UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of season and shrub-grass combination on the fodder quality of three agroforestry plant species… Mengich, Edward Kibet 1994

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1994-0135.pdf [ 4.47MB ]
JSON: 831-1.0075190.json
JSON-LD: 831-1.0075190-ld.json
RDF/XML (Pretty): 831-1.0075190-rdf.xml
RDF/JSON: 831-1.0075190-rdf.json
Turtle: 831-1.0075190-turtle.txt
N-Triples: 831-1.0075190-rdf-ntriples.txt
Original Record: 831-1.0075190-source.json
Full Text

Full Text

THE EFFECT OF SEASON AND SHRUB-GRASS COMBINATION ON THE FODDER QUALITY OF THREE AGROFORESTRY PLANT SPECIES GROWN IN MASENO, WESTERN KENYA BY EDWARD KIBET MENGICH B.Sc, Moi University, Kenya, 1988 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES (Department of Forestry) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA February 10, 1994 © Edward Kibet Mengich, 1994 In presentin g thi s thesi s i n partia l fulfilmen t o f th e requirement s fo r a n advance d degree a t th e Universit y o f Britis h Columbia , I  agre e tha t th e Librar y shal l mak e i t freely availabl e fo r referenc e an d study . I  furthe r agre e tha t permissio n fo r extensiv e copying o f thi s thesi s fo r scholarl y purpose s ma y b e grante d b y th e hea d o f m y department o r b y hi s o r he r representatives . I t i s understoo d tha t copyin g o r publication o f thi s thesi s fo r financia l gai n shal l no t b e allowe d withou t m y writte n permission. Department o f P i O c f i ^ The Universit y o f Britis h Columbi a Vancouver, Canad a Date I*  Izhl-DE-6 (2788 ) ABSTRACT ii An experimen t t o stud y th e effec t o f seaso n an d shrub-gras s combinations o n th e fodde r qualit y o f leucaen a (Leucaen a leucocephala), calliandra (Calliandra calothvrsus) and Napier grass (Pennisetum purpureum) was established at Maseno, western Kenya. The specie s wer e manage d a s hedgerow s o n fiel d bund s i n a randomized complete block design with seven treatments and four replications. Fresh and dry leafy biomass assessments, and sample collection wer e done a t two-month harvestin g interval s fo r 18 months. Percen t dr y matte r wa s determine d b y oven-dryin g approximately 50 0 g of fresh samples at 60 °C for 48-72 hours. Dried samples were ground to pass a 1 mm sieve and analysed for N (used for crude protein estimation), P, Ca, K, Mg, Zn, Cu and ADF (acid detergent fibre). Statistical analysis was done using SAS 6.04 at a=0.05 significance level. Napier grass was highest in fresh and dry biomass productivity. Biomass productivity, however, dropped significantly in the second year. Biomass productivity of shrubs was lower, but was maintained at similar levels throughout the study period. Leucaena was highest in crude protein, Ca and Cu, but lowest in Zn and ADF. Calliandra was highest in P, Zn and ADF, but was lowest in K and Mg. Napier grass was highest in K and Mg, but was lowest in crude protein, Ca, P and Cu. iii Except in the leucaena-Napier grass mixture, where differences were not significant, establishment in shrub-grass combinations caused significant increase s i n th e biomas s yield s o f Napie r grass. Biomass yield s o f th e wood y perennial s wer e eithe r increase d significantly or were not affected. Nutrient concentrations of the legume plant s wer e no t significantl y change d b y shrub-gras s combinations. Th e sam e i s tru e o f th e Napie r gras s fo r most nutrients except K and Mg. The former was increased significantly in combinatio n wit h bot h legumes , whil e th e latte r wa s significantly reduced in combination with calliandra, but remained unchanged in combination with leucaena. Except fo r K  conten t i n Napie r gras s (r=0.750) , biomas s an d nutrients were not significantly correlated with rainfall. Other correlations wer e no t significan t an d varied wit h specie s and parameter. It is suggested tha t the presence of at least some rainfall in all months maintained a reasonable level of moisture in the soil, so that adverse effects caused by prolonged drought in other areas were not observed at Maseno. Napier grass is suitable for providing the basic ration, while trees and shrubs have a significant potential as high nutrient supplements to conventional animal feeds. These can conveniently be established a s tre e o r shrub-gras s combinations . Attempt s t o diversify the genetic base of fodder trees and shrubs should be made to overcome problems related to toxicity, poor digestibility iv and the occurrence of pests and diseases. V TABLE OF CONTENTS Page ABSTRACT i i TABLE OF CONTENTS V LIST OF TABLES vii i LIST OF FIGURES i x LIST OF ANNEXES xi v ACKNOWLEDGEMENTS xvi i 1. INTRODUCTION 1 2. LITERATURE REVIEW 5 2.1. Agroforestry: concepts and potentials 5 2.1.1. An  overview 5 2.1.2. Agroforestry  in  Kenya  7 2.2. The role of woody perennials and their contributions in agroforestry 9 2.3. Woody plants for fodder in animal agroforestry .. 1 1 2.3.1. Rationale 1 1 2.3.2. The  role of  nutrients  1 3 2.3.3. Species 1 6 2.3.4. Benefits  to  livestock  1 8 2.4. Management for fodder 2 5 2.4.1. Grasses 2 5 2.4.2. Trees  and  shrubs 2 6 2.4.3. Alley  farming  3 0 vi 2.5. Limiting factors 3 1 3. MATERIALS AND METHODS 3 4 3.1. Study area 3 4 3.2. Site, experimental design and treatments 3 7 3.3. Plant species 3 8 3.4. Establishment and management 4 0 3.5. Biomass measurements, sample collection and weather records 4 1 3.6. Analytical procedures 4 2 3.6.1. Nitrogen  and  minerals  4 2 3.6.2. Acid  detergent  fibre  (ADF)  4 3 3.7. Statistical procedures 4 4 4. RESULTS 4 6 4.1. Leaf/stem ratios 4 6 4.2. Leafy biomass yields 4 8 4.2.1. Fresh  biomass  yields  4 8 4.2.2. Dry  matter  yields  5 1 4.3. Nutrient concentrations 5 4 4.3.1. General  5 4 4.3.2. Percent  dry  matter  (DMh)  5 8 4.3.3. Crude  protein  6 1 4.3.4. Calcium  6 4 4.3.5. Potassium  6 4 4.3.6. Magnesium  6 9 4.3.7. Phosphorus  7 2 4.3.8. Zinc  7 2 4.3.9. Copper 7 7 4.3.10. Acid  detergent  fibre  7 7 DISCUSSION 8 3 5.1. Leaf/stem ratios 8 3 5.2. Leafy biomass yields 8 4 5.3. Nutrient concentrations 8 7 5.3.1. Comparative  information from  literature . . 8 7 5.3.2. Percent  dry  matter  8 9 5.3.3. Crude  protein 8 9 5.3.4. Minerals 9 1 5.3.5. Acid  detergent  fibre  9 4 SUMMARY AND CONCLUSIONS 9 8 REFERENCES 10 2 APPENDICES 11 4 viii LIST OF TABLES Page Table 1. Average leaf/stem ratios for L. leucocephala and C. calothyrsus over 18 months 4 7 Table 2. Mean fresh and dry leafy biomass yields (kg/lOOm) of L. leucocephala. C. calothyrsus and P. purpureum over 18 months 4 9 Table 3. Comparative nutrient compositions of L. leucocephala. C. calothyrsus and P. purpureum over 18 months 5 5 Table 4. Spearman rank order coefficients between rainfall and nutrients in L. leucocephala. C. calothyrsus and P. purpureum 5 7 Table 5. The nutrient composition of L. leucocephala. C. calothyrsus and P. purpureum as found in some literature 8 8 ix LIST OF FIGURES Page Fig. 1. Average monthly rainfall and temperatures for Maseno research station (May /91-Nov.'92) 3 5 Fig. 2. Fresh leafy biomass yields of L. leucocephala, C. calothvrsus and P. purpureum over 18 months of repeated cutting 5 0 Fig. 3. Dry leafy biomass yields of L. leucocephala. C. calothyrsus and P. purpureum over 18 months of repeated cutting 5 2 Fig. 4. Variation in the dry matter content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothvrsus over 18 months 5 9 Fig. 5. Variation in the dry matter contents of L. leucocephala and C. calothvrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 6 0 X Fig. 6. Variation in the crude protein content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 6 2 Fig. 7. Variation in the crude protein contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 6 3 Fig. 8. Variation in the calcium content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 6 5 Fig. 9. Variation in the calcium contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over over 18 months 6 6 XI Fig. 10. Variation in the potassium content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 6 7 Fig. 11. Variation in the potassium contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 6 8 Fig. 12. Variation in the magnesium content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 7 0 Fig. 13. Variation in the magnesium contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 7 1 Xll Fig. 14. Variation in the phosphorus content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 7 3 Fig. 15. Variation in the phosphorus contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 7 4 Fig. 16. Variation in the zinc content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 7 5 Fig. 17. Variation in the zinc contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 7 6 Xlll Fig. 18. Variation in the Copper content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 7 8 Fig. 19. Variation in the copper contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 7 9 Fig. 20. Variation in the ADF content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 8 0 Fig. 21. Variation in the ADF contents of L. leucocephala and C. calothyrsus when grown in pure arragements and in combination with P. purpureum over 18 months 8 1 XIV LIST OF APPENDICES Page Appendix 1. Total rainfall received at harvesting and field data collection dates 11 4 Appendix 2. Layout of the experiment 11 5 Appendix 3. Dry matter and nutrient composition of L. leucocephala at different harvesting dates 11 6 Appendix 4. Dry matter and nutrient composition of C. calothyrsus at different harvesting dates 11 7 Appendix 5. Dry matter and nutrient composition of P. purpureum at different harvesting dates 11 8 Appendix 6. Recommended nutrient content of diets for dairy cattle 11 9 XV Appendix 7. Analysis of variance table for fresh leafy biomass production of L. leucocephala. C. calothvrsus. P. purpureum and their tree-grass combinations 12 0 Appendix 8. Analysis of variance table for dry leafy biomass production of L. leucocephala. C. calothyrsus. P. purpureum and their tree-grass combinations 12 0 Appendix 9. Results of the separation of means between pure and mixed treatments when the F values were significant 12 1 Appendix 10. A short description of the soil types common to western Kenya 12 6 Appendix 11. Spearman rank order coefficients between different leafy biomass and nutrients in L. leucocephala 12 7 Appendix 12. Spearman rank order coefficients between different leafy biomass and nutrients in C. calothyrsus 12 8 xvi Appendix 13. Spearman rank order coefficients between different leafy biomass and nutrients in P. purpureum 12 9 xvii ACKNOWLEDGEMENTS The accomplishmen t o f my studie s at the University o f British Columbia and field research at Maseno would not have been a reality without the support provided by a great number of individuals and institutions. I a m indebte d t o my researc h superviso r Dr . T.M. Ballar d fo r allowing me to use his laboratory facilities and for his excellent supervision, advice, counselling and guidance to the success of this programme. My advisory committee members, Drs. J.P. Kimmins, P.L. Marshall , H . Schreie r an d J . A. Shelford , constantl y an d willingly took time off their busy schedules in order to provide advice on both course work and thesis research. Dr. Mathew Koshy introduced me to the use of SAS when doing statistical analysis. Mr. Harr y Otien o (KEFRI' s Directo r o f th e Masen o Nationa l Agroforestry research centre), Mr. D.A. Hoekstra (ICRAF' s AFRENA East and Central Africa co-ordinator) and Dr. D. Nyamai (KEFRI's AFRENA National co-ordinator) provided technical supervision of my research while at the AFRENA site in Maseno. On-station scientists at Maseno allowed me to use some of their bio-physical data. The AFRENA on-station technicians M. Odongo, H. Wandabwa, Onesmus Karuno and H.M. Ngethe took care of the experiment at Maseno and collected and mailed all specimens and field data. xviii The assistance provided by Mr. Bernie Von Spindler (Soil Science) and Ms . Maureen Evan s (Anima l Science ) whe n doin g laborator y analysis at UBC is highly appreciated. I thank the Kenya government (through KEFRI) for granting a leave of absenc e durin g th e stud y period . ICRA F coordinate d th e procedures leading to my admission to the University of British Columbia. The International Development Research Centre (IDRC) of Canada provided funds for my studies at UBC. Part of my research activities (esp . those carried out at Maseno) were funded by the Kenya Governmen t an d ICRA F throug h th e KEFRI-KARI-ICRA F agroforestry research project. 1. INTRODUCTION 1 Livestock production is an integral part of all landuse systems in the highlands of western Kenya. Population densities in this part of the country have increased significantly since early in this century. Land holdings ar e now too small to provide all the needs of individual households. This has caused considerable changes in the traditional lifestyles of the local people (Conelly , 1992). Among other things, there is a general tendency to put more land under crops, leaving little or none for free grazing. Farmers have shifted their livestock management practices from tethering and herding o n communal an d fallo w land to zero grazing, the most intensive livestock production system, involving the cut-and-carry method. Zero grazing with an average of 2-3 animals is currently being practised by many of the farmers. The manure from the animals is utilized to improve soil fertility in crop plots and the crop residues (maize stover, sweet potato (Ipomoea batatas) vines, stem and foliage of banana plants, sugarcane tops and bean leaves) are used a s fee d supplement s durin g period s o f diminishe d fodde r supply. Crop residues, however, are low in digestibilities and are deficient i n fermentabl e energy , fermentabl e nitrogen , crud e protein and several of the required mineral elements (Nangol e et al. . 1983; Butterworth, et al. , 1984; Escobar and Parra, 1980, cited by Preston, 1982). 2 Napier grass (Pennisetu m purpureum) , a popular fodder species in the high-potential areas of Kenya (Abate et al. 1985), is the major livestock fee d supplyin g th e zero-grazin g units . Together with maize (Zea mays) stover, it accounted for 80% of all cut-and-carry feeds in Hamisi during the 1986-87 period (Conelly , 1992). Many farmers have established Napier grass either on terrace bunds (low-lying ridges of soil constructed along the contour) or as fodder banks. Napier grass is high in biomass production per unit area, but like many tropical pasture grasses, has a  few shortcomings: l) the protein content of the grass is not high (Gohl, 1981) and drops significantly durin g the dry season (Sand s et al., unpublished data, 1982), 2) during the dry season, availability of grass fodder is reduced because of a slow-down in growth and an increase in the crude fibre content, 3) biomass production from the grass starts to decline over time unless nutrients are added (Snyders, 1991). Resources to purchase high protein concentrates such as bone meal, and meat-and-bone meal (Butterworth et al., 1984) for supplementing livestock feeds are limited, and the majority of the farmers find these feeds expensive and outside their financial capabilities. Kenya, like many other third world countries, is looking for a cheaper source of protein (Semenye, 1990). Herbaceous legumes such as desmodiu m an d lucern e (Snyders , 1991 ) coul d improv e th e situation, but do not have the capacity to withstand periods of extreme moisture stress. Tropical leguminous woody forages are rich in protei n an d mineral s (Gohl , 1981 ) an d ma y b e possibl e alternatives. 3 The dry matter yield, protein quality and in-vitro digestibility of Leucaena leucocephala. for example, compare well with those of the finest forag e legume s suc h as alfalf a (NAS , 1977). Report s by Siebert e t al . (1976) , i n which steer s fe d choppe d sugarcan e (Saccharum officinarum ) supplemente d wit h Leucaen a leucocephal a gave the same liveweight gains (0.61 kg per head per day) as those supplemented wit h meat meal, support s this hypothesis. Studies involving th e us e o f multipurpos e tree s an d shrub s (MPTS ) a s livestock feed supplements have yielded similar results. Consequently, agroforestry research focusing on the identification of potential fodder species and suitable management options has been intensified . Du e to the lan d tenur e situatio n i n western Kenya, and to reduce labour requirements, fodder production systems in which component s are established a s tree-grass mixtures are prefered to pure planting arrangements. Earlier research has shown that Napier grass and some leguminous shrubs increase significantly in biomass production when grown in mixtures (Gil l et al., 1990; Otieno et al., 1991). However, information regarding the effect of these mixed plantings on the feeding value of the species involved is not available. The effect of seasonal changes on the nutrient composition (particularly of the shrub species) has also not been investigated. 4 The objectives of this study were to: 1) compare the fresh leafy biomass, dry matter, crude protein, acid detergent fibre (ADF) and macronutrient element contents of Leucaena leucocephala, Calliandra calothyrsus and Pennisetum purpureum grown in western Kenya; 2) study the fluctuations in the nutrient values of the three fodder crops with season; 3) study the nutrient content changes when the species are established as tree-grass mixtures or as pure stands. 5 2. LITERATURE REVIEW 2.1. Agroforestry: concepts and potentials 2.1.1. An overview Agroforestry is a new term for an old landuse system that has been in existence for many years (King , 1987). A preliminary overview for a  globa l inventor y o n agroforestr y system s (ICRAF , 1983 ) revealed man y example s o f prominen t agroforestr y system s an d practices in developing countries. Agroforestry permits multiple cropping in which woody perennials are integrated with crops and/or animals/pastures o n the sam e unit of lan d where the y interac t ecologically an d economicall y (Somarriba , 1992 ; Agroforestr y systems. vol. 1, pp. 7-12, 1982). Agroforestry an d it s potential s attracte d th e interes t o f scientists and development planners in late 1960's and the 1970's when increased population densities called for immediate remedies to pressing landuse problems. It has been pointed out that the increase in world population during the 20 years between 1965 and 1985 was equal to the total world population in 1900 and that 90% of this increase was in developing countries (Kidd and Pimentel, 1992). Despite declining birth rates in many developing countries, this population (about 4.8 billion in 1985) is estimated to increase by 37% to about 6.6 billion by about 2005. 