milk and 440 kg live weight from the calves of a total value of about 2 880 Ethiopian Birr (1 352 US dollar; based on common prices of 0.5 Birr kg-1 milk and 2.0 Birr kg-1 live weight). Instead, if the same land is used for wheat production the expected yield is about 2 500 kg ha-1 which is worth about Birr 1 375 (USD 664, based on .55 Birr kg-1, after Florell, 1987).

In the following tables the nutrient requirements of cross-bred dairy cows and their calves are shown and compared with the possible production of nutrients from a Rhodes grass-lucerne pasture.

Table 1.1. Daily nutrient requirements of a 350 kg crossbred dairy cow and a growing calf

tible crude
Maintenance 41.0 255.0 14.0 11.0
Pregnancy a 9.0 120.0 8.0 6.0
For production of 1 kg (4% fat) milk 5.0 60.0 2.6 1.8
A growing calf b 36.6 348.8 16.0 8.0

Source: Animal Husbandry Department, Asela, Ethiopia.

a Mean values for the last three months of pregnacy.

b Mean values for a growing calf at the rate of 0.5 kg day-1 during the first year.

Table 1.2. Total annual nutrient requirements of two dairy cows and two growing calves

tible crude
Maintenance 29930 186.15 10.22 8.03
Milk production 2000 kg cow-1 20000 240.00 10.40 7.20
Pregnancy (3 months) 1620 21.60 1.44 1.08
Growing calves up to one year 26718 254.62 11.68 5.84
Total amount yr-1 78268 702.37 33.74 22.15

Table 2.0. Nutrient content and total production from one hectare of Rhodes grass-lucerne pasture and balance with the total annual requirement in Table 1.2

ME MJ CP (kg) DCP (kg) Ca (kg) P (kg)
Content kg-1 10.4 .181 .130a .0059 .0032
Total production 8000 kg dm ha-1 yr-1 83200 1448 1040.0 47.2 25.60
Balance +4932 +337.38 +13.46 +3.45

a DCP (%) = 0.899 CP (%) - 3.25; according to Milford and Minson (1965).


This study was supported by

The Swedish Institute,

The Institute of Agricultural Research, IAR-Ethiopia,

The Swedish Royal Academy of Agriculture and Forestry,

The Development Cooperation Office, SIDA-Ethiopia,

The International Livestock centre for Africa and

The Department of Crop Procuction Science, SUAS.

All are gratefully acknowledged.

I wish to express my sincere gratitude to:


Association of Official Analytical Chemists, 1975. Official methods of analysis (12th ed). Association of Official Analytical Chemists, Washington, D.C.

Bogdan, A.V. 1969. Rhodes grass, Review article. Herbage abstracts 39 (1).

Boonman, J.G. 1973. On the seed production of tropical grasses in Kenya. Agr. rep. Verslagen van landbouwkundige onderzoekingen no. 794 Netherlands.

Buxton, D.R., Hornstein J.S., Wedin, W.F. & Marten G.C. 1985. Forage quality in stratified canopies of alfalfa, birdsfoot trefoil and red clover. Crop Science 25: 273-279.

Donald, C.M. 1963. Competition among crop and pasture plants. Advance in Agronomy 15: 1-117.

F.A.O 1978. Report on the Agro-Ecological Zones Project. Vol. 1. Methodology and results for Africa. Rome.

Florell, P.O. 1987. Data in agriculture. SEAD publication no. 5 Asela, Ethiopia.

Frankow-Lindberg, B.E. 1987. Lucerne-grass swards with different nitrogen application and grass components. 2. Competition. Swedish J. agric. Res. 17: 185-191.

French, M.H. 1957. Nutritional value of tropical grasses and fodders. Herbage abstract. Review article 27: 1-9.

Ford, C.W. & Williams, W.T. 1973. In vitro digestibility and carbohydrate compostition of Digitaria decumbens and Setaria anceps grown at different levels of nitrogenous fertilizers. Aust. J. Agric Res. 24: 309-316.

Goering, H.K. & van Soest, P.J. 1970. Forage fibre analysis (apparatus, reagents, procedures and some application). Agricultural Handbook no. 379 USDA-ARS, Beltsville, Maryland.

Gomide, J.A., Noller, C.H., Mott, G.O., Conrad, J.H. & Hill, D.L. 1969. Effects of plant age and nitrogen fertilization on the chemical composition and in vitro cellulose digestibility of tropical grasses. Agronomy J. 61: 116-120.

Hansson, A.A. (ed) 1988. Alfalfa and Alfalfa improvement Agronomy Monograph no. 29. Madison, USA.

Haque, I. & Jutzi, S. 1984. The potential role of forage legumes in African production systems. Paper presented at the joint FAO/AEA/IG coordinated research program on the use of nuclear techniques in pasture management held in Vienna, Austria 26-30 Nov. 1984.

Harper, J.L. 1964. The nature and consequences of interference amongst plants. Proc. of the XIth International Conference on Genetics. Genetics today 2: 465-482.

Håkansson, S. 1988. Competition in stands of short-lived plants. Density effects measured in three component stands. Crop Production Science 3. Swed. Univ. Agric. Scs. Uppsala, Sweden.

Kalu, B.A. & Fick, G.W. 1983. Morphological starge of development as a prediction of alfalfa herbage quality. Crop Science 23: 1167-1172.

Keftasa, D. 1985. Rhodes grass seed production. Forage network in Ethiopia newsletter no. 7: 19-21 ILCA, Addis Abeba, Ethiopia.

