Availability and use of dry season feed resources on smallholder dairy farms in central Kenya

Agroforestry Systems 50: 315–331, 2000.  2000 Kluwer Academic Publishers. Printed in the Netherlands. Availability and use of dry season feed resources on smallholder dairy farms in central Kenya THE LATE O. Z. NYAATA1, P. T. DORWARD2, *, J. D. H. KEATINGE2, 3 and M. K. O’NEILL4 1 Formerly of Kenya Agricultural Research Institute, Regional Research Centre – Embu, P.O. Box 27, Embu, Kenya; 2 The University of Reading, Department of Agriculture, Earley Gate, P.O. Box 236, Reading, RG6 6AT, England; 3 International Institute of Tropical Agriculture, PMB 5320, Ibadan, Nigeria; 4 Agricultural Science Center, New Mexico State University, P.O. Box 1018, Farmington, New Mexico 87499, USA (*Author for correspondence, E-mail: p.t.dorward@reading.ac.uk) Key words: Calliandra calothyrsus, forage, leguminous fodder, manure, Napier grass (Pennisetum purpureum) Abstract. A cross-sectional survey on 41 farms followed by six weeks monitoring of dairy cattle feeding on ten smallholder dairy farms in central Kenya was conducted to investigate the use, availability and quality of dry season feed resources. Fodder production was largely from Napier grass (Pennisetum purpureum) grown on small plots and contour strips where it acts both as a fodder source as well as a biological barrier to soil erosion. There is a need to broaden the choice of fodder crops on such farms to provide a wide range of harvesting management options and to avoid total loss in case of pest or disease outbreaks. Intercropping of Napier grass with leguminous fodder trees could boost the quantity and quality of herbage production especially during the dry season. Roughage from a variety of sources was utilised during the dry season in addition to Napier grass. Among the herbages, leguminous feeds had the lowest potential dry matter degradability while weeds harvested from cropland and roadsides had the highest. Energy and protein intake from the roughage fed to grade dairy cattle during the dry season may be insufficient to meet the requirements of these animals due to the high levels of fibre concentration (acid detergent fibre and neutral detergent fibre (ADF and NDF)) in them. It is recommended that the scope for alternative sources of improved roughage such as Napier/ calliandra mixtures, to boost the energy, protein and overall dry matter provision on the farms should be investigated further. Introduction In the recent past there has been a rapid shift towards a zero-grazing (cut and carry) system of dairy cattle management on smallholder farms (tea and coffee zones) in central Kenya. This has largely been brought about by the high rate of population growth resulting in less land being available for pasture production (Orodho, 1990) but is also due to other changes including disease management. The system involves permanently confining animals in units where they are fed on forage cut and carried to the feeding stalls. The main feed source for dairy cattle in central Kenya is Napier grass (Pennisetum purpureum) and other high yielding fodder grasses and legumes (Minae and Nyamae, 1988). Fodder is harvested from small plots or along soil conser- 316 vation contour lines. The amount and quality of fodder produced per hectare varies with species, season and soil fertility. In most farms, however, there is growing evidence that even during the rainy season the amount of fodder available for livestock is inadequate in both quality and quantity. This situation is acute during the dry season when animals are underfed and often malnourished (Minae and Nyamae, 1988; NARP II, 1993). The use of Napier grass with crop residues such as maize (Zea mays) and bean (Phaseolus spp.) stover, banana (Musa spp.) leaves and pseudostems, and sweet potato (Ipomoea batatus) vines as dry season supplements is one method adopted by farmers to alleviate nutritional problems of livestock (Murithi et al., 1993). Many farmers also use commercial concentrates as a supplement to basal fodder diets, but concentrate prices are high relative to milk prices. Understanding the various strategies used by farmers to manage dry season fodder shortages is important in guiding the development and dissemination of technologies to alleviate this problem. The objectives of the study were to assess the use and availability of feed resources on smallholder dairy farms of central Kenya as well as to identify the management strategies currently practised to alleviate dry season animal malnutrition and to recommend potential options. Materials and methods The study area The study was conducted in the Manyatta and Runyenjes Divisions of Embu District at an altitude range of 1300–1800 m. The soils are mainly humic Nitisols (typic Palehumult) derived from basic volcanic rocks. They are deep, well weathered with friable clay texture and are moderate to high in inherent fertility (Jaetzold and Schmidt, 1983). Rainfall distribution is bimodal, falling during the months of March to July and October to December, with a total annual average range of 1200–1500 mm. The temperatures are warm with a monthly average of 19 °C. The area has high population density with an average of 450 persons per km2 (based on 1979 national census). All land is demarcated and registered under the freehold system, with average farm size of 1.5 ha (Minae and Nyamae, 1988). Dairy production plays an important role in the area both as a food source and for cash generation. Cattle for instance, contribute 30% of the total household revenue of which 89% is from the sale of milk (Murithi, 1998). The main cattle breeds are grade (exotic) animals which have a genetic potential of 25 kg of milk per day, and their crosses with local East African Shorthorn. Milk productivity per lactating animal in the area is normally between three and 10 kg per day in the dry season, and five and 15 kg per day in the wet season. Another important product of the dairy industry is manure, which is a major input for Napier grass and cash crops like coffee (Coffea arabica). 