6 In tacklin g problem s associate d wit h reduce d lan d size s an d diminished per capita income, agroforestry is generally seen as a promising option, albeit a complex enterprise requiring knowledge and skills of experts and scientists from a diversity of academic disciplines (forestry, agriculture , soil science, animal science, etc.). The establishment o f the International Counci l fo r Research in Agroforestry (ICRAF ) in 1977 to promote and catalyse research in agroforestry (King , 1977 ) wa s a  mileston e i n th e histor y o f Agroforestry. After joining the Consultative Group on International Agricultural Researc h (CGIAR) , wit h globa l responsibilit y fo r research on agroforestry, ICRAF recently changed from a council to a centre . Wit h it s headquarter s i n Nairobi , Kenya , ICRAF' S institutional goals are to mitigate tropical deforestation, land depletion and rural poverty through improved agroforestry systems (ICRAF, 1991). The success of any agroforestry practice depends on the choice of plant species . Huxle y (1981a ) provide d a  lis t o f technical , managerial and socio-economic factors to be considered when trees (and other woody perennials) are being considered for agroforestry. Many leguminou s wood y perennial s hav e a n advantag e ove r othe r multipurpose trees and shrubs (MPTS) because of their ability to fix atmospheri c nitrogen . Consequently , a  literatur e surve y conducted by ICRAF prior to a global inventor y o f agroforestry 7 systems indicated that the majority of the species currently used in tropica l an d subtropica l agroforestr y belon g t o th e famil y Leguminosae [Fabaceae ] (Nai r e t al. , 1984) . Nitroge n fixatio n improves the nitrogen status of the soil and increases the yield of associated crops (Dommergues, 1987). 2.1.2. Agroforestry in Kenya Kenya is a developing country lying on the equator and having a land are a o f about 580,00 0 sq . km. A larg e proportion o f the country i s dominated b y arid an d semi-ari d climate . The high-potential area which consists of the Kenya highlands, the coastal strip and much of the Lake Victoria Basin occupies less than 20% of the are a (Getahun , 1989) , bu t support s th e majorit y o f th e country's 25 million people. In recent years, land use problems associated with increasing population densities (an d subsequent intensification of agriculture and livestock production) have been experienced, particularly in the medium-to high-potential zones. The period between 1970 and now (1994) has been one of accelerated agroforestry awareness, research and development in Kenya. This achievement has been attributed to the joint efforts of the Kenya government, donor agencies, the non-governmental (NGO ) community and the Kenyan farmers (Getahun, 1989). As Getahun further stated, the Keny a government' s initia l attentio n an d interes t i n agroforestry cam e a s a  resul t o f energ y suppl y problem s an d environmental degradation. 8 In 1981 , the the n Ministry o f Energ y an d Regiona l Developmen t (MOERD), in close collaboration with the Ministry of Agriculture (MOA), the Ministry of Environment and Natural Resources (MENR) , the Ministry of Livestock Development (MOLD ) and selected NGO'S such as Kenya Energy Non-Governmental Organizations (KENGO), began a nationa l effor t t o implemen t a  coordinate d programm e o f agroforestry research, training and extension (Getahun, 1989). By 1985, ther e were 13 major national organizations and 63 others active in agroforestry/social forestry and general tree planting activities i n Keny a an d thi s numbe r ha s bee n o n th e increas e (Energy Development International, 1985; cited by Getahun, 1989). The locatio n o f th e Internationa l Centr e fo r Researc h i n Agroforestry (ICRAF) in Nairobi has made a significant contribution to Kenya's progress in agroforestry. ICRAF's 40-ha field station at Machakos provide s facilitie s fo r fiel d experimentatio n an d demonstration (ICRAF , 1991). Since its creation in 1986, ICRAF's Collaborative Programmes Division (COLLPRO) , has conducted joint field research vis a vis the agroforestry researc h network for Africa (AFRENA) with the Kenya Forestry Research Institute (KEFRI) and the Kenya Agricultural Research Institute (KARI) at Maseno and Embu. AFRENA is a programme established by ICRAF and is mandated to develop appropriate agroforestry technologies for selected land use systems and to develop the regional and national capability to 9 plan, formulat e an d implemen t agroforestr y researc h i n th e participating African countries and regions. The programme covers, among other African regions, the highlands of East and Central Africa. ICRAF continues to play a leading role in coordinating, catalysing, promoting and encouraging agroforestry research, training and/or extension among local institutions. It also participates actively in the quest fo r appropriate agroforestry technologie s fo r the farming community . Agroforestr y i s no w taugh t a s par t o f th e curriculum i n som e o f th e universitie s an d agricultura l institutions. Agroforestr y researc h i s undertake n a t Mo i University, Universit y o f Nairob i an d Egerto n University . Experiences fro m agroforestry extensio n agencies indicate great interest an d enthusias m amon g small-scal e farmers . Impac t evaluation of the CARE-KEFRI extension project in the Siaya and South Nyanz a district s showe d considerabl e increase s i n tre e planting activities among the target groups (Scherr, 1992; Scherr and Alitsi, unpublished data, 1991). 2.2. The role of woody perennials and their configurations in agroforestry Depending on the intended principal end-uses, farmers integrate trees and shrubs with their crops and livestock in various ways (Burley, 1987) . Individua l woody plants may occu r regularly or 10 randomly at wide spacing on productive agricultural land. They may also be found as linear single or multiple-row plantings along boundaries, contours, roads, riverbanks o r railways. Trees and shrubs may be established as shelterbelts, windbreaks, or as live fences. Community and farm woodlots are common features in some areas. In many highly populated humid and subhumid tropics, woody perennials are established as hedgerows in small scale farms, where they are maintained at low cutting heights. In all these configurations, the products and services derived from woody perennials are manifold and include direct advantages and both environmenta l an d socio-economi c benefit s (Burley , 1987) . Extensive review s concernin g th e rol e o f wood y perennial s i n agroforestry have been undertaken (Huxley , 1981b; Nair et al.. 1984). Thes e role s ca n b e categorize d int o productiv e an d protective (Nai r e t al. , 1984) ; som e specie s pla y bot h role s simultaneously. Tree s an d shrub s ar e productive a s source s of consumables such as food, fodder, firewood, timber, green manure, fruits an d medicine . Th e protectiv e (service ) rol e o f wood y perennials in agroforestry stems from their soil improving and soil conserving functions, and their role as live fences, shelterbelts and windbreaks. Woody perennials improve and enrich soil conditions through nitroge n fixation , additio n o f organi c matter , an d improvement of soil structure and efficiency of nutrient cycling (Young, 1989). Physical soil conservation is the main protective function of woody perennials and can be conveniently exploited in 11 agroforestry if the chosen species can provide additional benefits and outputs such as fodder, fuel, etc. (Nair et al., 1984) . A large number of multipurpose woody perennials are being used as effective live fences at CATIE (Centro Agronomico Tropico de Investigacion y Ensenanza), Turrialba, Costa Rica (Budowski, 1983). Similarly, very encouraging result s o n shelterbelt s an d windbreak s hav e bee n obtained at the Pakistani Forestry Research Institute, Peshawar (Sheikh and Khalique, 1982). In animal agroforestry, fodder production is one of many productive and service roles played by woody perennials in systems supporting either or both domestic and wild herbivores (Torres, 1983). Browse in silvopastoral systems provides stability and productivity of livestock production, the major source of livelihood and income in arid and semi-arid African zones (Le Houerou, 1987). 2.3. Woody plants for fodder in agroforestry 2.3.1. Rationale Grasses for m th e mai n portio n o f tropica l an d sub-tropica l pastures. Thes e pasture s suppor t th e majorit y o f th e world' s herbivores: 66% of the cattle, 64% of the sheep and goats, 80% of the equines, and almost all the camels and the buffaloes (Jones, 1988). Individua l grasse s diffe r i n growt h habits , ecologica l requirements an d utilizatio n (Judd , 1979) . Whe n thei r fodde r 12 potentials are considered, grasses are generally highly productive and nutritiou s unde r favourabl e conditions ; bu t ar e highl y susceptible to environmental changes. There ar e severa l factor s tha t hav e bee n foun d t o limi t th e potentials of tropical fodder grasses. Grasses are generally low in protein and minerals (Gohl, 1981). Compared to temperate grasses, tropical grasses have lower feeding values because of lower protein levels which sometime s fal l below 6-8 % (depressin g dr y matter intake) and higher fibre contents, which lower voluntary intake and dry matter digestibility (Minson, 1981) . The lower digestibility of tropical grasses is caused by differences in anatomical structure associated with the different photosynthetic pathways (Laetsch , 1974) and the higher temperature at which tropical grasses are normally grown (Minson , 1990). Thes e feature s limit long-term animal performance to 0.7 kg liveweight gain per head per day for beef cattle , or 1 2 kg milk per head per day fo r dairy cattle (Humphreys, 1991) . Fluctuation s i n productivit y an d nutritiv e values associate d wit h environmenta l an d growt h change s ar e noteworthy. While grasses attain high productivity during periods of moisture sufficiency, there is a significant drop in quality (Sands et al., unpublished data, 1982) and quantity during the dry season. Animal feed requirements on the other hand remain relatively constant and must be met continuously. 13 Leguminous trees and shrubs are high in protein and minerals and are able to withstand adverse environmental conditions compared to grasses. Le Houerou (1980 ) found grasses in the dry season to be extremely deficien t i n protein, phosphorus an d carotene , while these nutrient s wer e adequat e fo r livestoc k maintenanc e requirements in a wide range of browse plants. Unlike grasses, the dry matter digestibilities o f tropical and temperate legumes are similar because both categories have the same photosynthetic pathwa y an d lea f anatom y (Minson , 1990) . Wood y perennials can provide alternative feeds during the dry season when grasses ar e scarc e o r absent . I n th e fac e o f hig h populatio n densities, agroforestry systems in which suitable woody perennials are managed along side crops and/or animals/pastures to serve this purpose are, therefore, being promoted among many small-scale farms in the tropics. 2.3.2. The role of nutrients A nutrient is a feed constituent, or group of feed constituents that are classified together, which contribute to the support of animal life. All nutrients in feeds are contained within the dry matter and include crude protein, energy, minerals, and vitamins. The quantity of each nutrient absorbed depends on (1) the quantity of forage dry matter eaten each day, and (2) the concentration and availability of that particular nutrient in each kilogram of forage 14 dry matter (DM) (Minson, 1990). Detailed information on individual nutrients and their functions in the animal body are provided in many standard texts. The following summary is an overview condensed from Huston and Pinchak (1991). Crude protein Crude protein is the basic structural material fro m which many animal bod y tissue s ar e forme d (e.g . muscles , nerves , skin , connective tissue s an d bloo d cells) . I t supplie s nitroge n (a s ammonium) and amino acids for intraruminal microbial activity and cellular-level tissu e metabolism . Twenty-fou r amin o acid s ar e generally thought to be constituents of proteins; some of these are "essential" because they cannot be formed in body tissues, and must be provide d i n th e die t (Pomeran z an d Meloan , 1987) . Protei n requirements i n ruminant s includ e protei n and/o r nitroge n requirements o f th e rumina l microbia l population . Generally , microbial requirements are met at 6-8% crude protein in the diet. Animal requirements range from 7-20% in the diet depending upon species, sex and physiologic state. Energy Energy is required primarily in making (anabolism), and sometimes breaking (catabolism), chemical bonds in metabolic processes which include muscle contraction, nerve impulses, and tissue synthesis. 15 Ruminants derive energy primarily from plant carbohydrates, lipids and proteins, though not all of it is captured in a form usable to the animal. The energy value of feeds and forages can be expressed in many ways. Net energy is the amount of energy available for maintenance and production. Minerals Minerals are required for tissue growth, repair, and the regulation of body functions. Twenty-two elements are important for animal nutrition an d ar e normall y categorize d int o tw o groups . Those required in relatively large amounts (grams/day) are refered to as "Macro" and include Na, Cl, Ca, P, Mg, K and S. Those required in small amounts are refered to as "Micro" (milligrams/day or less). They include Cu, Zn, Mn, Mo, Se, I, Fe, Co, F, V, Sn, Ni, Cr, Si, As (Little, 1985). Individual minerals have special functions as components of certain tissues or as cofactors for certain metabolic reactions. Vitamins Vitamins are "cofactors" or catalysts in metabolic reactions, in that they do not appear in the products of reactions, but must be present for reactions to occur. All vitamins or their precursors must b e absorbe d fro m th e digestiv e trac t a s the y canno t b e synthesized by mammalian tissue. With few exceptions, vitamin A is 16 the onl y vitamin tha t i s likel y t o limi t th e productivit y o f grazing ruminants. Vitamin A does not occur in plant tissue, but is synthesized b y th e anima l fro m chemica l precursor s i n plants, mainly beta carotene, but other plant pigments as well. Vitamin A deficiency is most likely to develop during an extended period of low temperature and/or drought when green plants are unavailable to the animal. 2.3.3. Species Not all woody perennials available in the tropics and subtropics are suitable for fodder (see section 2.5). A number of species have been identified and occur naturally or are widely cultivated in many ecological zones . A few of these species, the majority of which are legumes, have proved successful and have been introduced for wider adaptation throughout the tropics. Some introduced tree legume species such as Gliricidia sepium. Leucaena leucocephala. Sesbania grandiflora. Albizia falcataria and Calliandra calothyrsus have show n th e greates t potentia l a s a  forag e resourc e fo r increased animal production in Indonesia (Panjaitan, 1988). In attempts to widen the genetic base of browse, recent research has indicated the existence of many more potential fodder trees and shrubs (especially those indigenous to particular regions) that are traditionally used by local communities. Comparative studies into the fodder potentials of indigenous as opposed to exotic species 17 are being carried out to address potential problems that would result from continued dependence on the few known species. Out of eight indigenous and two exotic species evaluated for browse in southeastern Nigeria, mbom (Alchorne a cordifolia) was highest in dry matter production, while Leucaena leucocephala and Glyphaea brevis were the most prefered by goats (Larbi et al., 1993b). Though cultivated browse species are mainly members of the family Leguminosae, experience of cattle owners and researchers has shown that som e non-legume s ma y hav e tremendou s potentia l a s well . Observations of browsing cattle and interviews with pastoralists revealed that 39 plant species identified a s browse in central Nigeria alone belonged to several different plant familie s and varied greatly in appearance (Bayer , 1990). They included semi-woody legumes such as Adenodolichos paniculatus. twiners such as Mucuna poqgei. large trees such as the savanna mahogany (Khay a seneqalensis) and even a member of the graminae family (a bamboo species: Oxytenanthera abyssinica). Some species such as Tamarindus indica and Faidherbia albida were observed closer to the humid areas, but are usually more prominent in the drier parts of Africa. Since there are many species of fodder trees and shrubs, it is impossible to list all of them, and only examples are given here. In the Guinea and Sudan zones of Northern Nigeria, nutritious pods and seeds of the locust bean trees, Parkia clappertonia and Parkia filicoidea are utilized as livestock fodder and human food in the 18 dry season when other supplies are scarce in the region (Ichire, 1993) . Indigenous vegetation providin g foliag e o r frui t to supplemen t intake from pasture in Australia include mulga (Acaci a aneura), saltbrush (Atriple x spp. ) an d kurrajon g (Brachychito n spp. ) (Pratchett, 1989) . Usefu l introduce d specie s includ e tagasast e (Chamaecytisus palmensis), paulowni a (Paulowni a spp.) , popla r (Populus spp.) and sub-tropical shrubs such as leucaena (Leucaena leucocephala), caro b (Ceratoni a siliqua ) an d hone y locus t (Gleditsia triacanthos ) (Wilson , 1969) . O f these species , only leucaena (Leucaen a leucocephala ) an d tagasast e (Chamaecvtisu s palmensis) hav e achieve d an y degre e o f commercia l acceptanc e (Lefroy et al.. 1992) . Agroforestry with red calliandra (Calliandra calothyrsus) i n Bogor , Indonesi a provide s favourabl e economi c opportunities, particularly as a source of fuelwood, feed for bee keeping and cattle fodder (Satjapradja and Sukandi, 1981). 2.3.4. Benefits to livestock Trees and shrubs form a significant portion of the livestock diets in many tropical and subtropical zones. This is more so in arid and semi-arid areas compared to humid and subhumid climates, where dry seasons are relatively shorter and viable alternatives to browse are availabl e (L e Houerou , 1987) . Tree s an d shrub s i n th e silvopastoral production systems of Africa constitute the basic 19 feed resource of more than 500 out of 660 million head of livestock (FAO, 1985 ; cited by L e Houerou, 1987) . When carefully fe d to livestock, tree fodder has been shown to improve feed quality, availability, intake, digestibility, and animal production per se (Lefroy et al.. 1992; Kang e t al.. 1991). Quality and availability of feed resource In addition to their relatively low nutritional status, grasses are adversely affecte d b y drough t an d ar e no t availabl e durin g prolonged dry seasons. Trees and shrubs explore deep soil layers for moisture and nutrients and can better withstand dry periods. Some of these species, the legumes in particular, are excellent sources of high quality gree n fodder during periods of nutritional stress. In the arid and semi-arid silvopastoral systems of Africa, woody specie s ar e th e onl y sourc e o f protein , caroten e an d phosphorus for livestock and wildlife during the long dry season (Le Houerou, 1987). Effect on food intake and digestibility The quantity of net energy absorbed each day is the main factor controlling the growth rate and milk production in ruminants. The intake of net energy is controlled by the quantity of food energy eaten (intake ) , the proportio n o f eac h uni t o f fee d tha t i s digested (digestibility) and the efficiency of utilization of the 20 products of digestion (Minson , 1981). Intak e i s determined by, among other things (sward structure, availability, type of animal, etc.)