Keig, G. & McAlpine J.R. 1974. WATBAL. A computer system for the estimation and analysis of soil moisture regimes from simple climatic data. CSIRO division of land use research, Canberra. Technical monograph 74/4 Second edition.

Kilcher, M.R. & Heinrichs, D.H. 1974. Contribution of stems and leaves to the yield and nutrient level of irrigated alfalfa at different stages of development. Can. J. plant Sci. 54: 739-742.

Leach, G.T. 1978. The ecology of lucerne pastures. In: J.R. Wilson (ed). Plant relations in pastures pp: 290-308.

Leach, G.T. & Clements, R.J. 1984. Ecology and grazing management of alfalfa pastures in the sub-tropics. Advances in Agronomy 37: 127-154.

Lindgren, E. 1979. The nutritional value of roughages determined in vivo and by laboratory methods (English summary). Avdelningen för husdjurens näringsfysiologi. Sveriges lantbruksuniversitet. Rapport 45 Uppsala, Sweden.

McCown, R.L. 1973. An evaluation of the influence of available soil water storage capacity on growing season length and yield of tropical pastures using simple water balance models. Agricultural meteorology 11: 53-63.

McCown, R.L., Gillard, P. & Edye, L.A. 1974. The annual variation in yield of pastures in the seasonally dry tropics of Queensland. Aust. J. Exp. agric. Anim. Husb. 14: 328-333.

Milford, R. & Minson, D.J. 1965. The relation between the crude protein content and the digestible crude protein content of tropical pasture plants. J. Br. Grassl. Soc. 20: 177-179.

Milford, R. & Minson, D.J. 1968. The digestibility and intake of six varieties of Rhodes grass, Aust. J. Exp. agric. Anim. Husb. 8: 413-418.

Minson, D.J. 1971. The nutritive value of tropical pastures. J. Aust. Inst. Sci. 37: 255-263.

Minson, D.J. 1973. Effect of fertilizer nitrogen on digestibility and voluntary intake of Chloris gayana, Digitaria decumbens and Panicum clandestinum Aust. J. Exp. agric. Anim. Husb. 13: 153-157.

Minson, D.J. & McLeod, M.N. 1970. The digestibility of temperate and tropical grasses. Int. Grassl. Congr. Proc. 11th Surfers Paradise, Australia, pp. 719-722.

Moore, C.E. & Mott, G.O. 1973. Structural inhibitors of quality in tropical grasses. In: A.G. Matches (ed). Anti-quality components of forages. CSSA. Special publication no. 4. Crop Science Society of America INC. Madison Wisconsin, pp 53-98.

Nordkvist, E. & Åman, P. 1986. Changes during growth in anatomical and chemical composition and in vitro degradability of lucerne. J. Sci. Food Agric. 37: 1-7.

Olsson, A. 1986. Performance of dairy cattle at the research and dairy cooperative farms in the Arsi region of Ethiopia. Report from a minor field study. Int. rural dev. Center, Swed. Univ. Agric. Scs. Working paper no. 39, Uppsala, Sweden.

Rhodes, I. 1970. Competition between herbage grasses. Herbage abstacts. Review article 40 (2): 115-121.

Said, A.N. 1974. The nutritive value of some Kenya feed-stuffs and the effects of protein and energy rich concentrate supplementation on the Utilization of Chloris gayana hay by wether sheep. ph. D. thesis. University of Nairobi, Kenya.

Schaar, J., Brännäng, E. & Meskel, L.B. 1981. Breeding activities of the Ethio-Swedish integrated rural development project. Part II: Milk production of Zebu and cross-bred cattle. Wld Anim. Rev. (F.A.O) 37: 31-37.

Soneji, S.V., Musangi, R.S. & Olsen, F.J. 1972. Digestibility and feed intake investigations at different stages of growth of Brachiaria ruziensis, Chloris gayana and Setaria sphacelata using corriedale wether sheep. II. Chemical composition and yield. East Afric. Argic. Forest. J. 37: 267-271.

Taylor, M.S. 1984. Assistance to land use planning in Ethiopia. Livestock production and feed resources. Technical report no. 6. F.A.O. Rome.

Tothill, J.C. 1986. The role of legumes in farming systems. In: Potentials of forage legumes in sub-Saharan Africa. Proceedings if ILCA workshop. Addis Abeba, Ethiopia.

Vallis, I. 1978. Nitrogen relationships in grass-legume mixtures. In: J.R. Wilson (ed). Plant relations in pastures. CSIRO, Australia pp: 190-201.

Wilman, D. & Altimini, M.A.K. 1984. The in vitro digestibility and chemical composition of plant parts in white clover, red clover and lucerne during primary growth. J. Sci. Food Agric. 35: 133-138.

Wilson, J.R. 1973. The influences of aerial environment, nitrogen supply and ontogenetical changes on the chemical composition and digestibility of Panicum maximum Jacq var. Trichoglume Eyles. Aust. J. agric. Res. 24: 543-556.

Wilson, J.R. 1982. Environmental and nutritional factors affecting herbage quality. In: J.B. Hacker (ed). Nutritional limits to animal production from pastures. CAB, Farnham Royal, U.K. 111-131.

de Wit, C.T. 1960. On competition. Verslagen van landbouwkundige onderzoekingen 66: 1-82 Netherlands.

de Wit, C.T. & Bergh, van den. 1965. Competition between herbage plants. Neth. J. Agric. Sc. 13: 212-221.