317 Manure collection is made easier by the adoption of a zero-grazing system of cattle management by most farmers in the study area. Goats and poultry are also common in the area. An average household owns two livestock units (where a livestock unit is equivalent to one adult cow); usually one adult cow, one heifer and a calf (Minae and Nyamae, 1988). Farmer interviews A list of two hundred participating farmers in the National Dairy Development Project (NDDP) who are also members of two milk marketing co-operatives (Muriithi and Wang’ombe Dairy Co-operatives) was used to draw up a sample framework for the formal survey. Forty-six farmers were randomly selected from the list for interview. All farmers selected were from the tea and coffee zones of Embu District. Questionnaire pre-testing was conducted on five farms and to familiarise the enumerators with the exercise and to formulate hypotheses. The remaining forty-one farmers were interviewed using a structured questionnaire in mid-September 1995. Feeding management monitoring In addition to the formal interviews, 10 farms were randomly selected in which feeding was monitored. At each farm, feed was monitored over two week periods in October 1995, November 1995 and February 1996. October and February are dry season months associated with feed scarcity while November is the peak of the rainy season when feeds are more widely available. A 50 kg spring balance with units of 0.5 kg was used for weighing feeds at each farm. Samples of various varieties of feed were pooled separately and subsampled each week for chemical analysis. Refusals were weighed every afternoon, accumulated weekly, subsampled and prepared for chemical analysis. The types of feeds offered to cattle during the exercise were classified into: Napier grass (Pennisetum purpureum); weeds (roughage cut from road sides and cropland); fibrous crop residues (maize [Zea Mays] stover, banana [Musa spp.] stems and leaves and sweet potato [Ipomoea batatus] vines); fodder legumes (Calliandra spp., Desmodium spp. and lucerne [Medicago sativa]); and commercially processed feeds (concentrates) which included dairy meal, wheat or rice bran and pollard (a by-product from the oil extraction process from maize grain). Liveweights of each cow were estimated through heart girth measurements by the use of a ‘Weighband’ (Dalton Supplies Limited, Nettlebed, UK) at the start of the monitoring. Analyses Feed samples were dried at 65 °C to constant weight and ground to pass through a 1 mm sieve before analysis. Dry matter (DM), ash, crude protein (CP), neutral detergent fibre (NDF) and acid detergent fibre (ADF) were 318 determined using standard methods (AOAC, 1980; Van Soest, 1965). Potential degradability (Bailey and Hironaka, 1970) of all the feed was determined by incubating about 3 g of forage in dacron bags (with a pore size of 40µ) suspended for 48 hours in the rumen of a steer fitted with a permanent cannula. The fistulated animal was fed a diet of Napier grass and concentrate (16% CP). Data from the field exercise and chemical analyses of the feed samples were statistically analysed using SAS (SAS, 1985). Standard deviations and means were calculated for numbers of livestock owned, areas of fodder crops grown and quantities of roughage and concentrates offered to cattle during dry seasons. Analyses of Variance (ANOVA) were carried out to test for differences between farms and between seasons, in quantities of roughage and concentrates offered, and in chemical composition of feeds. Results from survey General Information Most respondents (83%) in the survey were the owners of the farms. A few (12%) of the people interviewed were sons of the owners of the land, while the rest (5%) were managers with full mandate to manage the land on behalf of the farmers. All the farmers interviewed kept dairy cows; either exotic or crossbreds, averaging 2.4 (s.d. 1.2) cows per farm. Friesian cows were the most popular among farmers, followed by Ayrshires, Guernseys, cross bred cows and Jerseys (Table 1). Indigenous goats were the commonest small ruminants and found on 42% of farms. Dairy goats (exotic goats) and sheep were recorded on 17% and 22% of the farms respectively. Ninety eight percent of farmers practised a zero-grazing or semi zero-grazing system of dairy cattle management. Table 1. Distribution of dairy cattle breeds owned by farmers in Embu District. % farmers* (n = 41) Number of cows Total %* 1 2 3 4 5 31.7 26.8 09.8 07.3 07.3 07.3 26.8 09.8 04.9 0– 04.9 0– 4.9 2.4 – – – – 4.9 – – – – – 2.4 2.4 – – – – Dairy breed Friesians Ayrshire Guernsey Zebu Crossbreed Jersey * Percentage do not sum to 100 because some farmers keep more than one breed. 