/ the feed quality. The appetite of animals, for example, is depressed by protein deficiency when the crude protein content of feeds falls below 6-8%, a common feature in tropical grass pastures (Minson, 1981). The availability of energy in forage is usually measured as dry matter digestibilit y (DMD) , whic h i s closely relate d t o othe r energy parameters: organic matter digestibility, digestible organic matter, tota l digestibl e nutrients , digestibl e energy , an d metabolizable energ y (Minson , 1990; Huston an d Pinchak , 1991) . Digestion o f forag e organi c matte r i s dependent , amon g othe r things, on the activities of rumen micro-organisms. These organisms require a  supply o f protein, amino acid s o r their precursors, including nonprotei n nitroge n an d sulfu r (Hungate , 1966) . Phosphorus, in addition to being important for skeletal formation, is also essential for proper functioning of rumen micro-organisms, especially those which digest plant cellulose (McDowel l et al., 1983). Deficiency of these substances may limit digestibility. Apart from chemical treatment of forage (e.g., with NaOH, ammonia or Magad i soda) , an d plan t breeding , voluntar y intak e an d digestibility can be improved cheaply and safely by including small quantities of high protein legumes as feed supplements (Minson , 1990). Rees et al. (1974) reported large increases in the appetite 21 of sheep when as little as 10% legume was added to a grass diet containing 3.6% protein. At ILCA in Addis Ababa (Butterworth and Mosi, 1985), average dry matter digestibility of the crop residue (oat, wheat, teff and maize straws) when fed alone was 48.3% and the intak e wa s 49. 8 g/(k g liveweight) 0-75. Th e provisio n o f an average o f 45 % legum e (Trifoliu m tembense ) ha y improve d digestibility to 65% and intake by 16% to 58 g/(kg liveweight)0-75. Digestibility of nitrogen and energy were improved in all cases, while th e cel l wal l constituents , ADF , NDF , cellulos e an d hemicellulose, were each improved i n at least one type of crop residue. Where both the grass and the legume component of a mixture contains sufficient crude protein and minerals, voluntary intake is linearly related to the proportion of legume in the mixture and there is no synergism between the two species (Moseley, 1974; cited by Minson, 1990) . Synergis m will occu r where one species i n a mixture is deficient in an essential nutrient and the other species contains a high level of this nutrient (Minson, 1990). The effect s o f legum e supplementatio n o n forag e intak e an d digestibility hav e als o bee n demonstrate d b y othe r researc h workers. Supplementing a basic diet of Pennisetum purpnreum with increasing level s o f Erythrin a abyssinic a leave s resulte d i n reduced intake of the grass, but increased total organic matter intake in both species (Larbi et al.. 1993a). Phiri et al. (1992) observed significant increases in total dry matter intake, diet dry matter digestibility, and diet organic matter digestibility when 22 maize husks were fed together with individual and mixed supplements of leucaen a an d calliandra . Bamuali m e t al . (1984 ) reporte d significantly greater organic matter intake when spear grass was supplemented wit h fres h leucaen a leaf . Van Ey s e t al . (1986 ) observed no difference among diets in total dry matter intake, intake of cell wall constituents, and digestibility when foliage from gliricidia, leucaena, and sesbania were used as supplements to Napier grass for growing goats. No significant differences in dry matter intak e wer e observe d whe n cross-bre d heifer s wer e fe d Gliricidia maculata leaves (Dharia et al.f 1987). Effect on production Foliage from multipurpose trees and shrubs can be fed either as sole feed s o r a s supplement s t o conventiona l livestoc k diets, though the forme r approach appears to have been accepted to a lesser extent. The increased animal productivity achieved when tree foliage is included in the diet is likely due, at least in part, to high intake of N and energy resulting from the increased total organic matter intake. Cattle feeding on leucaena near Brisbane in Australia gained an average of almost 1 kg each day for more than 200 day s (NAS , 1979 ) . Live weigh t gain s o f bul l calve s wer e significantly higher in animals fed leucaena ad lib, than in those fed a restricted amount of the forage (Sobale et al., 1978) . In central Jav a (NRC , 1983) , drie d calliandr a leave s ar e ofte n pulverized, pelleted (either alone or mixed with leucaena leaves) 23 and used for feed on Javanese chicken farms, while calliandra leaf meal has successfull y been used as chicken feed in amounts up to 5%. No significant differences in body weight gains were observed when cross-bred heifers were fed Gliricidia maculata leaves, but the cost of feeding was reduced (Dharia et al., 1987). Browse supplementation is popular and widespread among small-scale livestock owners in the tropics. A fee d consisting of Leucaena leucocephala alone caused loss of hair in sheep within 10 days, but combination with Setaria sphacelata grass up to about 60% shrub (DM) cause d n o il l effect s eve n afte r 2  month s (Josh i an d Upadhyaya, 1976) . The suggestion by Joshi and Upadhyaya (1976) that leucaena can be used as animal feed when fed in combination with any ordinary fodder is applicable to some other woody perennials as well. Liveweight gains of sheep and goats increased linearly with increasing level s o f Erythrin a abyssinic a lea f supplementatio n (Larbi et al., 1993a). In sheep trials using mixed diets of grass and calliandra, best growth was obtained with 40-60% calliandra (NRC, 1983). In Zambia, goats feeding on maize husks supplemented with leucaena , calliandr a an d leucaen a plu s calliandr a ha d significantly higher liveweight gains than the control (Phir i et al., 1992). In Natal, goats offered 100% elephant grass lost weight at the rate of 1 9 g/day , bu t th e tren d reverse d wit h increasin g leucaen a intakes, reaching a maximum liveweight gain of 43 g/day when the 24 forage was provided ad lib (Yates and Panggabean, 1985). Flores et al. (1979 ) reported a significant increase in milk production and quality when leucaena was fed to cows selecting a Rhodes-grass (Gloris gayana cv. Pioneer) diet containing 18.2% crude protein. When foliage from gliricidia, leucaena and sesbania were used as supplements to Napier grass for growing goats, average daily weight gain (21 g/day) was significantly higher than for the control (- 1 g/day) (Van Eys et al., 1986). At the Tumbi Research Institute in Tanzania, kids grazing on natural pasture showed higher liveweight gains than the control when provided fodder supplements from each of thre e multipurpos e tre e an d shru b specie s (Caianu s caian , Leucaena leucocephala and Sesbania sesban) (ICRAF, 1991). On the Kenya coast, an increase of 37% in milk production was realized by a farmer who was feeding his dairy cows a mixture of leucaena and Napier grass in his zero grazing unit (Getahu n and Reshid, 1988). Research in Rwanda has demonstrated that the weight gain of stall-fed livestock can be increased 60% (from 14 to 16 g/day to 20 to 30 g/day) by supplementing a ration of grasses with leaves fro m commonl y use d agroforestr y specie s (Nyirahabimana , 1985; cited by Kidd and Pimentel, 1992). A farmer who substituted a mixture of calliandra/leucaena leaves and stems for dairy meal reported no decline in milk production after one week of daily feeding. He was using 2 kg of this mixture together with 4 kg of Napier grass on a Friesian cow (Getahun and Reshid, 1988). 2.4. Management for fodder 25 From th e foregoing , ther e i s no doub t tha t tre e fodder , when provided as supplements to grasses, and sometimes when fed alone, can contribut e favourabl y t o anima l performance . T o sustai n productivity an d ensur e qualit y an d availabilit y o f fodder , appropriate management strategies must be identified and applied. Management options vary with the environment, plant species, growth characteristics, and planting arrangements. 2.4.1. Grasses Free grazin g an d regula r burnin g ar e tw o management practice s applied to maintain productivity and quality of fodder in tropical grass pastures. The high protein levels of Cynodon dactylon and a roadside mixture of Cynodon dactylon and Diaitaria scalarum were attributed to heavy grazing which maintained quality by preventing the production of substantial structural carbohydrates (Sand s et al., unpublished data, 1982). Grazing animal s ar e selectiv e an d usuall y g o fo r th e youn g nutritious plants or younger top portions of the plant. In systems where land use i s intensive and fre e movement o f livestoc k is restricted, these practices are no longer applicable and cut-and-carry practice s ar e more appropriate . As grasse s mature , they become mor e stemm y an d poore r i n qualit y becaus e ther e i s an 26 increase in the proportion of fibre and a decrease in the protein content, voluntary intake, and dry matter digestibility (Minson , 1981). The voluntary intake of Diaitaria decumbens decreased with increased cuttin g interva l (Chenost , 1975) . Th e dr y matte r digestibilities (in vivo) of five species of tropical grasses, each cut as 1-month regrowths, were higher than when cut as mature regrowths (Minson, 1972) . Consequently , the most effective way of maintaining quality is to maintain forage at a young vegetative stage of growth by regularly cutting or grazing at short intervals (Minson, 1990). Grasses readily sprout when cut back and this is usually done at ground level (Otien o et al. . 1991; ICRAF, 1991). There is very little informatio n i n th e literatur e whic h demonstrate s th e productivity of grasses at higher cutting heights; however, it is expected that the practice may increase the presence of injurious and low-quality woody parts in the swards. 2.4.2. Trees and shrubs In some silvopastoral and agrisilvopastoral systems of the arid and semi-arid tropics, trees and shrubs may be left to grow freely to provide seeds and pods for fodder or may be pollarded regularly to provide nutritious green foliage. The demand for browse increases during the dry season, when other forage is scarce, and diminishes during the rainy season. In a study of grazing activity and forage 27 resource use by settled Fulani cattle herders in the subhumid zone of Nigeria, it was found that cattle in a grazing reserve spent more than 10 % of annual feedin g time browsing, with a peak of almost 30% of monthly feeding time in the late dry season (Bayer, 1990) . In more intensive land use systems of the humid and subhumid areas, cultivate d fodde r tree s an d shrub s ar e integrate d int o arable farming systems and may be established as pure stands in pastures, pure hedgerows in agricultural land, or as tree-grass mixtures on hedgerows. Regular coppicing at lower cutting heights is the most common practice. Management practices revolving around cutting heights, frequencies, fertilization, planting densities, etc. are paramount to maintain high productivity, sustainability and quality. Researchers have reported varying forage yield and quality when different cuttin g frequencie s ar e impose d o n fodde r tree s and shrubs. Increased forage yields at lower cutting frequencies and higher percent forage fractions at higher frequencies have been reported for leucaena (Guevarra et al., 1978). Working on subabul (Leucaena leucocephala) and hybrid Napier, Maclaurin et al. (1982) found a harvest interval of six weeks to be significantly better in productivity than four or eight week intervals. Both leucaena and calliandra gave a higher leaf yield when the cutting interval was 12 weeks as compared to 6 weeks (Ell a et al . , 1989). Maximu m forage yield of Leucaena leucocephala was obtained with a 40 days interval compared to 60 and 120 days intervals (Patha k et al.. 28 1980). Reduced productivity due to more frequent cutting has been attributed to the increased number of recovery phases (Kari m et al., 1991). Whiteman and Lulham (1970) suggested that the severe check in growth caused by frequent cutting results in mobilization of sugars and amino acids from the roots to support development of new leaves, thus severely suppressing root formation, and further limiting the production of subsequent foliage. Too frequent cutting (30-day intervals), especially of short stools, is not recommended, not only because o f lowe r dry matter yields, but particularl y because it leads to accelerated depletion of soil nutrients, a decline in vigour with time and the eventual death of the trees (Hill, 1970; Osman, 1980). On the other hand, cutting intervals that are too long may result in higher total N yield, but poor quality forage due to aging (Tangendjaja et al., 1986). At cutting heights not exceeding 50 cm, no significant differences in yield are expected in Leucaena leucocephala (Karim et al.. 1991; Sampet and Pattaro, 1987; Gutteridge, 1988). Above this, however, there is a general tendency for the dry matter yields of Leucaena to increase with cutting height at least up to 1.5 m (Karim et al.. 1991; Turkel and Hatipoglu, 1989; Maclaurin et al.. 1982; Pathak et al., 1980 ; Herrera , 1967 ) excep t fo r a  fe w isolate d case s (Takahashi and Ripperton, 1949). Similar observations have been made for Calliandra calothyrsus (Otieno et al., AFRENA Maseno draft report, 1992) . Thes e observation s ca n b e attribute d t o mor e sprouting space and increased number of branches per tree at higher 29 cutting heights. Pathak et al. , (1980) reported significantly more branches at higher cutting heights than at lower ones. This trend is not clearly maintained above these heights. Catchpoole and Blair (1990) found that leaf production was unaffected by cutting at heights ranging from 1.5 m to 2.5 m above the soil surface. As is true wit h cuttin g interval , N yield s pe r tre e increase s wit h increasing height , bu t overal l nitroge n percentage s ar e no t affected (Karim et al., 1991). Higher plan t densitie s resul t i n highe r tota l production . B y increasing the density of plants from 12,346 to 444,444 plants/ha, significantly greate r productio n o f th e edibl e dr y matte r o f leucaena was obtained (Maclauri n et al., 1982). At the highest density of leucaena (15 x 50 cm) , slightly higher percent forage fractions were obtained compared to lower densities (Guevarra et al.. 1978) . Ella et al. (1989) reported increasing leaf and wood dry matter production u p to 40,00 0 trees/ha, but observe d an d postulated that yields would only increase marginally at higher densities. Thi s i s because a s number o f plants per uni t area increases, competition for moisture, nutrients, space, etc., starts to set in and the potential vigour of individual plants begins to become limited. Though they recorded higher total forage yields at higher plant population densities, Pathak et al. (1980) observed a relatively slight increase in number of branches per plant, which was not significant. 30 Fertilization has a positive effect on productivity and value of forage. Fertilizatio n cause s fas t growt h an d quic k suppl y o f required fodder . Fro m a n experimen t carrie d ou t t o stud y th e productivity o f grass-sesbania intercro p a s compared t o grass-nitrogen interactions , Mathuv a e t al . (1987 ) reporte d tha t fertilizer application gave the highest dry matter yield, while the intercropped grass gave DM above that of pure unfertilized grass. The fast growth also enables harvesting at an early age when the forage i s highl y nutritious . Nitroge n fertilizatio n increase d production to a greater extent with a shorter cutting interval (Maclaurin et al.. 1982). 2.4.3. Alley farming Recently, the International Livestock Centre for Africa (ILCA) has extended the concept of alley cropping to include animal production (Okali and Sumberg, 1985). In this system, termed alley farming, some portion of the tree foliag e i s utilized a s fodder. Alley farming is seen to be more beneficial to the small scale African farmers, compare d t o man y conventiona l agricultura l method s (Cashman, 1986). In order to maintain crop productivity and avoid deterioration in soil fertility, proper allocation of the foliag e is mandatory. Okali an d Sumber g (1985 ) proposed tha t 75 % of availabl e tre e foliage should be applied to the soil as mulch, while 25% should be 31 used a s animal feed . An economic analysi s of alley farmin g in Western Nigeria showe d tha t a t lo w crop yield levels , feeding ruminants with part of the foliage from hedgerow intercropping was profitable (Jabbar et al.. 1992). 2.5. Limiting factors The presenc e o f toxi c substance s (D'Mello , 1982 ; Minso n an d Hegarty, 1984; Onwuka, 1992) and other undesirable properties has led to animal health problems when tree and shrub forages are fed as sole feeds. The effects of toxicity are not apparent in free ranging system s because instinc t regulate s intak e of injuriou s species (Semenye, 1990). Semenye also noted that food selection is limited during drought and in cut-and-carry systems. Care must be taken when browse is being administered to livestock under these conditions. Th e non-protei n amin o acid , mimosine , beta-[N-(3-hydroxy-4-oxopyridyl) ] -alpha-amino propionic acid, occurs naturally in Leucaena leucocephala. Mimosine content in leucaena declines with ag e (Tangendjaj a e t al., 1986). I t i s les s than 1 % (dr y matter) in stems; and ranges from 1 % in old leaves to 9.6% in shoot tips (Jones , 1979; Pratchett et al. , 1991; Lowry et al., 1983; Guevarra et al., 1978; Joshi and Upadhyaya, 1976)). Guevarra e t al . (1978 ) observe d n o significan t difference s i n mimosine content between cultivars, spacings and cutting regimes of Leucaena leucocephala. The toxicity of mimosine has been attributed 32 to its metabolic product, 3-hydroxy-4-(lH) pyridone (DHP), a potent goitrogen which is produced during microbial degradation in the rumen (Hegarty et al., 1976). DHP causes goitre, alopecia, loss of appetite, loss of body weight and sometimes death (Jones , 1979; Hegarty et al., 1964; Jones and Megarrity, 1986; Semenye, 1990). The earlier hypothesis that goats fed on leucaena alone since birth would develop mimosine and DHP-degrading microbes was not supported by research conducted in western Kenya (Semenye, 1990). A bacterium capable of degrading DHP has recently been successfully transfered from Hawaiian goats to Australian ruminants (Jones and Megarrity, 1986) , but has not found its way into Kenya. However, according to Jones (1979 ) and NAS (1977 ) , no health problems are expected if leucaena form s les s tha n 30 % o f th e ruminan t feed . Th e objectionable odou r tha t ma y sometime s occu r i n th e mil k o f leucaena-fed cows disappears on boiling and pasteurizing and can be avoided entirely by eliminating leucaena from the animals' diets for 2 hours before milking (NAS, 1977). No toxic substances have been identified in calliandra (Calliandra calothyrsus). However, the condensed tannin content in calliandra is high (11.07%) compared to tha t in Leucaena leucocephala (4.32%) (Ahn et al., 1989). Tannins are water-soluble phenolic compounds found in legumes and oilseeds. They have molecular weights between 500 and 3000; they give the usual phenolic reactions, but also have special properties such as the ability to precipitate alkaloids, 33 gelatin and other proteins (Gupta and Halsam, 1979) . Their presence in legumes, and th e capacity o f tannins to precipitate forag e proteins and render them unavailable to the rumen microbes, has attracted considerable attention in recent years (Barry and Reid, 1984) . Ann et al. (1989) showed that calliandra and three other species (Acacia anqustissima. Acacia aneura and Albizia chilensis) which showe d hig h tanni n conten t wit h eithe r vanillin-HC l o r methanol-HCl method s o f analysi s exhibite d lo w nitroge n digestibilities. Th e concentratio n o f tannin , tota l phenol , flavonol glycoside and neutral detergent fibre in L^ leucocephala did not undergo significant changes over 10 weeks of leaf growth (Tangendjaja et al., 1986). Problems related to high tannin concentrations are still prevalent, but suitabl e strategie s to counte r these ar e being sough t and evaluated. Polyethylene glycol (PEG), for example, forms a soluble complex with condensed tannins (Jones and Mangan, 1977) and can be used either to prevent protein reacting with condensed tannins or to displace protein from pre-formed tannin-protein Complexes (Barry and Reid, 1984) . This principle was used by Pritchard et al. (1985) to displac e tannin s fro m tannin-protei n complexe s an d improv e nitrogen digestibility in mulga (Acacia aneura). 34 3. MATERIALS AND METHODS 3.1. Study area The study was done at the National Agroforestry Research Centre in Maseno, Western Kenya. Maseno (34 35' East and 0' North) is located 30 km northwest of Kisumu in Nyanza province. It has an altitude of between 1,500 m and 1,600 m above sea level. The bio-physical and socio-economic factors operating in and around Maseno are comparable to those in many other highland areas of Western Kenya. The long-term mean annual rainfall is about 1,750 mm. Total rainfall received during the study period was 2,995.2 mm. The distribution is bimodal with the long rainy seasons occuring between Marc h an d June , an d th e shor t rain s comin g betwee n September an d Novembe r (se e Fig . 1 ) . Average maximu m da y an d average minimum night temperatures between May '91 and November '92 were 26.6 °C and 17.8 °C respectively. These were comparable to the respective average annual maximum day and average annual minimum night temperatures of 28.5 °C and 15.6 °C in 1990, and 27.3 °C and 17.2 °C in 1991. 35 500 Rainfall/temp. month/yr Fig. 1. Average monthly rainfall and temperatures for Maseno research station (May '91-Nov. '92) 36 Predominant soil types vary with the position i n the landscape (uplands, ridges and valley bottoms) (Heineman et al., 1990). The soils around the research site are all based on Nyanzan basalt, granite and phenolitic lava s as the main parent materials. The major soil types are ferralsols, acrisols and lixisols (tropical luvisols) (see appendix 10). Population densitie s o f ove r 1,00 0 persons/km 2 ar e found , especially in some parts of South and East Bunyore locations. The natural vegetation has been replaced by cultivation and settlement. A dispersed cover of indigenous and exotic tree species found along roads and in and around farms includes Markhamia spp., Sesbania spp., Cassia spp., Aqrocarpus spp., Cuppressus spp., Eucalyptus spp.. Pinu s spp . an d Casuarin a spp . Far m size s ar e reduce d (sometimes to less than 0.5 ha per household) to the extent that they can no longer support local families, which are often 6 to 10 people per family. The land use system is foodcrop-based. The main food crops are maize, beans, bananas and vegetables. Others include sweet potatoes, cassava, yams and fruits. Annual food crops are sown twice a year (April and September) to coincide with the rainy seasons. Small-scale production of coffee and tea is done by a few farmers especiall y aroun d th e neighbouring Vihig a an d Maragoli divisions. The most important animals are cattle, but goats, sheep, pigs, poultry and bees are also kept. 37 3.2. Site, experimental design and treatments The experimental site was formerly under small bushes and couch grass (Digitari a scalarum). The slope is about 5% and is tilted both in north-south and west-east direction. Soils vary from clay loam to sandy clay, with a depth exceeding 1.5 m and a reddish brown colour . Chemical analysi s o f soi l sample s take n i n 1988 (Otieno et al., 1991) gave the following results: Average soil pH (in CaCl2; 1:2.5 ratio) 4.42 * Organic carbon (%) 0.9 3 Total available Nitrogen (%) 0.1 1 Phosphorus (by Bray no. 2 method) 13. 4 Sodium (me/100 g) 0.6 7 Potassium (me/100 g) 0.7 1 increase by 0.5 points to obtain pH in water (1:2.5) values The experiment was laid out as a randomized complete block design. The four blocks were field bunds constructed along the contour 3-4 m from each other down the slope. Each block had replications of five treatments, each of which was made up of two parallel lines of trees, grass or their shrub-grass combinations (se e appendix 2). Plot size was 4 m long and 1 m wide. The two lines were 4 m long and 0.5 m apart. The within-row spacing for trees was 0.25 m, while that of grass was 0.125 m. 38 The variou s plantin g arrangement s (treatments ) were : 1 ) pure leucaena (Leucaena leucaena, provenance: Melinda, Belize); 2) pure calliandra fCalliandr a calothyrsus . Guatemala); 3 ) pure Napier grass (Pennisetu m purpureum f Vet . farm, Maseno); 4) leucaena + Napier; 5) calliandra + Napier. 3.3. Plant species Three plant species, one grass and two shrubs, were used. Napier grass (Pennisetum purpureum) (Judd, 1979) is a perennial native to tropical Afric a an d ha s been introduce d t o man y part s o f the tropics and sub-tropics. It is propagated by stem cuttings and grows bes t o n dee p porou s soil s o f moderat e t o fairl y heav y texture. It occurs naturally in areas with not less than 40 inches of rainfall. Its fibrous root system can reach 4.5 m below ground. It is a major fodder resource in western Kenya and is grown by many farmers on terrace bunds or fodder banks, mainly for cut-and-carry purposes. Leucaena (Leucaen a leucocephala ) (NRC , 1984) is a fast growing multipurpose shrubby species native to central America, but has been introduced to the vast majority of tropical and subtropical countries. It withstands large variations in rainfall, sunlight, salinity and terrain, with reduced vigor at high elevations, low soil pH and prolonged drought. Leucaena grows best where annual rainfall i s 1,000-3,00 0 m m an d survive s dry season s lastin g 8 39 months an d occasionall y eve n 1 0 months . I t i s propagate d b y seedlings an d ofte n develops a lon g taproot syste m an d small, sparse lateral roots carrying nitrogen-fixing Rhizobium nodules. It coppices readily. Calliandra (Calliandra calothyrsus) (NRC, 1983) is a fast growing shrubby legume also native to central America. It is propagated by seedlings and has both superficial and deep-growing roots. Roots on seedlings only 4-5 months old can be as deep as 1.5 m and spread 2 m out from the stem. Roots fix atmospheric nitrogen because of symbiosis with Rhizobium bacteria. The species coppices well when lopped, but attains a maximum height of 10 m and a diameter of 20 cm when undisturbed an d when under favourabl e conditions (NAS , 1979). It is restricted to the tropics and can thrive in various soil types. It requires 1,000 mm of rainfall annually, and thrives at altitudes from 150 to 1,500 m. It can withstand drought of 3 to 6 months without losing its leaves. Leucaena and calliandra were introduced to Kenya in the 1980's (Getahun, 1989) and have high potential for agroforestry. In the Siaya and south Nyanza districts of western Kenya, leucaena now accounts for about 12% of all trees in cropland, while calliandra accounts fo r at leas t 2 % of al l tree s i n boundaries o r paths (Scherr and Alitsi, unpublished data, 1991). These species are rich in nutrients and are being promoted for animal fodder in the highly populated areas of the region. 40 3.4. Establishment and management The experiment was established in 1988 (Heineman et al., 1990). The area between bunds was utilized to grow food crops. Trees and/or grasses were planted in staggered lines in order to increase the effectiveness of the planted combination on the bund against soil erosion. Trees received 25 g of diammonium phosphate (DAP) (4.5 g of N  an d 11. 5 g  P 205) at planting . Th e pur e tre e an d gras s treatments, togethe r wit h th e tre e componen t i n th e mixe d treatments, were established in April. The grass component in the mixed treatments was established between August and November of the same year. Trees were managed as hedges and were cut back to 0.5 m in height. Grasses were cut back to ground level. Cutting frequency and timing of cutting between treatments varied initially (between August 1988 and January 1989) depending on the coppice regrowth. From April 1989 on, all treatments were cut on the same date and this was carried out at least twice every growing season at the time of sowing the food crop and around two months later. From May 1991 on (when cutting back had already been done 16 times), hedgerows were strictly harvested at intervals of 2 months, marking the beginning of this study. 41 3.5. Biontass measurements, sample collection and weather records At harvesting , onl y th e centr e 2  m  lengt h o f eac h lin e wa s considered. The outer 1 m length on both sides of each line was excluded in order to avoid border effects. The shrubby species were separated and measured as leaf and stem fractions. This ensured that quantitative data were collected specifically on edible parts and also provided the possibility for determining the leaf/stem ratios. At the same time, approximately 500 g of thoroughly mixed samples of fresh leafy material were collected and oven-dried at 60 °C for 48-72 hrs in order to calculate the proportion of dry matter (DM%). Dried sample s wer e groun d t o pas s a  1  m m sieve , packed an d transported t o th e University o f Britis h Columbia , Canad a fo r nitrogen, mineral and acid detergent fibre (ADF) analysis. Daily rainfall and temperature records were kept using the meteorological equipment available at the Maseno Agroforestry Research Centre tree nursery. The nursery is located about 200 m from the experimental site. 3.6. Analytical procedures 42 3.6.1. Nitrogen and minerals Before th e inorgani c constituent s o f plan t substance s ca n b e determined, i t i s generall y necessar y t o destro y th e organi c matter. This was achieved i n this experiment by using the wet oxidation procedures outlined by Parkinson and Allen (1975). Exactly 1.000 g samples of oven-dried foliage material were weighed into 100 mL digestion tubes. Five mL cone, sulfuric acid was added and immediatel y mixed with a vortex mixer. Four mL of lithium sulfate- peroxide mixtur e (mad e by dissolvin g 0.2 1 g  seleniu m powder and 7 g lithium sulfate crystals in 175 mL 30% H202) were added in 1 ml aliquots while applying discontinuous heating. After heating on a block for 1.5 h at 360 °C, 0.5 mL of H202 was added, and the tubes were then returned to the block for 0.5 h. The later step was repeated . Another 0. 5 m L H 202 and 1 0 more minutes of digestion were added whenever digests had not attained the expected pale yellow o r milky white colour . The resultan t digest s were cooled to room temperature, made up to 100 mL, transferred to 125-mL plastic bottles and stored in a refrigerator (abou t 5 °C) to await analysis. Appropriate dilutions were made before chemical analysis. Nitrogen (N) and phosphorus (P) were determined with a Lachat autoanalyser. 43 Calcium (Ca), magnesium (Mg), potassium (K), zinc (Zn) and copper (Cu) wer e determine d b y mean s o f a n atomi c absorptio n spectrophotometer. All analyses were read in mg/L and most were calculated and reported in percentages (cg/g). Elements that were relatively low in concentration (Zn and Cu) are reported in mg/kg (ppm). With the knowledge that crude protein contains about 16% nitrogen (Schneider and Flatt, 1975), the nitrogen contents of the samples were multiplied by 100/16 = 6.25 t o estimate the crude protein contents. 3.6.2. Acid detergent fibre (ADF) The micro-digestio n procedur e o f Walder n (1971 ) was utilized . Approximately 0.3 5 g of samples previously oven-dried at 105 °C were accurately weighed into 50 mL digestion tubes. Thirty-five mL of acid detergent solution [28 mL HgSO L^ of deionized water, plus 20 g Hexadecyltrimethylammonium Bromide (CETAB)/L of solution] and 1 mL of Decahydronaphthalene (Decalin ) were added to each tube. Each tube was covered with a large marble and the contents heated in a digestion block at temperatures up to 104 °C for exactly 1 hour. Contents were filtered under suction and washed twice each with hot demineralized water and acetone. The residue was oven-dried a t 10 5 °C , desiccate d an d weighed . Correcte d AD F wa s calculated by subtracting acid-insoluble ash from the uncorrected ADF value as follows: 44 %ADF =  (crucible + dry residue) - crucible wt. (uncorrected) x  100 Dry sample wt. The as h tha t ha d no t disolve d i n aci d (mainl y silica ) wa s determined by heating the residue in a muffle furnace at 475 °C for at least 2 hrs. % acid-insoluble ash = (crucible + ash) - crucible wt. x 100 dry sample wt. Corrected ADF % = ADF % (uncorrected) - % acid-insoluble ash 3.7. Statistical procedures All data were originally stored in Quattro (Borland International, 1993) and later imported into SAS 6.04 (SAS , 1985). An arcsine transformation was performed to improve normality of those figures expressed i n percentages. Bartlett's test (Nete r and Wasserman, 1974) wa s utilize d t o tes t homogeneit y o f variances . Lo g transformations were performed on data where this assumption was violated. Cu met this assumption without any transformation. DM% and CP% met the assumption after an arcsine transformation, while 45 Ca%, Mg% and K% did so after both arcsine and log transformations. Fresh leafy biomass, dry matter, P%, Zn and ADF% did not meet the assumptions even after the above transformations, and their results should be treated cautiously. Where season*treatment interactions were not significant, analysis of variance using the general linear models procedure (GLM) , was done fo r th e uppe r an d lowe r experimenta l row s combine d an d separately for all harvest seasons (Model : Yii = u + Si + Ti + Eii), wher e Yi i =observation, u=populatio n mean , Si=effee t o f season, Ti =treatment effect and Eii= error. Where season*treatment interactions were significant, separate analyses were done for each season (Model: Yii = u + T_j + Ej.) . When the F test was significant (a=0.05), means were compared by setting up and solving orthogonal contrasts at the a=0.05 significance level. Separate analyses were done (treating rainfall as a covariate) to eliminate effect s associate d wit h th e tota l rainfal l receive d during a growth period. Since temperatures in Maseno do not vary substantially from time to time (Fig. 1), rainfall was assumed to be the main factor determining seasonal effects. Linear correlation analyses wer e don e t o determin e th e strengt h o f relationshi p between rainfal l an d th e biomas s and/o r nutrien t data . Th e closeness o f th e relationshi p wa s expresse d a s th e linea r correlation coefficient (r). 46 4. RESULTS 4.1. Leaf/stem ratios Table 1 shows the proportion of fresh leaf and wood fractions for each of the two shrubby species over 18 months at two-month cutting intervals beginning in May, 1991. For ease of reference, harvesting dates are numbered as shown in Appendix 1. Leucaena leucocephala had a n averag e o f 73.8 % leave s an d 26.2 % wood ; Calliandr a calothyrsus was 70.9% leaves and 29.1% wood. Leaf/stem ratios were consistently higher for leucaena than for calliandra 89% of the time, averaging 2.96 and 2.58 respectively. Ratios obtained at the first cutting were similar to those obtained at the same time the following year, and were even higher at the end of the study period (November, 1992). Table 1. Average Leaf/stem ratios for Leucaena leucocephala and Calliandra calothyrsus over 18 months. 47 Harvesting Date 1 2 3 4 5 6 7 8 9 10 % Leaves 67.8 78.9 78.2 71.3 71.1 69.3 81.0 68.5 77.9 LEUCAENA % Stems 32.2 21.1 21.8 28.7 28.9 30.7 19.0 31.5 22.1 Ratio 2.11 3.74 3.59 2.48 2.47 2.26 4.26 2.17 3.52 CALLIANDRA % Leaves 64.9 78.2 76.8 70.8 69.4 66.0 59.7 75.6 76.3 % Stems 35.1 21.8 23.2 29.2 30.6 34.0 40.3 24.4 23.7 Ratio 1.85 3.59 3.31 2.42 2.27 1.94 1.48 3.10 3.22 Average 73. 8 26. 2 2.9 6 70. 9 29. 1 2.5 8 48 4.2. Leafy biomass yields 4.2.1. Fresh biomass yields The mean fresh and dry leafy biomass yields of the three fodder species and their combinations are presented in Table 2. Analysis of variance tables and the results of mean separations are provided in Appendices 7 and 8, and 9 respectively. Productivity from the pure treatments could be ranked from the highest to the lowest in the following order: Pennisetum purpureum. Leucaena leucocephala and Calliandra calothyrsus. The mixed treatments gave significantly higher total fresh biomass yields than the pure shrub treatments (p=0.0001) . The pure grass treatment had significantly higher fresh biomass yields (p=0.0001) than either of the shrubby species, but was significantly lower (p=0.0106) than the shrub-grass mixtures. Biomass yields from the two shrubs were not significantly different (p=0.6889) . Biomas s yields from three harvests in the first year were significantly (p=0.0025) higher than those from corresponding harvest times in the second year. Fresh biomass was not significantly correlated with rainfall (Table 4). When treated as a covariate, the latter's effects were found not to be significant (p=0.1268). 49 In figures 2-21, lines are drawn between data points, not to imply interpolated linea r trends, but to facilitate discrimination of data sets when reading the graphs. As illustrated by Fig. 2, the pure grass treatment declined in productivity over time, while the pure shrub treatments maintained relatively uniform productivity. It is suggested that the sharp drop associated with productivity in the mixed treatments may be due to the performance of the grass component in the mixture. Table 2. Mean fresh and dry leafy biomass yields (kg/lOOm) of Leucaena leucocephala f Calliandra calothyrsus and Pennisetum purpureum over 18 months. TREATMENT FRES H LEAFY BIOMASS DR Y LEAFY BIOMASS LI L 2 L I L 2 TOTA L L I L 2 TOTA L Leuc. +  Leuc. Call. +  Call . Napier + Napie r Napier +  Leuc. Napier + Call . 