70.7 41.4 14.7 07.3 12.2 07.3 319 Fodder production All the farms visited had Napier grass, either in small plots (blocks) or on the contour lines, where it acted both as a source of fodder for livestock and as a biological barrier to soil erosion. The average Napier area for the study farms was 0.5 ha (s.d. 0.3). Other forage crops found on the farms include Nandi setaria (Setaria splendida), sweet potato vines (Ipomoea batatus), desmodium (Desmodium spp., both green and silver leaf desmodium were observed on the farms), and Guatemala grass (Tripsacum laxum) (Table 2). A number of pasture grasses were also observed on the contour lines in coffee fields, with ruzi grass (Brachiaria ruziziensis) being the commonest. These grass species also form an important source of fodder. Planting and management of Napier grass Twenty-two percent of the farmers who grew Napier grass in blocks intercropped it with desmodium while 5% of the farmers who grew Napier grass strips intercropped it with Calliandra calothyrsus. The numbers of farmers using cattle manure and inorganic fertilisers in Napier production are given in Table 3. Average application rates for NPK fertiliser at planting of Napier were 100 kg ha–1. For cattle manure average application rates at planting were 40 t ha–1 (equivalent to 240 kg N, 216 kg P2O5 and 260 kg K2O according to Taiganides, 1987). The average rate of calcium ammonium nitrate application as an annual maintenance was 192 kg ha–1 (about 50 kg N). Application of manure as a maintenance fertiliser to Napier was very variable and only reported by 37% Table 2. Distribution of species of fodder crops grown by farmers in Embu District. Fodder spp. Napier grass (Pennisetum purpureum) Nandi setaria (Setaria splendida) Sweet potato vines (Ipomoea batatus) Desmodium spp. Pennisetum/Desmodium (mixtures) Guatamala grass (Tripsacum luxam) Ruzi grass (Brachiaria ruziziensis) Neg. = Negligible area. Area in ha Minimum Maximum Mean s.d. Number of farmers n = 41 % < 0.02 1.6 0.46 0.31 41 100 < 0.02 0.1 0.04 0.03 07 017.1 < 0.02 < 0.02 0.4 0.4 0.10 0.15 0.14 0.16 09 05 022.0 012.2 < 0.02 0.4 0.14 0.17 04 009.8 0 Neg – – – 01 002.4 0 Neg – – – 01 002.4 320 Table 3. Farmers use of cattle manure and inorganic fertiliser in Napier grass production in Embu District. % of farmers (n = 41) Inorganic fertiliser (mainly NPK 20:20:20) Cattle manure Cattle slurry Calcium ammonium nitrate (26% N) At planting Annual maintenance 29 54 14 0– 0– 37 17 75 Some farms utilise more than one source. of farmers. A small proportion (17%) of farmers applied slurry as a maintenance fertiliser and this was mainly by gravity flow of the washings from zero-grazing units to the nearest areas of Napier grass plots. Overall, more than half of all cattle manure produced in 59% of the farms visited was reported to have been applied to fodder production plots. However the percentage of manure applied to fodder crops varied greatly across farms (0 to over 90%). The timing of applications of maintenance fertiliser varied considerably across farms (Table 4), probably as a result of the different strategies adopted by farmers to increase dry season feed, or because other activities take priority in terms of labour resource allocation. Forty-nine percent of farmers reported that the amount of manure produced on their farm did not satisfy their needs. When asked how they deal with the shortage, 20% said they rotate applications among fields, 15% buy from other farms (import), while 15% supplemented the manure with inorganic fertilisers. The majority of farmers (81%) interviewed collected and piled manure outside the zero-grazing units to await application during planting time. A large number (73%) of the farmers interviewed harvested (cut back to ground level) their Napier grass when it attained a height of 1.5 m (which Table 4. Time of application of maintenance dressings of fertiliser/cattle manure to Napier grass in Embu District. Time Onset of the rainy season Early part of the rainy season Peak of the rainy season Whenever there is enough labour Onset and end of the rainy season Agro-ecological zone Tea zone Coffee zone No. of farmers n = 38 8 8 0 3 1 11 03 01 01 00 19 11 01 04 01 %* 50 29 03 11 03 * Percentages do not sum up to 100 because some farmers do not apply maintenance fertilisers/ manure. 321 during the rainy season usually coincides with a period of eight weeks regrowth). Twenty-two percent of these farmers however, sometimes harvested their grass more often than this because of fodder scarcity. Feed types and availability during the dry season February and September were reported by the majority of farmers to be the most critical months in terms of fodder shortage while April, May, June and November have least fodder problems (Figure 1). Maize stover and other crop residues were used as dry season supplements by all farmers. Use of banana leaves and pseudostems was reported at 98% of the farms visited. Other feeds reported by fewer farmers include weeds (from coffee fields and road sides) sweet potato vines, tree leaves, Nandi setaria and ruzi grass (from soil conservation strips in coffee fields) and concentrates (mainly wheat and rice bran) (Table 5). Some farmers purchased Napier grass and/or maize stover from neighbours. Forage management for dry season feeding Sixty-six percent of farmers reported that they produced excess fodder in some months of the year. However, fodder conservation was not a common practice in this area. Of the farmers who occasionally get excess fodder 63% reported that they either leave it in the field until the need arises or sell it to their neighbours. Other uses of excess feed include compost making, mulching and bedding for cattle and only one farmer reported making hay with Napier grass. Various forms of maize stover storage were observed on the farms including stacking it outside around ornamental or fruit trees and under cover within the zero-grazing units. Figure 1. Yearly fodder distribution in Embu District, Kenya. 