54.53 39.60 125.21 222.89 222.81 59, 36. 98. 82. 58. .06 .35 .75 .62 ,49 113.59 75.95 225.30 305.51 281.30 17.06 14.62 25.54 45.46 46.58 17.59 13.48 19.47 20.94 21.85 34, 28. 45. 62. 68. .65 .11 .86 .40 ,32 LI = upper line of treatment L 2 = Lower line of treatment 50 700 Fib (kg/100m ) 4 5  6  7 time of harvest Pig. 2. Fresh leafy biomass yields of L  leucocephala, C. calothyrsus  an d P.  purpureum  ove r 18 months of repeated cutting 51 A separate assessment was done on individual lines. Planting Napier grass in combination with calliandra increased the yield of both components significantly (appendix 9) . However, the Napier-leucaena combination benefited the former (p=0.0125), but did not cause any significant difference in the latter (p=0.4170). 4.2.2. Dry matter yields The mixe d treatment s wer e significantl y highe r i n dry biomas s production than each of the pure treatments (Table 2). Dry matter productivity fro m th e pur e Napie r gras s wa s significantl y (p=0.0459) higher than that of calliandra but not significantly different (p=0.8921) from that of leucaena treatments. Yields from leucaena an d calliandr a wer e no t significantl y different . Dr y matter yields from the first three harvests in the first year were significantly (p=0.0002) higher than the corresponding harvests in the secon d year . Covarianc e analysi s indicate d n o significan t (p=0.7472) effects associated with rainfall. Productivity in all treatments was higher at the beginning than at the end of the study period, but the drop was substantial for those treatments with a grass component in them (Fig. 3). Planting Napier grass in combination with leucaena did not cause any significan t difference s i n the dr y matte r yield s o f bot h species (see Appendix 9) . Both components in the Napier-Calliandra mixture benefited significantly. 52 160 120 80 40 Dm (kg/100m) —~~ Leucaen a —\~ Calliandr a Napier Leuc. •  Napie r ~~*~ Call . +  Napie r 4 5  6  7 time o f harves t 8 10 11 Fig. 3. Dry leafy biomass yields of L. leucocephala, C. calothyrsus and P. purpureum over 18 months of repeated cutting 53 54 4.3. Nutrient concentrations 4.3.1. General The nutrient composition of the three fodder species as determined across the entire study period is summarized in Table 3 (also see Appendices 3, 4 and 5). All nutrients except zinc and copper are expressed in percentages (cg/g) . Zn and Cu were relatively low in concentration, and are presented i n parts per million (mg/kg) . Acid-insoluble ash was negligible in the shrub species, but was high (6.3%) in Napier grass. The acid detergent fibre (ADF) values provided i n th e tabl e ar e correcte d value s obtaine d afte r subtracting the acid-insoluble ash content from all samples. 55 Table 3. Comparative nutrient compositions of L. leucocephala, C. calothyrsus and P. purpureum over 18 months. Nutrient Leucaen a Calliandr a Pennisetu m leucocephala calothyrsu s purpureu m Dry matter, % Crude protein, % Potassium, % Calcium, % Magnesium, % Phosphorus, % Zn (ppm) Cu (ppm) ADF, % Acid-insol. ash, % 32.0 27.2 1.62 0.757 0.313 0.178a 19.2 8.8b 29.4 -38.5 21.7 1.08 0.552 0.218 0.182a 27.6 8.6b 54.1 -21.9 10.6 1.98 0.275 0.365 0.148 24.0 5.4 39.7 6.3 Means followed by the same superscript on the same row were not significantly different (a=0.05) As shown by Table 3, both the shrubs had a higher proportion of dry matter than the grass. This means that consumption of the same amount of fresh material results in a comparatively higher dry matter intake of the legumes. L. 1eucocepha1a was highest in crude protein, Ca and Cu, but it was lowest in Zn and ADF. C. calothyrsus 56 was highest in P, Zn and ADF, but was lowest in K and Mg. P. purpureum was highest in K and Mg, but was lowest in crude protein, Ca, P and Cu. Since ther e wer e significan t interaction s (a=0.05 ) betwee n treatment and time of harvest (season), it was necessary to perform separate statistical analyses for each harvest time in order to compare the means at the a=0.05 significance level. The results are summarized in Appendix 9. Correlation analysis showed that most of the nutrients were not significantly relate d t o rainfall under Maseno conditions. As shown by the Spearman rank order coefficients presented i n Table 4 , only the K conten t o f Napier grass was significantly (r=0.750) correlated with rainfall. The time to time variation of each of the dry matter and nutrient concentrations are shown for the upper and lower lines of each treatment in Figures 4-21. 57 Table 4. Spearman rank order coefficients between rainfall and nutrients in L. leucocephala, C. calothyrsus and P. purpureum. Significant correlation between rainfall and nutrient FLB = Fresh leafy biomass DM = Dry matter Biomass/ nutrient FLB, kg/100 m DM, kg/lOOm Dry matter, % Crude protein, % Calcium, % Magnesium, % Potassium, % Phosphorus, % Zinc, (ppm ) Copper, (ppm ) Acid det. fibre, % Leucaena leucocephala -0.283 -0.250 -0.024 0.500 -0.067 0.234 0.000 -0.300 -0.500 -0.059 -0.444 Calliandra calothyrsus -0.100 -0.393 -0.286 0.385 0.500 0.467 -0.217 0.343 -0.075 -0.059 0.033 Pennisetum purpureum 0.183 -0.071 -0.595 0.360 -0.600 -0.400 0.750* 0.326 -0.100 0.259 0.517 58 4.3.2. Percent dry matter (DM%) Calliandra was significantly higher than both leucaena and Napier grass in DM%. Likewise, leucaena was significantly higher than Napier grass in eight out of nine harvests. DM% varied as shown in Figures. 4 and 5. The DM% of both shrub s and the grass component i n the Napier-calliandra mixtur e wer e no t significantl y affecte d b y mixe d planting at any one time. In the Napier-leucaena mixture, the DM% of the grass was not significantly different in six out of eight harvests. It was significantly lower in two. 5 9 80 70 60 50 40 30 20 10 0 Dm% ~~*~ Leucaen a (L ) H — Calliandr a (C ) - * - Napie r (N ) N (i n NXL ) 4 5  6  7 Date o f harves t 8 10 11 Fig. 4. Variation in the dry matter content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyraus over 18 months 60 Dm% 80 70 60 50 40 -30 20 10 0 — *~ Leucaen a (L ) -+- Calliandr a (C ) - * - Napie r (N ) - H - L  (I n NXL ) - * - C  (i n NXC ) 4 5  6  7 time o f harves t 10 11 Fig. 5. Variation in the dry matter contents of L. leucocephala  an d C.  calothyrsus  whe n grown in pure arrangements and in combination with P.  purpureum  ove r 18 months 61 4.3.3. Crude protein All species were significantly different from each other in crude protein concentration. At the end of the study period, the crude protein conten t o f eac h specie s wa s simila r t o tha t a t th e beginning (Figures 6 and 7) . A relatively uniform level of crude protein was maintained throughout the study period. No significant difference in the protein concentration of Napier grass wa s observe d whe n i t wa s plante d i n combinatio n wit h leucaena. In combination with Calliandra, there was no significant difference in eight out of nine harvests and a significant decrease in only one. Apart from calliandra, which was significantly higher in only one out of nine harvests, shrub-grass mixtures did not cause any significant difference in the crude protein concentration of the legumes. 62 60 55 50 45 40 35 30 25 -20 15 10 5 -0 CP% —*~~ Leucaen a (L ) - + " Calllandr a (C ) - * - Napie r (N ) - B - N  (i n NXL ) -*- N  (i n NXC ) 4 5  6  7 Date o f harves t 8 10 11 Pig. 6. Variation in the crude protein content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 63 CP% " _ " ^ H - ^ a , 1 5^—~^__^_ ^ ^ ^ B - u ^^=^r== i i  i —*~ Leucaen a (L ) - 4 - Calliandr a (C ) - * - Napie r (N ) - B - L  (i n NXL ) - * - C  (i n NXC ) ^ / ^ ^ i ^ v ^ ^ ^ ~ ~ ~ * — * — * i i  i  i  i 4 5  6  7 Date o f harves t 8 9  1 0 11 Fig. 7. Variation in the crude protein contents of L. leucocephala  an d C.  calothyrsus  whe n grown in pure arrangements and in combination with P.  purpureum  ove r 18 months 4.3.4. Calcium 64 Both shrubs were significantly higher in Ca concentration than the grass. Leucaena was significantly higher than calliandra in seven out of nine harvests. Napier grass maintained a uniform level of Ca throughout the study period, while both legumes showed a gradually increasing trend (Figures 8 and 9) . Ca concentration in Napier grass was not significantly affected when planted i n combination with th e legumes . When mixed with Napier grass , th e concentratio n o f C a i n bot h leucaen a an d calliandra were not significantly different in eight out of the nine harvests. 4.3.5. Potassium Leucaena wa s significantl y highe r tha n calliandr a i n K concentration. K content of Napier grass was significantly higher than that of leucaena in six out of nine harvests, but was not significantly different in the other three harvests. Figures 10 and 11 show all species to have returned to about the initial level of K by the end of the study period. 1.6 1.2 0.8 0.4 Ca% —*~~ Leucaen a (L ) — t - Cal l landr a (C ) - * - Napie r (N ) - B - N  (i n N X L ) - X - N  (i n NXC ) 4 5  6  7 Date o f harves t 10 11 Fig. 8. Variation in the calcium content of P. purpureum when grown in pure arrangements and in combination with L. leucocephala and C. calothyrsus over 18 months 66 1.6 1.2 0.8 0.4 I Ca% — *~ Leucaen a (L ) - 1 - Calliandr a (C ) - * - Napie r (N ) - H - L  (i n NXL ) - * - C  (i n NXC ) 4 5  6  7 Date o f harves t 8 10 11 Fig. 9. Variation in the calcium contents of L. leucocephala  an d C. calothyrsus  whe n grown in pure arrangements and in combination with P.  purpureum  ove r 18 months 67 K% "~*~" Leucaen a (N ) -4— Cailiandr a (C ) - * - Napie r (N ) - S - N  (i n NXL ) - * - N  (I n NXC ) 4 5  6  7 Date o f harves t 8 10 11 Fig. 10. Variation in the potassium content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C.  calothyrsus  ove r 18 months 68 6 5 4 3 2 1 n K% _ -H i —&— I • — B i —a-——s&^=== i i ~*~~ Leucaen a (L ) - + - Calllandr a (C ) - * - Napie r (N ) - B - L  (I n NXL ) -*- C  (I n NXC ) I I  I  I  I 4 5  6  7 Date o f harves t 10 11 Fig. 11. Variation in the potassium contents of L. leucocephala  an d C.  calothyrsus  whe n grown in pure arrangements and in combination with P.  purpureum  ove r 18 months 69 K concentration of Napier grass was higher in the mixed treatments. The difference was significant in five out of nine harvests in the Napier-leucaena mixture and in seven out of nine harvests in the Napier-calliandra mixture . Excep t fo r calliandra , whic h wa s significantly lowere d i n only one harvest, none of the legumes differed significantly in the concentration of K when planted with Napier grass. 4.3.6. Magnesium Mg concentration of Napier grass was significantly higher than that of leucaena in all except two out of nine harvests. Mg content of calliandra was significantly lower than that of Napier grass and leucaena in eight out of nine harvests. Mg levels of all species at the end of the study period were similar to those at the beginning (Figures 12 and 13). Mg concentratio n o f Napier gras s decrease d whe n combine d wit h either legume. The decrease was not significant in seven out of nine harvests with leucaena and three out of nine harvests with calliandra. None of the woody perennials, however, was affected significantly by planting in combination with the grass. 70 Mg% —"— Leucaena (L ) —I— Calliandr a (C ) - * - Napie r (N ) - S - N  (i n NXL ) - * - N  (i n NXC ) 0 1 4 5  6  7 Date o f harves t 10 11 Fig. 12. Variation in the magnesium content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C.  calothyrsus  ove r 18 months Mg% 8 ~ — Leucaen a (L ) —t— Calliandr a (C ) - * - Napie r (N ) - B - L  (i n NXL ) - * - C  (I n NxC ) 0 1 4 5  6  7 Date o f harves t 10 11 Pig. 13. Variation in the magnesium contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 72 4.3.7. Phosphorus Both legumes were higher in P concentration than Napier grass. However, the difference was significant in only four out of nine harvests wit h leucaen a an d fiv e ou t o f nin e harvest s wit h calliandra. I n al l cases , th e legume s wer e no t significantl y different from each other. The P concentration of all species were at similar levels at the end as at the beginning of the study period (Figures 14 and 15). P concentration in all species was not significantly altered by planting in shrub-grass mixed arrangements. 4.3.8. Zinc Calliandra was either similar or higher than the other species in Zn concentration in all harvests. The difference was significant in eight out of nine harvests with leucaena, and in four out of nine harvests with Napier grass. Napier grass was significantly higher in Zn concentration than leucaena in four out of nine harvests. Figures 16 and 17 show the variation of the concentration of Zn in all species over 18 months. Planting Napier grass with either of the legumes did not cause any significant difference in the zinc concentration of all species. 73 0.4 0.3 0.2 0.1 P% i i  i — *~ Leucaen a (L ) —f~" Calliandr a (C ) - * - Napie r (N ) - B - N  (i n NXL ) - X - N  (i n NXC ) 7 I I '  I--  '  ~yt— "^^?«fe<r ^ ^ > ^ ^ ^ I I  I  I  I  I 4 5  6  7 Date o f harves t 10 11 Fig. 14. Variation in the phosphorus content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C.  calothyrsus  ove r 18 months 74 P% 0.4 0.3 0.2 0.1 —*— Leucaen a (L ) —I— Calliandr a (C ) - # - Napie r (N ) - B - L  (I n NXL ) - * - C  (i n NXC ) 4 5  6  7 Date o f harves t 8 10 11 Fig. 15. Variation in the phosphorus contents of L. leucocephala  an d C. calothyraus  whe n grown in pure arrangements and in combination with P.  purpureum  ove r 18 months 75 Zn ppm -— T 7 ^ ^ \ # ^^^^-^Si^ \ . 1 1  1 — -+-- * -- B -^__ | _ H  - * ; g ^ ' ^ ^ " ~ - ^ f c > ^ i i  i i  i Leucaena (L ) Calliandra (C ) Napier (N ) N (i n NXL ) N (I n NXC ) ^ ^ L 1 1 4 5  6  7 Date o f harves t 10 11 Pig. 16. Variation in the zinc content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C.  calothyrsus  ove r 18 months 76 Zn (ppm ) 60 r 50 -40 -30 -20 -10 -0 -0 1 2 3 4 5 6 7 8 9 1 0 1 1 Date o f harves t Fig. 17. Variation in the zinc contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months —"— Leucaena (L) 77 4.3.9. Copper Cu concentrations were the lowest among the nutrients studied. Both woody perennials were higher in the concentration of copper than the grass. The differences were significant in seven out of nine harvests fo r leucaen a an d eigh t ou t o f nin e harvest s fo r calliandra. The concentrations of Cu in the two legumes were not significantly different in seven out of the nine harvests. By the end of the study period, Cu concentrations had returned to levels similar to those at the beginning, but both shrubs showed peaks at the 6th harvesting date (Figures 18 and 19) . The concentrations of Cu in Napier grass and calliandra were not significantly altere d by establishment i n shrub-grass mixtures. Similarly, the Cu concentration in leucaena was not significantly different in eight out of nine harvests. 4.3.10. Acid detergent fibre (ADF) Calliandra was significantly higher than leucaena in fibre (ADF) content. ADF content in calliandra and leucaena was, respectively, significantly higher, and significantly lower than that of Napier grass in eight out of nine harvests each. The shrubs tended to gradually increas e while the grass tended to remain at similar levels of ADF throughout the study period (Figures 20 and 21). 78 Cu (ppm) --i i  i  i ~— Leucaen a (L ) — 1 - Calliandr a (C ) - * - Napie r (N ) / \ -Q - N  (in NXL ) / \  - * - N  (In NXC) / #-^i^-* ^ ^ ^ ^ ^ i i  i  i  i  i 0 1  2 4 5  6  7 Date o f harves t 10 11 Fig. 18. Variation in the copper content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C.  calothyraus  ove r 18 months 79 20 15 10 Cu (ppm) 5 -—— -+-a -3K -/ \  _ a _ /^-^\ \  ~* ~ i i  i  i  i  i  i  i Leucaena (L ) Calliandra (C ) Napier (N ) L (i n NXL) C (in NXC) i i 4 5  6  7 Date o f harves t 10 11 Pig. 19. Variation in the copper contents of L. leucocephala and C. calothyrsus when grown in pure arrangements and in combination with P. purpureum over 18 months 80 _ + -~"""K \ / * - » = ^ * ^ — ^H^= -^H--1 1  1  1 — — 1 ftT"2 5*" i — — Leucaen a (L ) - + - Calliandr a (C ) - * - Napie r (N ) - B - N  (i n NXL ) - X - N  (i n NXC ) A 1  1  - + ^ ^ ^ -i i  i  i 4 5  6  7 Date o f harves t 10 11 Fig. 20. Variation in the ADF content of P.  purpureum when grown in pure arrangements and in combination with L. leucocephala  an d C. calothyrsus  ove r 18 months 8 1 120 100 80 60 40 20 ADF% ~~*~~ Leucaen a (L ) - 4 - Calllandr a (C ) - * - Napie r (N ) - H - L  (i n NXL ) C (i n NXC ) 0 0  ~ B 1 2  3 4 5  6  7 Date o f harves t 10 11 Fig. 21. Variation in the acid detergent fibre Contents of L. leucocephala  an d C.  calothyrsus  whe n grown in pure arrangements and in combination with P. purpureum  ove r 18 months 82 The AD F content s o f Napie r gras s an d leucaen a wer e no t significantly changed by their shrub-grass combinations. In the Napier-calliandra mixture , th e proportio n o f AD F wa s no t significantly affected eight out of nine harvests both ways. 5. DISCUSSION 83 5.1. Leaf/stem ratios Knowledge of leaf/stem ratios provides an opportunity for making quick estimations of edible biomass in woody forag e plants. It allows optimum rationing of tree-based feeds while at the same time avoiding tim e an d labou r cost s tha t woul d b e incurre d whe n separating leaves and stems for feeding or measurement purposes. The results obtained i n this study show that leucaena regrowths have a higher proportion of leaves than calliandra regrowths. Phiri et al . (1992 ) suggeste d tha t th e muc h highe r AD F valu e i n calliandra than in leucaena could be due to the lower leaf to stem ratio in the former. The information available in the literature concerning thi s aspec t i s limite d an d appear s t o sho w sligh t differences from place to place. Pathak et al. (1980) reported an average leaf/stem ratio of 2.01 (1.53-2.71) for leucaena. This ratio decreased significantly with increasing cutting interval and increasing planting density, but was not significantly affected by changes in cutting height. In his literature review, Jones (1979) reported regrowing leucaena shoots, 0.5 to 1 m long, to consist of 75 to 80 percent nutritious leaves. Leaf/stem ratios based on dry matter yields (tonnes/ha/year ) of leucaena wer e 2.0 7 fo r cultiva r K34 1 an d 1.8 7 fo r cultva r K 8 (Guevarra e t al. , 1978) . Youn g calliandr a tree s plante d a t 84 different spacings yielded 59-68% leaves when first cut back to 1 m after five months of growth (Baggio and Heuveldop, 1984). When new sprouts were cut back to the same height, five months later, the proportion of leaves ranged from 57-63%. Most of these values are lower than the results of the present study and the differences may be attributed either to the bio-physical environment, the plant varieties, or management strategies. 5.2. Leafy biomass yields The high biomass productivity shown by Napier grass in this study has long been recognized by Kenya's Ministry of Agriculture and has been widely utilized to benefit small and large scale dairy farmers in the high potential areas (Abat e et al., 1985). The apparent setback observed in this experiment was the significant loss in productivity in the second year. Similar results were obtained in Kitale, western Kenya, where in the absence of fertilizer, the dry matter yiel d o f Napie r grass , Guine a gras s an d Rhode s gras s decreased substantiall y i n th e secon d yea r afte r plantin g (Sheldrick and Thairu, 1974; cited by Wolfgang Bayer, 1990). This phenomenon may be attributed to the continued depletion of the soil nutrient resource s resultin g fro m repeate d harvestin g withou t replenishment. Wolfgang Bayer (1990 ) further state d tha t while unfertilized Napier swards became unproductive after 3-4 years, those which received doses of appropriate fertilize r showed no decline in yields even after 30 years. 85 Though the study period was relatively short, the maintenance of biomass productivity by the woody perennials supports reports from many sources that woody legumes have advantages over grasses in this respect. In Indonesia where it was introduced in 1936 with seeds fro m Guatemala , calliandr a plantation s fo r firewoo d production are harvested annually for 15-20 years and produce 35-65 m3 per hectare pe r year (NAS , 1979) . Similarly , som e leucaen a hedges in Hawaii have been trimmed at least twice a year for more than 40 years (NRC, 1984). These observations may be attributed to the high coppicing ability and the extensive taproot systems which tap water and essential nutrients from far below the soil surface, sustaining biomas s productivit y eve n durin g drie r periods . Leucaena, for example, is so resilient that pastures near Brisbane in Australia have been browsed almost continuously fo r about 20 years without requiring replanting (NAS, 1979) . Aspects relating to appropriate management of individual shrub and grass species are discussed in section 2.4. Establishing an d managin g appropriat e tree s an d shrub s i n association with grasses fo r fodder has potential quantit y and quality advantage s fo r extende d periods . Th e relativel y hig h productivity shown by the mixed treatments were a direct result of significant increases in biomass production (of at least one of the components in the mixture) when components were grown as mixtures rather tha n pur e hedgerows . Thes e demonstrate d th e positiv e interactions tha t occurre d whe n Napie r gras s wa s plante d i n 86 combination with the two woody legumes. From earlier studies in the same experiment (Otieno et al., 1991) , Napier gras s gained i n fres h biomass productivity b y 48 % when planted in combination with leucaena and by 9% when in combination with calliandra. Both shrub species benefited from the combination with calliandr a gainin g b y 43 % an d leucaen a b y 35% . Thes e advantages have long been appreciated elsewher e by farmers (who graze their animals in tree-grass mixed pastures) and researchers alike. In South-East Queensland, Australia, for example, pastures combining leucaen a an d Rhode s gras s (Glori s aayana ) have been grazed fo r more than 20 years and have retained a good balance (NRC, 1984) . Similarly, leucaena/Brachiaria brizantha pastures in Malaysia and leucaena/pangola grass pastures in Northern Australia have remained in balance after being grazed for several years (NRC, 1984). Research in India showed significantly higher total green forage and dry matter yield in the treatment where hybrid Napier and subabul (Leucaen a leucocephala) were planted in paired rows (Gill et al., 1990). These mutual benefits may be attributed to: 1) the nitrogen-fixing and nutrient-cycling ability of the legumes, resulting in improved soil nutrient status, 2) the reduction of direct competition for moisture and nutrients by adjacent fodder lines; grasses explore upper layers of the soil profile, while trees and shrubs have long taproot systems which occupy much deeper soil layers, and 3) the 87 combined protectiv e effec t agains t run-of f an d soi l erosion , thereby ensurin g maximu m utilizatio n o f availabl e moisture . Information on soil conservation aspects of different agroforestry practices have been dealt with in detail by Young (1989). 5.3. Nutrient concentrations 5.3.1. Comparative information from literature The nutrient composition of leucaena, calliandra and Napier grass as found in the literature are provided for comparison purposes in Table 5. 88 Table 5. The nutrient composition of L. leucocephala. C. calothyrsus and P. purpureum as found in some literature. N u t r i e n t Dry m a t t e r , % Crude p r o t e i n , % Potas s ium, % Calcium, % Magnesium, % Phosphorus , % Zinc (ppm ) Copper (ppm ) Acid d e t e r g e n t f i b r e , % Leucaena l e u c o c e p h a l a 2 8 . 0 s , 2 8 . 3 b , 2 5 . 2 4 a , 2 6 . 9 b , 2 5 . 9J 1.7-2.6 1 ' 2 . 3 6 j , 0 . 5 - 1 . 1 8 ' 0 . 1 7 - 0 . 4 1 1 0.16% 0.23 J" 0 . 1 5 - 0 . 3 8 ' 2 8 . 0 - 4 4 . 0 ' 7 . 0 - 1 1 . 0 ' 2 2 . 6 4 a , 2 2 . 6 b , C a l l i a n d r a c a l o t h y r s u s 3 3 . 0 a , 3 9 . 09 2 4 . 4 7 a , 2 2 . 0 f , 2 1 . 6 3 9 0 .69 k 1 .03 k 0.37 k 0 . 1 5 e , 0 . 1 1 k --4 6 . 2 7 a , 7 0 . l k Pennisetum purpureum 1 6 . 6 b , 2 0 . 0 d , 1 1 . 9 b , 1 0 . l c , 8 . 7 d , 1 . 31 d 0 . 6 0 d 0 .26 d 0 . 4 1 d 1 9 . 0 - 2 8 . 0 h 4 . 0 - 8 . 0 h 4 4 . l b , 3 8 . 3C aPhiri e t al . (1992) , 'Va n Ey s e t al . (1986) , c Sands e t al . (unpublished, 1982), "NRC (1988), eAhn et al. (1989), fNRC (1983), 9Baggio and Heuveldop (1984) , hNjwe and Kom (1988) , 'Jones (1979), jNAS (1977), kBlair et al. (1988) 89 5.3.2. Percent dry matter The product of DM% and fresh biomass/100 m length of hedgerow gives dry biomass/100 m. The DM% values obtained are similar to most of those provided in the literature (Table 5). A change in DM% causes a corresponding change in absolute dry matter values. The amount of dry matter in any fresh feed material determines the amount of nutrients in that material since all nutrients are contained within the dry matter. A kg of silage, for example, is worth less than a kg o f goo d ha y largel y becaus e i t contain s les s dr y matte r (Schneider and Flatt, 1975). DM% is expected to increase in the dry season and with age because of increase d lignificatio n i n bot h cases . However , littl e information, i s know n abou t th e effec t o f managemen t o n th e proportion of dry matter. It can be inferred from this experiment that establishing the three species as shrub-grass mixtures does not impose significant changes on the proportion of dry matter. 5.3.3. Crude protein The crude protein values determined in this study were similar to those obtained from research elsewhere (see Table 5). All species had protein contents well above the 7% level known to limit the intake of tropical forage (Milford and Minson, 1966) . However, only the two legumes were above the 19% level recommended for lactatin g 90 dairy cattle by the National Research Council (Appendix 6). This supports the general consensus that many multipurpose tree and shrub legume s hav e hig h potential s a s protei n supplement s t o conventional livestock feeds (e.g. Devendra, 1990; Bamualim et al., 1984, Flore s et al., 1979). The high quality protein associated with leucaena is attributed to the balanced nature of its amino acid compositio n (NAS , 1977) . Calliandra' s rol e a s a  protei n supplement ma y b e limite d b y th e presenc e o f hig h level s o f condensed tannins which have the capacity to bind feed protein, reducing its solubility in the ruminant digestive system (An n et al., 1989 ) . The results are of particular interest to the small scale dair y owner s o f western Keny a where protein i s a major limiting facto r i n many livestoc k production system s and where household income is generally too low to allow purchase of high protein concentrates. Apart from type of species, the protein content of forage plants is expected t o vary wit h th e environment an d typ e o f management. According t o Le Houerou (1980) , protei n conten t i n grasses is expected to drop significantly during prolonged dry periods, while that i n tree s an d shrub s behave s likewis e bu t t o a  lesse r magnitude. The protein content of leucaena fodder (mainly leaves along with some soft twigs) was found to range from 20-24% during different season s (Josh i an d Upadhyaya , 1976) . I n Siay a an d Kakamega districts , th e declin e i n crud e protei n an d genera l quality of different feedstuffs during the dry season was reflected 91 in the weight changes of livestock (Sands et al., unpublished data, 1982). Adverse effects on the weight changes of sheep and goats were minimal, and it was suggested that their selective feeding habits, compared t o cattle, allowed them to selec t the better quality forage under these conditions. However, these observations were not apparent in this experiment, where protein content of all species showe d a  positiv e bu t insignifican t relationshi p wit h rainfall. The results showed that the three species used could be established as shrub-gras s mixture s o n hedgerow s withou t causin g advers e effects on the protein content of any of them. This is advantageous in that farmers can incorporate mixed forage as hedgerows in their small farms, which not only serve as cut-and-carry fodder, but also help i n minimizin g soi l erosion . However , optimu m harvestin g frequencies mus t b e establishe d sinc e folia r protei n conten t declines with age (Jones, 1979; Snyders, 1991), causing a decline in anima l productivit y (Snyders , 1991) . Crud e protei n i s no t affected by cutting height (Pathak et al., 1980). Higher planting densities resulted in slightly lower protein percentage in Leucaena leucocephala (Pathak et al., 1980). 5.3.4. Minerals When dietary crude protein and energy are present in sufficient quantities, mineral deficiencies may depress forage utilization and 92 intake, an d anima l performanc e ca n b e increase d b y minera l supplementation (Prabowo et al., 1983) . It has been stated that the concentrations of mineral elements in forages are dependent upon the interactio n o f a number of factors , including soil , plant species, stage of maturity, yield, forage management and climate (McDowell, et al., 1983) . Under Australian conditions, on a variety of soils, Ca concentration in leucaena rarely exceeds 1% in the dry matter, whereas in the material grown in India and Malawi, values of more than 2% Ca are reported in leucaena leaf (Jones , 1979) . Jones (1979 ) found the concentrations of P, S, Ca, Mg and Na to decrease with aging leucaena leaves, while those of K, Cu and Zn did not show clear trends. According to the recommendations made by the National Research Council (1988 ) for diets fed to dairy cattle, all species in the present experiment were sufficiently high in K. Njwe and Kom (1988) found adequat e level s o f K o n fou r dominant gras s specie s o f natural pastures of the west region of Cameroon. Except in the case of calliandra, Mg levels were similarly adequate. K is not known to be a limiting nutrient in many parts of the tropics (Minson, 1990), while Mg , together wit h K , F e an d M n ma y b e deficien t unde r specific conditions i n some tropical region s (McDowel l et al., 1983). Though the legumes used in this experiment were generally higher in minerals than the grass, they did not confirm their full potentials as likely supplements to ordinary feeds. Contents of Ca, P, Cu and Zn were generally low compared to those recommended for 93 dairy cattl e (NRC , 1988) . Thi s ma y suppor t th e statemen t b y McDowell, et al. (1983) that the mineral elements most likely to be lacking under tropical conditions are Ca, P, Na, Cu, I, Se and Zn. Ca and P contents were lower than those available in the literature (Table 5). Likely causes include the effect of soil chemistry and soil characteristics. Soils at the experimental site were acidic with a pH (in water) of 4.9 (Heinema n et al., 1990). According to Andrew (1978) , low pH in tropical soil systems is accompanied by low Ca and P supply, and relatively high concentrations of Al. Under these conditions, high levels of Al cause reduced uptake of Ca an d interact s wit h P  i n o r o n th e roots , renderin g i t unavailable for absorption by plants. Zn and Cu concentrations were similar to those of other authors (Table 5 ) . Concentrations o f thes e element s wer e foun d t o b e deficient in four dominant pasture grass species studied in west Africa (Njw e an d Kom , 1988 ) . Both Z n an d C u ion s ar e highly sensitive t o soi l acidit y an d decreas e 100-fol d fo r eac h unit increase in pH (Lindsay , 1978). However, it should be noted that the determination of Cu in the forage has limited diagnostic value because it interacts with other elements, particularly Mo and S (McDowell, et al., 1983). Therefore, it is important that these elements are also determined if the validity of the results is to be improved. Since this was not done in this experiment, it is difficult to judge whether the levels of these elements were of any 94 consequence to the results obtained. There is limited information in the literature regarding the effect of management on mineral content of forages. Pathak et al. (1980) found highest foliar Ca and P concentrations at 120 (compared to 40 and 60) day cutting intervals and 20 (compared to 10 and 30) cm cutting heights of Leucaena. When 1.5, 3 and 4 plants/m2 planting densities were compared , Ca content was highest a t the medium population density, while P was highest at 4 plants/m2. The results of the present experiment showed that leucaena and calliandra can be grown i n association with Napie r gras s without causin g any negative effects on the species composition of minerals, except Mg. The Mg content of Napier grass was generally lower in the mixture, while the composition of K was generally higher. Given appropriate management, total mineral quantities in the mixed arrangements will be muc h highe r becaus e o f th e significan t increase s i n tota l biomass production. As plants mature, mineral elements such as P, K, Mg, Na, CI, Cu, Co, Fe, Se, Zn and Mo are expected to decline due to a natural dilution process and translocation of nutrients to the root system (McDowell et al., 1983). 5.3.5. Acid detergent fibre (ADF) Fibre can be defined nutritionally as the insoluble organic matter which i s indigestible by proteolytic an d diastatic enzyme s and which cannot be utilized except by microbial fermentation in the 95 digestive tracts of animals (Van Soest, 1967). Ruminants require adequate coarse fibre for normal rumen function and maintenance of normal milk fat in dairy cattle (Van Soest, 1990). Fibre quality is determined by particle size, buffering capacity, cation exchange capacity, and fermentation rate. The acid-insoluble ash, which was negligible in the legumes, and highest in Napier grass (6.3%) is mainly made of silica, and was subtracted from all sample values to get the ADF values recorded in the present study. The results agree with the statement made by Goering and Van Soest (1970) that silica content i n man y grasse s i s a  principa l facto r i n reducin g digestibility. Research workers i n recent years have argued i n favour of the neutral detergent procedure as a better method of analysing total fibre compare d t o th e traditional AD F an d crud e fibr e methods (Goering and Van Soest, 1970; Van Soest et al., 1991). NDF (neutral detergent fibre ) ca n b e correlate d wit h fill , intake , digestibility, and energy content, all of which are important for animal performance . ADF , however , bein g a  rapi d metho d fo r lignocellulose estimatio n i n feedstuffs (Va n Soest, 1963) , ca n better b e relate d t o digestibilit y tha n intak e becaus e th e lignified matrix in it is the most unavailable feed fraction (Van Soest, 1990 ) . The work done by Lohan et al. (1980 ) in Himachal Pradesh showed a significant negative correlation between in vitro dry matter digestibility an d NDF, ADF and cellulose. Van Soest (1963) foun d th e know n dr y matter digestibilitie s o f eightee n 96 forages, including ten different species of grasses and legumes to decrease wit h increasin g ADF . Wilso n e t al . (1966 ) showe d a negative relationship i n sheep between fodde r intak e and crude fibre, NDF, and ADF contents. Bamualin et al. (1980) also found that the nylon-bag dry matter digestibility of browse legumes was negatively correlated with NDF, ADF, and lignin, and positively correlated with nitrogen and ash contents. Consequently, the high ADF fraction determined for Calliandra, as compared to the other species, is expected to have a detrimental effect on its intake, digestibility, an d genera l feedin g value . A  lo w dr y matte r digestibility of 35.43% for calliandra was obtained by Baggio and Heuveldop (1984) . Foliage of Leucaena, which had the lowest ADF value i n thi s study , i s know n t o exhibi t hig h dr y matte r digestibilities, e.g 70% (NRC, 1984) and 62.2% (Van Eys et al., 1986) . The dry matter digestibility of Napier grass is intermediate at 51% (Van Eys et al., 1986). The ADF contents of all species examined in this experiment were well above the minimum (21% ) recommended by the NRC (1988 ) for dairy cattle. The values are slightly higher than most of those available in the literature (see Table 5). The differences may be associated with environmental conditions, stage of growth, plant variety, an d method s o f analysis . I n influencin g fee d fibr e composition, these factors also affect the general character of feeds. The decline in in vitro organic matter digestibility and in vitro cell wall digestibility after the start of the rains was 97 attributed to increases in ADF and lignin that occured in feed resources during the dry season (Sands et al., unpublished data, 1982). Accordin g t o thes e authors , ADF level s are expected t o increase wit h age . I n th e presen t experiment , ther e wa s n o significant relationship between rainfall and ADF. The results indicated tha t all the plant specie s used i n this experiment ca n b e establishe d i n shrub-gras s mixture s withou t affecting thei r fibr e content , an d b y inference , thei r digestibilities and feed utilization relative to ADF. It is not clear why ADF contents of both shrubs showed a gradual increase over the study period, while the grass maintained a  relatively constant level. 