322 Table 5. Common feeds utilised during the dry season in Embu District. Type of feeds No. farmers reporting n = 41 %* Maize stover Banana leaves and pseudostems Weeds Sweet potato vine Tree leaves (both indigenous and exotic) Nandi setaria Purchase of cereal concentrate Purchase of maize stover/Napier grass 41 40 21 09 05 02 11 08 100 098 051 022 012 005 027 020 * Some farms utilise more than one source of fodder. Results from feed monitoring Background The ten participating farmers had a variable number of grade cows of mixed breeds (mainly Friesians, Ayrshires, Guernseys and Jerseys). There were an average of 2.9 cattle per farm and their liveweights ranged from 260 to 508 kg. In all farms animals were fed from concrete troughs. A variety of forage was utilised in each farm and were hand-chopped into 3–5 cm lengths, weighed and fed ad-libitum depending on availability. A range of types and amounts of concentrates were fed by farmers at milking times (morning and evening). Some farmers provided the animals with a mineral supplement when it was available while water was made available at all times. Dry season nutrient supply to dairy cows In all farms, Napier grass and other roughages contributed most DM offered to cattle in the three seasons studied. The amount of roughage offered to animals varied significantly (P < 0.01) amongst farms (5.4–18.2 kg cow–1 day–1) and seasons (9.7–14.6 kg cow–1 day–1). The amount of concentrates offered to each lactating animal per day (Table 7) also varied significantly (P < 0.01) amongst farms (0.6–4.4 kg cow–1 day–1) as well as between seasons (0.9-2.1 kg cow–1 day–1). The average was 1.7 kg per animal per day across farms. The number of farmers not offering cattle concentrate feed increased to six by the third season of the study. The dry matter content of most feeds varied significantly (P < 0.05) amongst seasons as well as farms (Table 8). Higher DM content in the dry season was recorded consistently only for Napier grass and legume fodder. Feeds are variable and the nutrient content of feeds varied significantly (P < 0.05) between seasons and farms. Crude protein content among legu- 323 minous fodder was consistently higher than in other roughages over seasons as well as farms. Variation in ADF with season did not reach significant (P > 0.05) levels in Napier grass and weeds. Similarly, the ash content of by-products, the NDF levels in weeds and the CP content of leguminous fodder were not significantly affected by seasons. However, dry matter degradability varied significantly (P < 0.05) between seasons and farms. Weeds attained the highest average degradability while legume fodder had the least among the roughages. The ash content of legume fodder was consistently lower than for all other feeds. Concentrate feeds had the highest average CP content and dry matter degradability across seasons and farms. However, variation in nutrient content of concentrates was apparent across farms (P < 0.05). Dry matter content of the concentrates was significantly (P < 0.05) affected by both seasons and farms. Discussion and conclusions The average size of the 41 farms involved in the survey was 2.2 ha (s.d. 1.8) and is slightly larger than reported in two other surveys conducted in central Kenya (Minae and Nyamae, 1988; Murithi et al., 1993). Participants in the National Dairy Development Project may have slightly greater access to resources, including contact with extension, which should be borne in mind in interpreting the results. Findings from the survey provide useful information on the availability of feed. A variety of forages are used but most farmers depend on Napier grass to supply the bulk of the forage requirement for their animals in this agro-ecological zone, which is consistent with the findings of Potter (1987). Forage shortages in the dry seasons are clearly a constraint and were reported by more than 60% of farmers. However, some farmers experienced this constraint throughout the year and more than 25% reported that forage was scarce for seven months of the year. Results from the feed monitoring survey give a useful insight into the quantities and qualities of feed supplied to animals during the dry season. Quantities offered were similar to those observed by Potter (1987) from similar systems which gave an average availability of less than 5 kg DM per cow. Three percent of liveweight is frequently used as an estimate of potential dry matter intake (McDonald et al., 1995) but is unrealistic with poor quality feed. In low input and output systems few farmers would be expected to meet these requirements