98 6. SUMMARY AND CONCLUSIONS Many small scale farmers in western Kenya keep a minimum number of animals in keeping with the land tenure situation associated with high human population densities in the region. Napier grass is a high biomass producing fodder crop and forms the main portion of many livestock diets. Its low levels of crude protein and most minerals, which are expected to drop significantly during extended dry periods, and its reported biomass fluctuation with season and drop overtime make it an unreliable fodder crop at some points in time. Woody perennials maintain biomass productivity for a longer time, some are high in crude protein and some minerals, and may provide supplements to conventional feeds. Initial biomass production was higher for the Napier grass than for the two shrubs. The overtime drop in productivity of the grass may be attributed to frequent cutting which weaken plant regeneration capacity, and depletion of soil nutrients; and emphasizes the need for fertilize r applicatio n o r replantin g a t shor t intervals . Resources to purchase inorganic fertilizers are limited and farmers may prefe r th e late r approach . Th e abilit y o f th e shrub s t o maintain productivity for longer periods may partly be due to their possession o f lon g taproo t system s which , unlik e i n grasses, explore deep soil layers for moisture and nutrients. Where the production of large fodder quantities is not the main objective, the low productivity of woody perennials is not of major concern 99 because smal l quantitie s o f shru b ca n provid e th e require d nutrients. The results were consistent with most of those available in the literature in that the two shrubs were higher in the concentration of most nutrients than Napier grass. All species were at similar levels of most nutrients at the end of the study period as at the beginning. Leucaena's possession of high levels of most nutrients, low levels of ADF and high dry matter digestibilities makes it superior to calliandra in fodder quality, except for the presence of mimosine in its tissues. The negative effects of Mimosine can be reduced by administering feeds at the right proportions. No toxic effects relate d t o mimosine , fo r exampl e wil l b e expecte d i n ruminants, pigs and chicken if diets contain less than 30%, 10% and 5% o f leucaen a lea f respectively . Feed s supplemente d wit h Calliandra have improved livestock performance. But calliandra's high fibre and condensed tannin contents, and consequent low dry matter and protein digestibilities, make it s relative value as animal fodder questionable. Introduction of appropriate woody perennials into the small-scale farming systems of western Kenya can alleviate problems related to fodder qualit y an d quantity . In recognitio n o f this fact , and considering the relative land tenure and logistic advantages of combining wood y perennial s an d grasse s i n smal l scal e farms , management of fodder as tree or shrub-grass combinations in small 100 holder livestoc k system s ar e bein g studied . Accordin g t o th e results o f thi s experiment , leucaen a an d calliandr a ca n b e established a s shrub-grass combinations without causing adverse effects o n biomass productivit y an d mos t nutrien t content s i f proper managemen t strategie s ar e applied . Th e shrub s fi x atmospheric nitrogen and improve the soil nutrient status, provide a microclimate for enhanced growth of both components, and provide a more effective barrier to soil erosion (Young, 1989). Of all biomass and nutrients considered in this study, only K in Napier gras s was significantl y correlate d wit h rainfall . Other correlations were not significant and varied in extent from species to species and from one parameter to another. Expected correlations between rainfall and biomass (especially of Napier grass) may have been overshadowed , i n part , b y th e significan t dro p i n th e productivity of the later over the study period. It is suggested that the presence of some rainfall in each month of the year, and the absenc e o f lon g dr y seasons , kee p soi l moistur e level s reasonably high so that adverse quality and quantity effects that have been caused by prolonged drought in dry areas elsewhere are not clearly observed at Maseno. It is also noted that nutrients, especially minerals, are highly influenced by soil physical and chemical conditions, and different recommendations might apply to different edaphic and climatic conditions. 101 These results are of particular interest to the small scale farmers of western Kenya where farms are too small to supply sufficient livestock fodder, per capita income is too low to allow purchase of high protein concentrates, and where the need to establish and manage high nutrient fodder trees and shrubs in farming systems is rising. Though the survey conducted by CARE in 1991 (Scher r and Alitsi, unpublished data, 1991) showed animal fodde r to be the primary use of less than 1% of 29 main species (comprising about 90% of trees on farms) in Siaya and South Nyanza districts, fodder trees and shrubs have continued to gain popularity in the recent past. Continued research programmes to increase pest and disease resistance an d t o reduc e toxicit y an d anti-nutritiona l characteristics of the species studied in the present experiment would be of tremendous additional benefit. 102 7. REFERENCES Abate, A., H. Kayongo-Male and M. Wanyoike. 1985. Fodder for high-potential areas in Kenya. In Animal feed resources for small-scale livestoc k producers . Proc . secon d PANES A workshop , Na i rob i, Kenya. Ann, J.H., B.M. Robertson, R. Elliott, R.C. Gutteridge and C.W. Ford. 1989. Quality assessment of tropical browse legumes: Tannin content and protein degradation. Animal feed science and technology 27: 147-156. Andrew, C.S. 1978. Mineral characterization of tropical legumes. In C.S. Andrew and E.J. Kamprath [Eds] . Mineral nutrition of legumes in tropical and subtropical soils: Proceedings of a workshop hel d a t CSIR O Cunningha m laboratory , Brisbane , Australia-January 16-21, 1978. Baggio, A . an d J . Heuveldop . 1984 . Initia l performanc e o f Calliandra calothyrsu s meissm . i n liv e fence s fo r th e production of biomass. Agroforestry systems 2(1): 19-29. Bamualin, A., R.J. Jones and R.M. Murray. 1980. Nutritive value of tropical browse legumes in the dry season: Proceedings of the Australian society of animal production 13: 229-232. Bamualim, A., S. Stachiw, R.J. Jones and R.M. Murray. 1984. The effect of fresh Leucaena leucocephala as a supplement on the utilization of pasture hay by goats. Animal prod, in Austr. 15: 259-262. Barry, T.N. and C.S.W. Reid. 1984. Nutritional effects attributable to condensed tannins, cyanogenic glycosides and oestrogenic compounds in Newzealand forages. In R.F. Barnes, P.R. Ball, R.W. Brougham , G.C . Marten an d D.J . Minso n [Eds] . Forag e legumes for Energy-efficient anima l production. USDA, ARS, CSIRO, DSIR, pp. 251-259. Bayer, W. 1990. Use of native browse by fulani cattle in central Nigeria. Agroforestry systems 12: 217-228. Blair, G.J. , MAS'U D Panjaitan , D.A . Ivory , B . Palme r an d M . Sudjadi. 1988 . An evaluatio n o f tre e legume s o n a n aci d ultisol in South Sumatra, Indonesia. J.agric. Sci. Ill: 435-431. Borland International. 1993. Quattro pro for Dos: User's guide. Borland International, Inc. Budowski, G. 1983. An attempt to quantify some current agroforestry practices in Costa Rica. In P.A. Huxley [Ed]. Plant research and agroforestry- ICRAF, Nairobi, Kenya, pp. 43-62 103 Burley, J. 1987 . Exploitation o f the potential of multipurpose trees and shrubs in agroforestry. In H.A. Steppler and P.K.R. Nair [Eds] . Agroforestry: a  decade of development, ICRAF, Nairobi, Kenya, pp. 273-287. Butterworth, M.H., N.C.C. Chintsanya , K.M.J . Phir i an d P.W.S . Mitengo-Gama. 1984 . Stall-feedin g bee f cattl e wit h agricultural by-product s i n Malawi . Tropica l Agricultur e (Trinidad) 61(1): 25-28. Butterworth, M.H. and A.K. Moshi. 1985. Practical evaluation of crop residues in agro-industrial by-products for ruminants in developing countries with emphasis on East Africa. In T.R. Preston, V.L. Kossila, J. Goodwin and S.B. Reed [Eds]. Better utilization o f cro p residue s an d by-product s i n anima l feeding: research guidelines 1. state of knowledge. Proc. of the FAO/ILCA expert consultation 5-9 march 1984. Cashman, K. 1986. The benefits of alley farming for the African farmer and her household. In P. Allen and D. Van Dusen [Eds]. Global perspective s o n agroecolog y an d sustainabl e agricultural systems. Proceedings of the sixth International conference of the INFOAM. Catchpoole, D.W. an d G.J. Blair. 1990. Forage tree legumes, I. Productivity and Nitrogen economy of Leucaena, Gliricidia. Calliandra an d Sesbani a an d tree/gree n pani c mixture . Australian journal of agricultural research 41(3): 521-530. Chenost, M. 1985. Ann, zootech. 24: 327-349. Conelly, W.T. 1992. Population pressure and changing agropastoral management strategies in western Kenya. In CM. McCorkle [ed]. Plants, animals and people: Agropastoral system s research. Westview press, Inc. pp. 155-174. Dann, P. and Low, S. 1988. Assessing the value of browse plants as alternative fodder sources. Agricultural science 1(2): 20-27. Devendra, C. 1990. The use of shrubs and tree fodders by ruminants. In C. Devendra [Ed]. Shrubs and tree fodders for farm animals: Proceedings of a workshop in Denpasar, Indonesia, 24-29 July, 1989. Dharia, S.K. , S.V . Majgaonka r an d V.A . Toro . 1987 . Effect o f feeding Gliricidia maculata leave s on growth o f crossbred heifers. Indian journal of animal nutrition 4(4): 260-263. D'Mello, J.P.F. 1982. Toxic factors in some tropical legumes. World review of Animal production, vol. xviii, No. 4. 104 Dommergues, Y.R. 1987. The role of biological nitrogen fixation in agroforestry. I n H.A . Stepple r an d P.K.R . Nai r [Eds] . Agroforestry: A decade of development, ICRAF, Nairobi, Kenya, pp. 245-271. Ella, A.C., Jacobsen, W.W. Stur and G. Blair. 1989. Effect of plant density and cutting frequency on the productivity of four tree legumes. Tropical grasslands 23(1): 28-34. Energy Developmen t Internationa l (E/DI) . 1985 . A director y o f organizations workin g o n tre e plantin g an d woo d fue l conservation in Kenya. (Compiled by W. Mathu), 99 pp. Escobar, A. and Parra, R. 1980. Los residuos agricolas fibrosos y su procesamiento. In Estratecrias para el uso de residuos de cosechas en la alimentacion animal. Turrialba, CATIE. FAO. 1985. FAO production yearbook. Rome: FAO. FAO. 1990. FAO-UNESCO soil map of the world: revised legend. World soil resources report 60. FAO, Rome Flores, J.F., T.H. Stobbs and D.J. Minson. 1979. The influence of the legume leucaena and formal-casein on the production and composition of milk from grazing cows. J. Agric. Sci. 92: 351-357. Getahun, A . 1989 . Agroforestr y fo r developmen t i n Kenya : A n overview. In W.W. Budd, I. Duchhart, L.H. Hardesty an d F. Steiner [Eds]. Planning for Agroforestry. Elsevier, Amsterdam, The Netherlands, pp.183-202. Getahun, A. and K. Reshid. 1988. Agroforestry in Kenya: A field guide. Nairobi, Kenya. Gill, A.S., S.N. Tripathi and K.S. Gangwar. 1990. Effect of stand geometry on productivity of hybrid Napier Subabul association under agroforestry system. Indian J. Aaron. 35(4): 434-437. Goering, H.K. an d P.J . Va n Soest . 1970. Forage fibr e analyses (Apparatus, reagents , procedure s an d som e applications) . Agric. Handbook No. 379. ARS-USDA, Washington, D.C. Gohl, B . 1981 . Tropical feeds . Feed informatio n summarie s an d nutritive values. FAO Animal production and health series no. 12. Guevarra, A.B., A.S. Whitney and J.R. Thompson. 1978. Influence of intra-row spacing and cutting regimes on the growth and yield of Leucaena. Agronomy Journal 70: 1033-1037. 105 Gupta, R.K. and E. Halsam. 1979. Vegetable tannins-structure and Biosynthesis. I n Huls e [Ed] . Polyphenols i n Cereal s an d legumes; Proceedings of a symposium held during the 36th annual meeting of the institute of food technologists, St. Louis, Missouri, 10-13 June 1979. IDRC, Ottawa, Canada. Gutteridge, R.C. 1988. Alley cropping Kenaf (Hibiscus cannabinus) and Leucaen a leucocephal a i n sout h easter n Queensland . Australian journal of experimental agriculture 28(4) : 481-484. Heineman, A.M., E.K. Mengich, A.D. Olang and J.H.O. Otieno. 1990. AFRENA project Maseno, Kenya progress report for the period January 1988 to January 1990. AFRENA Report no. 27. ICRAF, Na i rob i, Kenya. Hegarty, M.P . 1976 . Mimosin e i n Leucaen a leucocephal a i s metabolized to a goitrogen in Ruminants. Austr. Vet. J. 52: 490. Hegarty, M.P., R.D. Court and P.M. Thorne. 1964a. The determination of Mimosine and 3,4-Dihydroxypyridine in biological material. Austr. J. Aar. Res. 15: pp. 168-179. Herrera, P.G. 1967. Effect of height of cutting on pigeon pea and Koahaole. Agricultural tropical 23(1): 34-429. Hill, G.D. 1970. Studies on the growth of Leucaena leucocephala 1. Effect o f clea n weeding an d nitroge n fertilize r o n early establishment. Papu a an d Ne w Guine a agricultura l journa l (Papua New Guinea) 22(1): 29-30. Huggan, R.D. and S.B. Westley. 1989. Agroforestry in Kenya: Fitting the pieces together. Agroforestry today 1(1): 1-4. Humphreys, L.R . 1991 . Tropica l pastur e utilization . Cambridg e univesity press Hungate, R.E. 1966. The rumen and its microbes. Academic press, Newyork. Huston, J.E. and W.E. Pinchak. 1991. Range animal nutrition. In R. K. Heitschmidt and J.W. Stuth [Eds] . Grazing management: An ecological perspective. Timber press, Inc. Huxley, P.A. 1981a. Some characteristics of trees to be considered in Agroforestry. I n P.A. Huxle y [Ed] . Plant Research an d Agroforestry. ICRAF, Nairobi, Kenya, pp. 3-12. Huxley, P.A . 1981b . Th e rol e o f tree s i n agroforestry : som e comments. I n P.A . Huxle y [Ed] . Plan t Researc h an d Agroforestry. ICRAF, Nairobi, Kenya, pp. 257-270. 106 Ichire, O. 1993. Trees for food and fodder production in dryland Nigeria. Agroforestry today 5(1): 12-13. ICRAF. 1983. A global inventory of agroforestry systems: A project announcement. Agroforestry systems 1(3): 269-273. ICRAF. 1991. Annual report. ICRAF, Nairobi, Kenya, pp.148. Jabbar, M.A., J. Cobbina and L. Reynolds. 1992. Optimum fodder-mulch allocation of tree foliage under alley farming in South western Nigeria. Agroforestry systems 20: 187-198. Jones, R.J. 1979. The value of Leucaena leucocephala as a feed for ruminants i n the tropics. World anima l review . FAO, Rome, Italy. Jones, R.J. 1988. The future of the grazing herbivore. Tropical grasslands 22: 97-115. Jones, W.T. and J.L. Mangan. 1977. Complexes of condensed tannins of Sainfoin (Onobrychis viciifolia scop) with fraction 1 leaf protein and with submaxillary mucoprotein, and their reversal with polyethylene glycol and pH. J. sci. food agric. 28: 126-136. Jones, R.J. and R.G. Megarrity. 1986. Successful transfer of DHP-degrading bacteria from Hawaiian goats to Australian ruminants to overcome the toxicity of Leucaena. Australian vet. Journal 63(8): 259-262. Joshi, D.C. an d R.B. Upadhyaya. 1976. Leucaena leucocephala-A n evergreen protein-rich tre e fodde r and th e possibility o f using it in the dietary of animals. Indian Vet. J. 53: 606-608. Judd, B.I . 1979. Handbook o f tropical forag e grasses. Garland publishing, INC. Kang, B.T., L. Reynolds and A.N. Atta-Krah. 1990. Alley farming. Advances in agronomy 43: 315-359. Karim, A.B., E.R. Rhodes and P.S. Savill. 1991. Effect of cutting height and cutting interval on dry matter yield of Leucaena leucocephala (Lam) De wit. Agroforestry systems 16: 129-137. Kidd, C.V . an d D . Pimente l [Eds] . 1992 . Integrate d resourc e management: Agroforestry for development. Academic press inc. King, K.F.S. 1987. The history of agroforestry. In H.A. Steppler and P.K.R. Nair [Eds]. Agroforestry: a decade for development. ICRAF, Nairobi, Kenya, pp. 3-11. 107 Larbi, A., D. Thomas and J. Hanson. 1993a. Forage potential of Erythrina abvssinica; Intake, digestibility and growth rates for stall-fe d shee p an d goat s i n souther n Ethiopia . Aaroforestry systems 21: 263-270. Larbi, A., M.A. Jabbar, E.J. Orok, N.B Ichon g and J. Cobbina. 1993b. Alchornea cordifolia. a promising indigenou s browse species adapted to acid soils in south eastern Nigeria for integrated crop-livestoc k agroforestr y productio n systems. Aqroforestry systems 22: 33-41. Lefroy, E.C., P.R. Dann, J.H. Wildin, R.N. Westley-smith and A.A. McGowan. 1992 . Trees an d Shrub s a s source s o f fodde r i n Australia. Agroforestry systems 20: 117-139. Le Houerou, H.N. 1980. Chemical composition and nutritive value of browse i n tropical wes t Africa. In H.N. L e Houerou [Ed]. Browse in Africa. The current state of knowledge. ILCA, Addis Ababa, Ethiopia, pp. 261-289. Le Houerou , H.N . 1987 . Indigenou s shrub s an d tree s i n th e silvopastoral systems of Africa. In H.A. Steppler and P.K.R. Nair [Eds] . Agroforestry: A decad e o f development, ICRAF, Nairobi, Kenya, pp. 139-156. Lindsay, W.L. 1978. Chemical reactions affecting the availability of micro-nutrients in soils. InC.S. Andrew and E.J. Kamprath [Eds]. Minera l nutritio n o f legume s i n tropica l an d subtropical soils: Proceedings of a workshop held at CSIRO Cunningham laboratory , Brisbane , Australia-Januar y 16-21 , 1978. Little, D. 1985. The dietary mineral requirements of ruminants: Implications o f th e utilizatio n o f tropica l fibrou s agricultural residues. In Doyle, P. [Ed]. The utilization of fibrous agricultural residues as animal feeds: Proceedings of fourth annua l worksho p o f th e Australian-Asia n fibrou s agricultural residues research network. Lohan, O.P., D. Lall , R.N. Pal and S.S. Negi. 1980 . Cell wall constituents and in vitro dry matter digestibility of some fodder trees in Himachal pradesh. Forage research 6: 21-27. London, J.R. [Ed]. 1984. Booke tropical manual: A handbook for soil survey an d agricultura l lan d evaluatio n i n th e tropics . Longman Inc. pp. 231-250. Lowry, J.B. , maryant o an d B . Tangendjaja . 1983 . Autolysis o f Mimosine t o 3-hydroxy-4-l(H)pyridon e i n gree n tissue s o f Leucaena leucocephala. J. Sci. Food Agric. 34: 529-533. 108 Maclaurin, A.R., D.I. Bransby and N.M. Tainton. 1982. The influence of cutting height, planting density, harvesting frequency and fertilization o n edibl e dr y matte r yiel d o f Leucaen a leucocephala (Lam.) De wit. Grassland society of south Africa 17: 61-64. Mathuva, M., J.F.M. Onim , K. Otien o an d H.A . Fitzhugh . 1987. Productivity of grass-sesbania intercrop as compared to grass-nitrogen interaction : Proceeding s o f th e sixt h Keny a Veterinary association/Small Ruminant CRSP workshop, November 4, 5 and 6, 1987. pp. 81-92. McDowell, L.R., J.H. Conrad, G.L. Ellis and J.K. Loosli . 1983. Minerals for grazing ruminants in tropical regions. University of Florida, Gainesville, Florida, pp.86. Milford, R. an d Minson, D.J. 1966 . Intake of tropical pasture species. Proceeding s o f th e 9 th Internationa l grasslan d congress, Sao Paulo, pp. 815-822. Minson, D.J. 1972. The digestibility and voluntary intake by sheep of six tropical grasses. Australian journal of experimental agriculture and animal husbandry 12: 21-27. Minson, D.J. 1981 . Nutritional differences between tropical and temporate pastures . I n Morley, F.H.W . [Ed] . World anima l science. Bl. Grazing animals. Elsevier scientific publishing co. New York. Minson, D.J. 1990 . Forage in ruminant nutrition. Academic press Inc. Minson, D.J. an d M.P. Hegarty. 1984. Toxic factors in tropical legumes. In R.F. Barnes, P.R. Ball, R.W. Brougham, G.C. Marten and D.J. Minson [Eds] . Forage legumes for Energy-efficient animal production. USDA, ARC, CSIRO, DSIR, PP. 246-250. Moseley, G. 1974. British Journal of nutrition 32: 317-326. Nair, P.K.R., E.C.M Fernandes and P.N. Wambugu. 