and Table 6 therefore includes calculations at 2.3, 2.6 and 2.9 % of liveweight. At 2.9% of liveweight seven of the ten farmers offered less than the potential dry matter intake in one or both dry seasons and five farmers in the wet season. At 2.3% of liveweight this reduced to five farmers in one or both dry seasons, and four farmers in the wet season. The quality of feeds offered was variable but generally poor. The DM 324 Table 6. Variation across seasons and farms in roughage DM offered to cattle (kg cow–1 day–1) during the dry season in Embu District. Farmer Season 1 Season 2 Season 3 Farm No. (Dry) (Wet) (Dry) mean Average Potential dry matter intake* liveweight of 2.6% 2.9% cows farm–1 2.3% 01 02 03 04 05 06 07 08 09 10 05.1 10.0 06.4 05.9 09.4 21.7 16.5 17.1 04.2 11.7 06.1 13.0 06.1 09.4 10.4 12.8 16.4 11.6 05.1 06.4 05.0 12.0 13.2 13.1 06.4 20.2 19.2 18.8 18.2 09.7 345 414 328 289 334 459 398 338 281 395 Mean 10.8 b 09.7 c 14.6 a 05.4 11.8 08.6 09.5 12.1 18.2 17.4 15.8 09.2 09.2 e c d d c a a b d d 07.9 09.5 07.5 06.6 07.7 10.6 09.2 07.8 06.5 09.1 09 10.8 08.5 07.5 08.7 11.9 10.3 08.8 07.3 10.3 10 12 09.5 08.4 09.7 13.3 11.5 09.8 08.1 11.5 Note: Means with different letters along the same columns or row are significantly different (P < 0.05). * Calculated as % of the average liveweight. content and nutrient concentration in Napier grass reported in this study are similar to those given by Anindo and Potter (1994) for the dry season in Muguga, Kenya. The high levels of NDF and ADF found in Napier grass are indicative of the fibrous nature of the grass and are associated with slow fermentation and long rumen retention. However, the measures of degradability given for the high fibre feeds like Napier grass should be interpreted with caution given the relatively short period that samples were incubated in the rumen (48 hours). Ideally this and rates of protein breakdown should be measured over a period of at least 72 hours (AFRC, 1993). High NDF values were also reported in most of the other feeds, including concentrates, and will affect energy and protein intake. In addition levels of crude protein (CP) reported in all feeds were low. Maize stover which was widely used in the dry season had high levels of NDF (83%) and low levels of CP (4.5%) and these are close to values reported by Methu et al. (1996) of 81–91% and 4.8–8.3% respectively. Similarly, percentage CP for maize and concentrates agreed with values given by FAO (1981, 1991). The intake of energy, and protein, is therefore unlikely to be sufficient to sustain satisfactory animal production levels by the large breed animals (Meissner et al., 1991) commonly found in this area. Despite the benefits of sampling and measuring feed offered by farmers rather than relying on farmers estimates, limitations of feed monitoring need to be recognized when interpreting results. Farmers may be influenced by the process and offer better than normal forage when observed by an outsider, particularly if the monitoring is for a short period of time. The timing of the monitoring (duration and period within seasons) is also important and normally 325 constrained by costs and logistics. Types and qualities of forage offered may vary within seasons and weeds for example are offered whenever weeding is carried out (Massawe, 1999). In this study, although monitoring on each farm occurred for a total of six weeks (three two-week periods), some fluctuations may not have been observed. The importance of short-term fluctuations in types of feed offered is evident from recent research in Kenya that reported poor performance of animals when the quality of forage was frequently changed (Sanda et al., 1999). Feed studies have generally focused on the quality and quantities of forage supplied but variations in forage type experienced in many smallholder systems may affect efficiency of digestion and overall animal performance. It is evident therefore that overall, inadequate provision of digestible dry matter as well as poor feed quality are likely to be the main limiting factors to milk production in most smallholder farms during the dry season. The remainder of the paper focuses on means of addressing the constraints identified, and in particular on the potential role of agroforestry. Areas warranting further investigation are identified. The use of cattle manure as a fertiliser is the cheapest means of increasing the productivity and quality in existing forage production systems on most smallholder dairy farms. Findings from the survey showed that the proportion of available manure applied to fodder crops varied considerably between farmers. This was determined by, amongst other factors, the importance of dairying as a source of farm income. However, only 20% of farmers reported that they attempted to reduce nutrient losses from the manure either by storing it under a roof or in concrete pits with covers. Slurry applications are also not usually incorporated into or covered with soil on most farms. Studies on cheap and simple means of on-farm manure management and means of reducing nutrient losses through leaching and volatilisation are therefore warranted. Despite the low availability of fodder experienced by many farmers in the area during the dry season, 63% said that they occasionally had surplus feeds during the rainy season. The feasibility of on-farm feed conservation strategies (e.g. simple grass and legume hay-making techniques using wooden boxes, [Onim et al., 1986, Massawe et al., 1999]) to make surplus rainy season feeds available during the dry season should be investigated. Maize stover was reported to be plentiful in the dry season and low cost methods of treatment and storage (e.g. urea treatment [Methu et al., 1997]) to increase its nutritive value could also be explored. Agroforestry systems may offer means to improve the supply and quality of forage in the dry season and to reduce requirements for inorganic and organic fertilisers. Several leguminous trees offer potential including Leucaena spp., Sesbania spp. and Gliricidia spp. as do natural occurring or naturalised tree species that are already fed to livestock in the area, such as Morus alba and Manihot glaziovii. However Calliandra calothyrsus warrants particular consideration as it is already grown on some farms in the study area and is 326 being promoted by the National Agroforestry Research Project to provide fodder for cattle and as a means of stabilising soil along contour lines (O’Neill et al., 1994). Furthermore, there has been considerable interest in the species and a recent adoption study involving farmers who participated in the on-farm trials in the area found that over four-fifths of the farmers expanded their calliandra plantings after their first planting (Franzel et al., 1996). Both the low levels of concentrate used and the poor quality of forages (Tables 7 and 8) suggest the need for low cost alternatives for sources of protein. Calliandra forage with its high crude protein content (23.5–31% of dried foliage [Kaitho et al., 1993, Thijssen et al., 1993]) can offer an important protein supplement. However, the low potential DM degradability coefficient often found in calliandra may indicate low protein digestibility, probably because of the high levels of tannins reported in calliandra leaves (Ahn et al., 1989; Kaitho et al., 1993). Nevertheless, feed trials conducted in central Kenya have demonstrated the potential of calliandra as a protein supplement to low quality forages. Results from on-farm trials showed that the same increase in milk production resulted from feeding either 1 kg of commercial concentrate (16% CP) or 3 kg of fresh calliandra to Ayrshire and Friesian cows. The butterfat content of the milk also increased from 4.0 to 4.5% by the addition of calliandra (Paterson et al., 1996). Besides improving voluntary feed intake and animal performance, inclusion of leguminous leaves in grass based diets greatly enhances the quality of manure of the animals. For example, the N concentration of litter (wood shavings plus manure) of goats feeding on Panicum maximum, supplemented with leaves of Leucaena leucocephala, increased with the increase in the percentage of leucaena in the diet (Cobbina et al., 1989). However without Table 7. Variation with seasons and farms in concentrate DM (kg cow–1 day–1) offered to cattle in ten farms in Embu District. Farmer No. Season 1 (Dry) Season 2 (Wet) Season 3 (Dry) Farm mean 01 02 03 04 05 06 07 08 09 10 1.7 3.3 – 1.7 2.0 1.5 4.4 3.1 1.4 2.3 1.7 2.5 – – 0.6 1.7 4.4 4.3 2.6 1.7 1.5 0.3 – – – 1.3 – 5.6 – – 1.6 2.0 – 0.6 0.9 1.5 2.3 4.4 1.3 1.3 Mean 2.1 a 1.9 b 0.9 c d c g f ed b a e e Means with different letters along the same column or row are significantly different (P < 0.05). 327 Table 8. Variation with farms in the chemical composition of feeds offered to cattle during the dry season in Embu District. Feed Farmer No. 1 2 3 4 5 6 7 8 9 Napier grass % DM (2.199) % Ash (1.584) % CP (1.650) % NDF (2.491) % ADF (2.583) % DMD (2.352) 15.1 13.3 10.3 75.8 41.5 42.3 16.8 12.9 12.6 73.1 39.0 45.2 19.0 11.1 10.8 68.3 35.2 44.1 16.6 14.4 11.8 72.5 41.8 40.0 15.8 13.0 13.4 71.5 38.2 45.1 16.0 14.6 13.3 69.7 39.0 45.7 17.5 12.7 11.5 73.8 43.9 39.4 15.7 13.9 11.1 73.3 42.5 39.3 16.1 14.6 12.3 72.1 39.9 42.4 Concentrates % DM (5.410) % Ash (4.087) % CP (3.686) % NDF (6.873) % ADF (9.970) % DMD (10.75) 90.9 06.7 20.0 52.5 24.6 52.8 84.4 09.4 14.1 75.2 51.7 47.9 84.3 07.1 13.9 50.1 14.1 59.7 85.0 05.8 14.3 54.8 14.0 60.0 79.8 06.0 15.7 58.9 16.8 63.0 83.6 11.0 15.5 64.9 14.1 63.5 Fodder legume % DM (2.719) % Ash (0.611) % CP (2.638) % NDF (4.480) % ADF (4.893) % DMD (3.376) 26.3 06.9 14.9 58.6 35.1 32.2 28.3 05.4 17.4 54.3 37.2 31.0 28.3 05.2 18.6 49.3 36.8 31.8 27.3 05.6 18.8 55.3 42.2 33.1 25.6 06.3 19.9 56.1 40.2 35.2 Weeds % DM (4.06) % Ash (3.133) % CP (2.412) % NDF (5.511) % ADF (4.802) % DMD (5.838) 19.1 12.2 11.7 70.9 39.2 47.7 19.6 13.2 15.0 66.2 34.3 48.5 19.9 11.1 14.6 69.8 35.7 48.5 By-products % DM (16.28) % Ash (2.271) % CP (6.641) % NDF (9.930) % ADF (3.512) % DMD (8.742) 27.1 10.5 12.4 66.2 42.8 41.6 14.7 11.5 11.9 64.0 29.2 47.4 31.8 06.4 08.7 64.8 33.0 47.7 Note: LSDs are included in parenthesis against each component; DM = Dry matter; CP = Crude protein; NDF = Neutral detergent fibre; ADF = acid detergent fibre; DMD = Dry matter degradability. improvements to manure storage and management noted above, the benefits of improved manure quality may not be realised. Franzel et al. (1996) reported that most calliandra planting by farmers involved in on-farm research in the Embu area, was on contour lines, home- 328 stead boundaries or intercropped with food crops and coffee. Intercropping calliandra with Napier grass however could potentially boost the quantity and quality of forage per unit area of land. A further benefit of grass/ leguminous tree intercropping is the maintenance of soil fertility through biological nitrogen fixation. According to Bethlenfalvay and Phillips (1977), removal of large quantities of nitrogen from the legume, as in the cut and carry system, will increase the demand of the plant for available nitrogen, thus stimulating the rate of nitrogen fixation. This could reduce the need for fertiliser and manure on forage and allow the reallocation of these scarce resources to other crops on farms. However, although the removal of large quantities of legume foliage may stimulate N fixation, it may also result in depletion of K and other nutrients, and in the need for applications of these for levels of production to be sustained. Work is particularly required on calliandra/Napier intercrops to investigate the effects of spacing and density of calliandra, cutting frequencies and timing, and of organic and inorganic fertiliser application on forage yield and quality. It is evident that on most farms in the study area there is competition for scarce resources such as manure, labour and cash. Decisions regarding their allocation are likely to be based on a range of factors including the availability and prices of the resources, and the relative prices of different farm products e.g. coffee vs milk at a given time. New land uses such as agroforestry will have implications for resource allocation on farms. These may be positive e.g. reduced fertiliser requirements and increased forage supply, or negative e.g. increased labour demands. Some agroforestry systems may also reduce farmers flexibility to change crops e.g. in response to prices in subsequent years. The potential resource and management implications of potential changes to farming practices therefore need to be carefully explored with farmers in designing and adapting systems, prior to conducting on-farm research. Recently developed Participatory Farm Management approaches have been successfully used with farmers in Ghana to explore and evaluate the potential of green manuring (Galpin et al., 2000) and could offer means to achieve this for agroforestry systems. This paper has provided information on feed availability and agronomic management practices in small-scale dairy farms in central Kenya. From the findings of the survey and feed monitoring, it is evident that solutions are needed to the constraints of poor quality and quantities of dry season forage. Agroforestry systems involving tree legumes appear to offer one means of addressing these needs within the resource constraints and conditions that small-scale dairy farms operate. Research is particularly warranted on the design and management of Napier/calliandra intercropping to enable the benefits of the system to be maximised, and its potential disadvantages minimised. As part of this, socioeconomic considerations regarding farmers access to and allocation of scarce resources need to be addressed. 329 Acknowledgements The work was funded by Swedish International Cooperation Development Agency (Sida). Early drafts of this paper were commented on by Drs Peter Cooper and Ben Dzowela (ICRAF) and a later draft by Professor Emyr Owen (The University of Reading, UK). Further information was kindly provided by Ralph Roothaert. Agricultural extension staff in Runyenjes and Manyatta Divisions of Embu District assisted during the survey and feeding monitoring exercise in their divisions. References AFRC (1993) Energy and protein requirements of ruminants. Advisory manual prepared by the AFRC Technical Committee on Responses to Nutrients. CAB International, Wallingford, UK Ahn JH, Robertson BM, Elliot R, Gutteridge RC and Ford CW (1989) Quality assessment of tropical browse legumes: tannin content and protein degradation. Anim Feed Sc Tech 27: 147–157 Anindo DO and Potter HL (1994) Seasonal variation in productivity and nutritive value of Napier grass at Muguga, Kenya. East Afr Agric For J 59: 177–185 AOAC (1980) Official Method of Analysis. 13th Ed. Washington DC Bailey CB and Hironaka R (1970) Maximum loss of feed from nylon bag in rumens of steers as related to apparent digestibility. Can J Anim Sc 50: 325–330 Bethlenfavay GJ and Phillips DA (1977) Ontonetic interactions between photosynthesis and symbiotic nitrogen fixation in legumes. Plant Physiology 60: 419–421 Cobbina J, Atta-Krah AN and Kang BT (1989) Leguminous browse supplementation effect on the agronomic value of sheep and goat manure. Biological Agriculture and Horticulture 6: 115–121 FAO (1981) Tropical Feeds. Feed Information Summaries and Nutritive Values. FAO Animal Production and Health Series No. 12. FAO, Rome, Italy FAO (1991) Tropical Feeds. Database, Version 2.1. Software developed by Oxford Computer Journal. FAO, Rome, Italy Franzel S, Arimi H, Karanja J and Murithi F (1996) Boosting milk production and income for farm families: the adoption of Calliandra calothyrsus as a fodder tree in Embu District, Kenya. In: Mugah JO (ed) People and Institutional Participation in Agroforestry for Sustainable Development, pp 421–438. Proceedings of the First Kenya Agroforestry Conference. Kenya Forestry Research Institute, Muguga, Nairobi, Kenya Galpin M, Dorward P and Shepherd D (2000) Participatory Farm Management methods for agricultural research and extension: a training manual. The University of Reading and the Department for International Development (DFID) Jaetzold R and Schmidt H (1983) Farm Management Hand Book of Kenya Vol. II parts B and C; Eastern and Coast Provinces. Ministry of Agriculture, Nairobi, Kenya Kaitho RJ, Tamminga S and Bruchem J (1993) Rumen degradation and in vivo digestibility of dried Calliandra calothyrsus leaves. Anim Feed Sc Tech. 43: 19–30 McDonald P, Edwards RA, Greenhalgh JDF and Morgan CA (1995) Animal nutrition. Longman Scientific and Technical, Harlow, UK Massawe NF (1999) Strategies based on participatory rural appraisal for improving the utilisation of forages to increase profitable milk production on smallholder farms on Tanzania. PhD Thesis, Department of Agriculture, The University of Reading, UK Massawe NF, Owen E, Mtenga LA, Romney DL, Ashley SD and Holden SJ (1999) The economics of maize stover transportation in northern Tanzania: 1. Cost effectiveness of loose 330 vs manual baling and using different sizes of trucks. Proceedings of the 25th Scientific Conference of Tanzania Society of Animal Production 25: 5–12 Meissner HH, Koster HH, Nieuwouldt SH and Coertze RS (1991) Effect of energy supplementation on intake and digestion of early and mid-season rye grass and panicum/Smuts finger hay, and on in sacco disappearance of various forage species. S Afr J Anim Sc 21: 33–42 Methu JN, Owen E, Abate A, Mwangi DM and Tanner JC (1996) Smallholder dairying in central Kenya highlands: practices in the utilization of maize stover as a feed resource. In: Fungoh PO and Mbadi GCO (eds) Focus on Agricultural Research for Sustainable Development in Changing Economic Environment, pp 243–251. Proceedings of the 5th KARI Scientific Conference. Kenya Agricultural Research Institute, Nairobi, Kenya Methu JN, Owen E, Abate A, Scarr MJ and Tanner JC (1997) Effect of offering three amounts of maize stover to dairy cows on intake and selection. Proceedings of the British Society of Animal Science, 85 Minae S and Nyamae D (1988) Agroforestry Research Project Proposal for the Coffee Based Land use System in the Bimodal Highlands Central and Eastern Provinces, Kenya. AFRENA Report No. 16. International Centre for Research in Agroforestry (ICRAF). Nairobi, Kenya Murithi FM (1998) Economic evaluation of the role of livestock in mixed smallholder farms of the central highlands of Kenya. Unpublished PhD thesis, The University of Reading, Department of Agriculture Murithi F, O’Neill MK, Thijssen HC, Mugendi DN, Mwangi JN and Nyaata OZ (1993) Agroforestry technologies for fodder production and soil fertility improvement in Meru District, Kenya. AFRENA Report No. 76. International Centre for Research in Agroforestry (ICRAF). Nairobi, Kenya NARP II (1993) Chapter 7 Livestock Research. National Agricultural Research Project (NARP). Kenya Agricultural Research Institute (KARI), Nairobi, Kenya O’Neill MK, Murithi FM, Nyaata OZ, Mugendi DN, Mwangi JN, Gachanja SP, Tuwei P and Thijssen HC (1994) National Agroforestry Research Project, pp 11–18 and pp 38–39. KARIRegional Research Centre-Embu. Annual Report: March 1993–March 1994. AFRENA Report No. 81. International Centre for Research in Agroforestry (ICRAF). Nairobi, Kenya Onim JF, Mathuva M, Otieno K and Fitzhugh HA (1986) Forage resources for small ruminants in the humid zones of Africa, pp 145–155. In: Proceedings of a workshop on improvement of small ruminants in Eastern and Southern Africa, OAU/IBAR, Nairobi, Kenya Orodho AB (1990) Dissemination and utilisation of research technology on forage and agriculture by-products in Kenya. In: Dzowela BH, Said AN, Asrat Wendem-Agenehu and Kategile JA (eds) Utilisation of Research Results on Forage and Agriculture By-Product Materials as Animal Feed Resources in Africa, pp 70–90 Paterson RT, Roothaert RL, Nyaata OZ, Akeyeampong E and Hove L (1996) Experience with Calliandra calothyrsus as feed for Livestock in Africa. In: Evans DO (eds) International Workshop on the Genus Calliandra, pp 195–289. Proceedings of a workshop held in Bogor, Indonesia. 23–27 January 1996. A Winrock International Publication Potter HL (1987) Inventory of feed resources for the smallholder farmer in Kenya, pp 2–22. Proceedings of the second PANESA Workshop held in Nairobi, Kenya. 11–15 November 1985. IDRC. Nairobi, Kenya Sanda I, Romney DL, Tanner JC, Thorne P and Leaver JD (1999) Effect of abrupt and frequent changes in forage quality on digestibility and performance of crossbred cattle offered napier grass (Pennisetum purpureum) and barley straw, p 88. In: Proceedings of the British Society of Animal Science SAS (1985) Proprietary Software release 6.02 (SAS User guide statistics). SAS Institute Inc. Cary, NC, USA Taiganides EP (1987) Animal waste management and wastewater treatment. In: Strauch D (ed) Animal Production and Environmental Health, pp 91–151. Elsevier Science Publishers, Oxford 331 Thijseen HJC, Murithi FM, Nyaata OZ, Mwangi JN, Aiyelaagbe IOO and Mugendi DN (1993) Report on an ethnobotanical survey of woody perennials in the coffee zone of Embu District, Kenya. AFRENA Report No. 62. International Centre for Research in Agroforestry (ICRAF). Nairobi, Kenya Van Soest PJ (1965) Non nutritive residues: a system of analysis for the replacement of crude fibre. J Assoc Off Chem 50: 50–55