1984. Multipurpose leguminous trees and shrubs for agroforestry. Agroforestry systems 2(3):145-163. Nangole, F.N. , H . Kayongo-Mal e an d A.N . Said . 1983 . Chemica l composition, digestibility and feeding value of maize cobs. Animal feed science and technology 9: 121-130. National Academy of Sciences. 1977. Leucaena: promising forage and tree crop for the tropics. Washington, D.C. 109 National Academy of Sciences. 1979. Tropical legumes: Resources for the future, National academy press, Washington, D.C., 331 PP. National Research Council. 1983. Calliandra: A versatile small tree for the humid tropics. National academy of sciences press, Washington, D.C. National Research Council. 1984. Leucaena: promising forage tree crop for the tropics. Second edition. National academy press, Washington, D.C. Neter, J. and W. Wasserman. 1974. Appied linear statistical models: Regression, Analysis of variance and experimental designs. Richard D. Irwin, Inc., Illinois. Njwe, R.M. an d J . Kom . 1988 . Survey o f the mineral statu s of pastures and small ruminants in the west region of Cameroon. Tropicultura 6(4): 150-152. Nyirahabimana, P . 1985 . Les Arbres Fourrager s e t Fruitier s a u Rwanda. In Compte-Rendu des iournees d1Etudes Forestieres et Aqroforestieres. 9-12 October, 1984. Ruhande, Rwanda: Institut des sciences agronomigues du Rwanda. Okali, C . an d J.E . Sumberg . 1985 . Sheep an d goats : Househol d relations and Small Ruminant development in southwest Nigeria. Agricultural systems 18: 39-59. Onwuka, C.F.I. 1992. Tannin and saponin contents of some tropical browse species fed to goats. Trop. Aqric. (Trinidad) 69(2). Osman, A.M. 1980 . Protein fro m crops fo r cattle production i n Mauritius with emphasis on Acacia, Leucaena leucocephala (Lam de wit) . In Peerally, A et al. [Eds] . Proceeding s of the National agricultura l productio n conference , pp . 387-406. University of Mauritius, Reduit, 10-15 December 1979, Reduit, Mauritius. Otieno, H.J.O., A.M. Heineman, E.K. Mengich and B. Amadalo. 1991. AFRENA on-station Maseno, Kenya progress report for the year 1990. AFRENA report no. 41, ICRAF, Nairobi, Kenya. Otieno, J.H.O., B. Amadalo and A.M. Heineman (Draft AFRENA report, 1992). AFRENA Maseno progress report for the period January, 1991 to January, 1992. KEFRI, KARI, ICRAF. Panjaitan, M. 1988. Nutritive value of tree legumes introduced in Indonesia. IARD Journal 10(3): 73-80. 110 Parkinson, J.A. and Allen, S.E. 1975. A wet oxidation procedure suitable fo r th e determinatio n o f nitroge n an d minera l nutrients in the biological material. Communication in soil science and plant analysis 6: l-ll. Pathak, P.S., p. Rai and R.D. Roy. 1980. Forage production from koo-babool (Leucaena leucocephala (Lam.) De wit.). 1. Effect of plan t density , cuttin g intensit y an d interval . Forage research 6: 83-90. Phiri, D.M., B. Coulman, H.A. Steppler, C.S. Kamara and F. Kwesiga. 1992. Th e effec t o f brows e supplementatio n o n maize husk utilization by goats. Agroforestry systems 17: 153-158. Pomeranz, Y. an d C.E . Meloan. 1987 . Food analysis : Theory and practice; second edition, Av. publ. Co. Prabowo, A., J.E. Van Eys, I.W. Mathius and L. Soekanto. 1983. Trace mineral status and effect of mineral supplementation in Javanese thin-tail sheep. Proceedings Vth world conference on Animal production. Tokyo, Japan 2: 389-390. Pratchett, D. 1989. The prospects for Leucaena on the ord. J. Dept. aaric. western Australia 30: 2. Pratchett, D. , R.J . Jone s an d F.X . Syrch . 1991 . Use o f DHP -degrading rume n bacteri a t o overcom e toxicit y i n cattl e grazing irrigated Leucaena pasture. Tropical grasslands 25: 268-274. Preston, T.R. 1982 . A strategy for the efficient utilization of crop residue s an d agro-industria l by-product s i n anima l production based on their nutritional constraints. In Crop residues and agro-industrial by-products in animal feeding. Proc. o f th e ILCA/FA O workshop , Dakar , Senegal , 21-2 5 September, 1982. Pritchard, D.A. , G . Weide r an d P.S . Hopkins . 1985 . Us e o f Polyethylene glycol to improve the nutritive value of Mulga (Acacia aneura). In Recent advances in animal nutrition in Australia. Univ. of New England, Armidale. Rees, M.C. , D.J . Minso n an d F.W . Smith . 1974 . The effec t o f supplementary an d fertilize r sulphu r o n voluntary intake , digestibility, retentio n tim e i n th e rume n an d sit e o f digestion of Pangola grass in sheep. Journal of agricultural science, camb., 82: 419-422. Sampet, C . an d V . Pattaro . 1987 . Sesbani a sesban , Gliricidi a maculata an d Leucaen a leucocephal a a s fodde r crop . Tha i iournal of agricultural science 20(4): 303-313. Ill Sands, M.W., H.A. Fitzhugh and S. Chema. (Unpublished data, 1982). Seasonal variation in feed resources on small farms of western Kenya. Paper presented at the 74 th annual meeting of American society o f anima l science , Augus t 8-11 , Guelph , Ontario , Canada. SAS. 1985 . User's guide . Statistical analysi s system s institut e INC, Cary , North Carolina, USA. Satjapradja, 0 . an d T . Sukandi . 1981 . Agroforestr y wit h re d Calliandra (Calliandr a calothyrsus). Indonesian agricultural research and development journal 3(3): 85-88. Scherr, S.J . 1992 . Th e rol e o f extensio n i n agroforestr y development: evidence from western Kenya. Agroforestry systems 18: 47-68. Scherr, S.J. and Alitsi, E. [Eds] (Unpublished data) . The impact of the Care agroforestry extension project. Report to extension staff, ICRAF and CARE. Schneider, B.H . an d W.P . Flatt . 1975 . Th e evaluatio n o f feed s through digestibility experiments. Athens, Georgia, University of Georgia press. Semenye, P.P . 1990 . Toxicity respons e o f goat s fe d o n Leucaen a leucocephala forage only. Small ruminant research 3: 617-620. Sheikh, M.I. and Khalique, A. 1982 . Effect of tree belts on the yield o f agricultura l crops . Pakista n journa l o f Forestr y 32:21-23. Sheldrick, R.D. and D.M . Thairu . 1974. Seasonal gras s growth a t Kitale. Pastur e Researc h Project , Technica l repor t no . 3 . Kitale: Nationa l Agricultura l Station , Ministr y o f Agriculture, Kenya. Siebert, B.D., R.A. Hunter and P.N. Jones. 1976. The utilization by beef cattl e o f sugarcan e supplemente d wit h anima l protein, plant protein or non-protein nitrogen and Sulphur. Austr. J. Exp. Aaric. Anim. Husb. 16: 789-794. Snyders, P.J.M. 1991 . Fodder managemen t an d th e protei n ga p o n zero-grazing farms based on Napier grass. Introduction for the NDDP fodder tree seminar, Nakuru, Kenya. Sobale, B.N., S.T. Kharat, V.L. Prasad, A.L. Joshi, D.V. Rangnekar and S.S . Deshmukh . 1978 . Nutritiv e valu e o f Leucaen a leucocephala for growing Bull calves. Trop. Anim. Hlth. Prod. 10: 237-241. 112 Somarriba, E. 1992. Revisiting the past: an essay on agroforestry definition. Agroforestry systems 19(3): 233-240. Takahashi, M. and J.C. Ripperton. 1949. Koahaole (Leucaena qlauca). its establishment, culture and utilization as a forage crop. University o f Hawai i agricultura l experimenta l station , Bulletin 100, 59 pp. Tangendjaja, B., J.B. Lowry and R.B.H. Wills. 1986. Changes in mimosine, phenol , protei n an d fibr e conten t o f Leucaen a leucocephala leaf during growth and development. Austr. J. Exp. Agric. 26: 315-17. Torres, F. 1983. Role of woody perennials in animal agroforestry. Agroforestry systems 1(2): 131-163. Turkel, T . an d R . Hatipoglu . 1989 . Th e effect s o f intra-ro w spacings an d cuttin g height s o n th e yiel d o f Leucaen a leucocephala i n Adana, Turkey. Journal of range management 42(6): 502-503. Van Eys J.E., I.W. Mathius, P. Pongsapan and W.L. Johnson. 1986. Foliage of the tree legumes Gliricidia, Leucaena and Sesbania as supplement to napier grass diets for growing goats. Journal of agricultural science 107: 227-233. Van Soest, P.J. 1963. Use of detergents in the analysis of fibrous feeds. II. A rapid method for the determination of fibre and lignin. Journal of the A.O.A.C. 46(5): 829-835. Van Soest P.J. 1990. Fibre utilization. In Proceedings of the 26th nutrition conference for feed manufacturers, April 24 and 25, Guelph, Ontario, Canada. Van Soest, P.J. an d R.H. Wine. 1967. Use of detergents i n the analysis of fibrous feeds. IV. Determination of plant cell-wall constituents. Journal of the A.O.A.C. 50(1): 50-55. Van Soest, P.J., J.P. Robertson and B.A. Lewis. 1991. Symposium: Carbohydrate mthodology , metabolis m an d nutritiona l implications i n dair y cattle . Methods fo r dietar y fiber , Neutral detergen t fibe r an d Nonstarc h polysaccharide s i n relation to animal nutrition. Journal of dairy science 74: 3583-3597. Waldern, D.E. 1971. A rapid micro-digestion procedure for neutral and acid detergent fibre. Canadian J. Anim. Sci. 51: 67. Whiteman, P.C. and A. Lulham. 1970. Seasonal changes in growth and nodulation of perennial tropical pasture legumes in the field. 2. Effect of controlled defoliation of Desmodium intortum and Phaseolus atropurpurens. Australian iournal of agricultural research. 21: 195-206. 113 Wills, R.B.H and B. Tangendjaja. 1981. Effect of pH and temperature on the degradation of Mimosine and 3-Hydroxy-4(1H)-pyridone. Phytochemistry. 20(8): 2017-2018. Wilson, A.D. 1969. A review of browse in the nutrition of grazing animals. Journal of range management 22: 23-28. Wilson, R.K., T.A. Sillane and M.J. Clancy. 1966. The influence of fibre content on herbage intake by ruminants. Irish journal of agricultural research 5: 142-143. Wolfgang Bayer , G . 1990 . Napie r grass-Apromisin g fodde r fo r smallholder livestoc k productio n i n th e tropics . Plan t research and development 31: 103-111. Yates, N.G. and T. Panggabean , 1985. The performance o f goats offered Elephan t gras s (Pennisetu m purpureum ) with varied amounts of Leucaena or Concentrate. Tropical grasslands 22(3) : 126-131 Young, A . 1989 . Agroforestr y fo r soi l conservation . C.A. B International. 114 APPENDIX 1. Total rainfall received at harvesting and field data collection dates. Harvesting Harvestin g date Tota l rainfall date no. (Month ) (mm ) May, 1991 July, 1991 September, 1991 November, 1991 January, 1992 March, 1992 May, 1992 July, 1992 September, 1992 November, 1992 495.8 209.2 212.8 382.5 71.0 130.2 413.6 514.9 205.8 359.4 APPENDIX 2. Layout of the experiment THE EXPERIMENT 115 T4 T 5 T l T 2 T 3 T5 T3  T l T 4 T 2 Rl R2 T3 T4  T l T 2 T 5 R3 T2 T 4 T 5 T l T3 R4 T = Treatment (plot ) R  = Replicate THE PLOT II * * * * * * * * * * * * * * * || *** Grass or shrub || 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 || 000 Grass or shrub li i l 116 APPENDIX 3. Dry matter and nutrient composition of Leucaena leucocephala at different harvesting dates. Date DM% CP % Ca % Mg % K % P % Z n C u Adf % ppm pp m 1 2 3 4 5 6 7 8 9 1 0 2 9 . 3 0 2 8 . 8 2 4 1 . 1 6 -3 0 . 6 0 3 0 . 3 3 3 2 . 9 3 3 2 . 7 8 2 9 . 7 5 3 0 . 2 2 7 . 8 2 5 . 2 2 4 . 7 2 6 . 2 3 0 . 8 2 6 . 7 2 4 . 5 2 9 . 0 0 . 7 6 9 0 . 6 9 2 0 . 6 1 3 0 . 7 2 2 0 . 9 3 3 0 . 6 9 8 0 . 8 5 9 0 . 7 6 0 0 . 7 7 0 0 . 3 4 6 0 . 2 9 3 0 . 2 7 2 0 . 3 1 5 0 . 3 6 0 0 . 2 8 1 0 . 3 5 5 0 . 2 7 2 0 . 3 1 9 1 . 6 9 1 . 6 5 1 . 6 3 1 . 7 0 1 . 9 0 1 . 7 4 1 . 5 9 1 . 1 5 1 . 5 6 0 . 1 8 1 0 . 1 9 8 0 . 1 7 9 0 . 1 5 3 0 . 1 7 5 0 . 1 8 3 0 . 1 6 4 0 . 1 9 4 0 . 1 7 7 1 9 . 1 2 0 . 8 1 7 . 5 1 8 . 4 2 0 . 2 1 8 . 1 1 9 . 3 1 9 . 5 2 0 . 0 8 . 9 6 . 8 6 . 6 8 . 8 1 3 . 5 8 . 8 7 . 8 8 . 6 9 . 0 2 5 . 3 2 8 . 9 2 3 . 0 2 8 . 0 3 4 . 1 3 1 . 4 2 9 . 8 3 2 . 1 3 2 . 1 Mean 31.96 27. 2 0.75 7 0.31 3 1.6 2 0.17 8 19. 2 8. 8 29. 4 117 APPENDIX 4. Dry matter and nutrient composition of Calliandra calothyrsus at different harvesting dates. Date DM % CP % Ca % Mg % K % P % Z n C u Adf % ppm pp m 0 . 2 0 3 3 1 . 0 9 . 8 5 1 . 9 0 . 2 1 6 3 2 . 3 7 . 3 5 4 . 4 0 . 1 7 0 2 7 . 0 6 . 7 3 9 . 8 0 . 1 6 5 2 4 . 5 5 . 1 5 6 . 1 0 . 1 6 1 2 6 . 6 1 0 . 5 5 7 . 3 0 . 1 7 9 2 7 . 0 9 . 8 5 4 . 3 0 . 1 9 4 2 6 . 8 9 . 3 5 9 . 0 1 . 3 8 0 . 1 8 1 2 7 . 4 9 . 6 5 5 . 6 0 . 1 6 5 2 5 . 5 9 . 1 5 8 . 6 Mean 38.4 5 21. 7 0.55 2 0.21 8 1.0 8 0.18 2 27. 6 8. 6 54. 1 1 2 3 4 5 6 7 8 9 10 3 6 . 3 8 3 4 . 0 4 5 0 . 4 2 -4 1 . 0 4 3 0 . 0 5 3 8 . 1 4 3 7 . 8 5 3 9 . 6 6 2 4 . 7 2 3 . 7 1 9 . 5 1 8 . 9 2 0 . 6 2 5 . 3 2 2 . 6 1 9 . 5 2 0 . 5 0 . 6 1 1 0 . 4 6 6 0 . 4 5 8 0 . 3 9 5 0 . 6 0 1 0 . 6 3 3 0 . 6 6 3 0 . 5 6 8 0 . 5 6 9 0 . 2 4 2 0 . 2 0 8 0 . 1 7 9 0 . 1 8 7 0 . 2 3 3 0 . 2 3 1 0 . 2 6 1 0 . 2 0 7 0 . 2 1 8 1 . 0 8 1 . 1 8 1 . 0 3 0 . 9 8 1 . 0 4 1 . 0 2 1 . 1 1 .  0 . 9 1 118 APPENDIX 5. Dry matter and nutrient composition of Pennisetum purpureum at different harvesting dates. Date DM % CP % Ca % Mg % K % P % Z n C u Adf % ppm pp m 1 1 7 . 0 9 9 . 9 0 .25 6 0 .32 0 2 . 2 7 0 .16 3 2 5 . 4 6 . 0 4 2 . 8 2 1 7 . 4 4 1 0 . 5 0 .25 9 0 . 3 9 1 2 . 0 9 0 . 1 7 1 2 8 . 8 5 . 0 4 3 . 4 3 3 1 . 2 1 1 0 . 9 0 .24 9 0 .43 6 1 .9 3 0 .16 4 2 9 . 2 6 . 5 3 6 . 8 4 -  9 . 7 0 .30 2 0 .37 5 1 .7 7 0 . 1 4 0 1 5 . 8 3 . 3 3 9 . 9 6 2 7 . 1 1 9 . 1 0 .33 0 0 .34 9 1 .6 2 0 . 1 0 0 2 4 . 8 4 . 8 3 8 . 0 7 1 8 . 7 0 1 2 . 5 0 .25 3 0 . 2 9 6 2 . 4 5 0 . 1 4 8 2 1 . 3 6 . 0 3 9 . 5 8 1 8 . 3 4 1 1 . 5 0 .24 7 0 .37 2 2 . 2 2 0 .16 4 2 6 . 8 6 . 6 4 0 . 8 9 2 3 . 0 2 1 1 . 4 0 .27 6 0 .37 8 1 . 6 1 0 . 1 4 1 2 1 . 9 6 . 4 3 7 . 8 10 2 1 . 9 5 9 . 9 0 . 3 0 1 0 .36 6 1 .8 5 0 .13 9 2 1 . 7 3 . 8 3 8 . 3 Mean 2 1 . 8 6 1 0 . 6 0 . 2 7 5 0 . 3 6 5 1 .9 8 0 .14 8 2 4 . 0 5 . 4 3 9 . 7 119 APPENDIX 6. Recommended nutrient content of diets for dairy cattle (NRC, 1989). Nutrient Crude protein, ADF, % Calcium, % Phosphorus, % Magnesium, % Potassium, % Copper, (ppm) Zinc, (ppm) Early lactation % 1 9 21 0.77 0.48 0.25 1.00 10 40 Dry pregnant cows 12 27 0.39 0.24 0.16 0.65 10 40 Mature Bulls 10 19 0.30 0.19 0.16 0.65 10 40 120 APPENDIX 7. Analysis of variance table for Fresh leafy biomass production of L. leucocephala. C. calothyrsus. P. purpureum and their shrub-grass combinations Season D F Typ e I SS Mea n square F  value Pr> F Season 8  4.9509027 7 0.6188628 5 7.6 2 0.000 1 Treatment 4  8.7348149 9 2.1837037 5 26.9 0 0.000 1 SSN*TRT 3 2 1.9061118 7 0.0595660 0 0.7 3 0.845 1 APPENDIX 8. Analysis of variance table for dry leafy biomass production of L. leucocephala, C. calothyrsus, P. purpureum and their shrub-grass combinations Source D F Typ e I SS Mea n square F  value Pr> F Season 6  4.5215187 1 0.7535864 5 10.7 2 0.000 1 Treatment 4  2.8762838 6 0.7190709 6 10.2 3 0.000 1 SSN*TRT 2 4 0.7023907 6 0.0292662 8 0.4 2 0.991 8 121 APPENDIX 9. Results of the separation of means (p values) between pure and mixed treatments when the F values were significan t Nutrient Seaso n L Napier c Napier Leucaen a Calliandr a FLB (kg/lOOm) 1-10 > 0.0125 "0.008 3 0.417 0 "0.002 6 DM (kg/lOOm) 1-1 0 0.064 9 "0.017 5 0.780 7 "0.001 8 1 2 3 4 5 6 7 8 9 10 0.4365 0.6319 <0.0023 0.8692 0.9753 0.3501 0.2804 <0.0431 0.7670 0.9141 0.0968 0.1946 0.1235 0.6242 0.0705 0.9074 0.7192 0.1485 0.1179 0.5659 0.9993 0.9656 0.7207 0.5981 0.6702 0.5951 0.5381 0.9821 0.3032 0.9027 0.4051 0.8617 CP% 1  0.767 4 0.636 7 0.472 5 0.134 1 2 0.786 8 ^.009 3 0.219 0 0.655 1 3 0.439 6 0.689 7 0.071 1 0.209 8 4 0.355 2 0.559 3 0.489 1 0.249 4 122 5 6 7 8 9 10 — 0.9174 0.7738 0.6918 0.0720 0.6257 — 0.8642 0.2699 0.2707 0.1034 0.9627 — 0.0501 0.4752 0.2589 0.2184 0.0933 -0.2908 0.6151 0.9343 ^.0163 0.2618 Ca% 1 2 3 4 5 6 7 8 9 10 0.3519 0.5521 0.6091 0.7628 0.0884 0.9095 0.2762 0.8427 0.2230 0.4551 0.1102 0.1363 0.3089 0.7438 0.0743 0.9940 0.3953 0.5767 0.3112 0.6773 0.7944 <0.0687 0.5853 0.5015 0.1530 0.1860 0.8195 0.9889 0.7660 0.8403 0.6357 0.3219 0.3944 0.8933 "0.0243 0.8119 K% 1 2 3 4 5 6 0.2933 0.0787 >0.0001 >0.0423 >0.0006 0.4738 >0.0100 >0.0004 ^.0122 >0.0001 0.8541 0.7115 0.9698 0.8795 0.6328 0.8481 0.9535 0.7227 0.3971 0.9492 7 8 9 10 0.1312 0.2343 "0.0288 "0.0333 "0.0205 0.0742 "0.0032 "0.0399 0.9588 0.9759 0.2818 0.7568 123 0.1441 0.9593 *0.0213 0.6500 Mg% 1 2 3 4 5 6 7 8 9 10 0.3569 0.4917 <0.0394 0.7489 0.0652 0.1352 0.1238 0.1148 <0.0036 0.9652 0.0704 <0.0336 <0.0101 <0.0128 0.0991 ^.0163 ^.0315 <0.0284 0.1585 0.9987 0.6531 0.0667 0.4626 0.7446 0.6209 0.6596 0.5832 0.8492 0.7235 0.8592 0.5835 0.8736 0.3219 0.5814 0.1384 0.9891 p% 1 2 3 4 5 6 7 8 0.8217 0.5193 0.7790 0.2142 0.8785 0.9058 0.8858 0.9670 0.1311 0.8818 0.0788 0.3692 0.2884 0.9121 0.5721 0.1488 0.2812 0.3818 0.3293 0.9992 0.8911 0.3364 0.6638 0.4061 0.1068 0.0849 0.6650 0.3070 Zn (ppm) 9 10 1 2 3 4 5 6 7 8 9 10 0.2488 0.7124 0.1603 0.3583 0.0966 0.9750 0.2913 0.2943 0.1073 0.6492 0.4947 0.9524 0.2581 0.6926 0.2469 0.9336 0.1186 0.8645 0.2634 0.2193 0.4098 0.2525 0.6611 0.9655 0.0628 0.5021 0.5938 0.8674 0.8489 0.2297 0.7316 0.9566 0.6909 124 0.0882 0.2093 0.9807 0.4943 0.6890 0.7483 0.9880 0.7781 0.8528 0.8499 0.1375 Cu (ppm) 1 2 3 4 5 6 7 8 9 10 0.0958 0.7780 0.5025 0.5104 0.5603 0.4708 0.5723 0.5427 0.6956 0.6632 0.2688 0.8220 0.1126 0.5603 0.2376 -0.5427 0.7693 0.7140 0.1783 0.4056 0.2203 ^.0123 0.2606 0.2216 0.5018 0.8162 0.5835 0.1783 -0.8340 0.6123 0.2922 0.4774 0.7152 0.3590 125 ADF% 1 2 3 4 5 6 7 8 9 10 0.6822 0.5769 0.2098 0.8109 0.7953 0.6458 0.5732 0.5822 0.5462 0.8584 0.4263 0.5354 0.0515 >0.0293 0.9929 0.4594 0.8122 0.6016 0.2415 0.0632 0.8801 0.2530 0.6379 0.2971 0.4656 0.4673 0.0934 0.3722 0.5311 0.0842 0.9031 0.9819 *0.0373 0.7196 0.7436 0.4770 L Napier grass in combination with leucaena c Napier grass in combination with calliandra * Significantly lower than in monoculture * Significantly higher than in monoculture 126 APPENDIX 10. A short description (FAO, 1990; London, 1984) of the soil types common to western Kenya SOIL NAME BRIEF DESCRIPTION Ferralsols: Acrisols: Strongly weathered soils of the humid tropics with oxic horizons. CEC < 16 me/100 g clay. Acid low base status lessives; more strongly leached than luvisols, but insufficiently weathered for ferralsols. Base saturation of B horizon <50% (cf luvisols) Luvisols: Lessive s - soils having argillic B horizons, with high base status. Base saturation of B horizon >50% (cf acrisols) Lixisols: Soil s having an argillic B horizon which has a CEC < 24 cmol (+) kg"1 clay .Base saturation of B horizon >50% 127 APPENDIX 11. Spearman rank order coefficients between different leafy biomass and nutrients in L.  leucocephala FLB DM DM% CP% Ca% Mg% K% P% Zn Cu ADF FLB - 0.857 -.357 0.619 0.167 0.347 0.607 0.262 0.071 0.455 0.192 DM - 0.143 0.107 0.000 0.306 0.829 -.286 -.214 0.143 -.126 DM% - -.667 0.024 -.216 -.107 -.476 -.500 -.333 -.012 CP% - 0.050 0.268 -.024 0.250 -.000 0.293 -.017 Ca% - 0.837 0.071 -.500 0.450 0.695 0.619 Mg% - 0.095 -.577 0.410 0.613 0.282 K% - 0.048 -.071 0.347 0.190 P% - 0.200 -.318 0.000 Zn - 0.276 0.594 Cu - 0.563 ADF * Significant positive correlations existed between FLB and DM (r=0.857), K% and DM (r=0.829), Mg% and Ca% (r=0.837), and Cu and Ca% (r=0.695) 128 APPENDIX 12. Spearman rank order coefficients between different leafy biomass and nutrients in C.  calothyrsus FLB DM DM% CP% Ca% Mg% K% P% Zn Cu AOF% FLB - 0.786 0.000 0.647 0.429 0.214 -.143 -.132 0.263 0.611 -.357 DM - 0.143 0.000 -.357 -.464 -.357 -.321 0.270 0.144 -.821 DM% - -.766 -.333 -.190 -.262 -.667 -.611 -.240 -.167 CP% - 0.736 0.703 0.192 0.529 0.517 0.597 -.201 Ca% - 0.917 0.150 0.234 0.100 0.745 0.267 Mg% - 0.183 0.243 0.059 0.703 0.383 K% - 0.695 0.728 0.243 -.083 P% - 0.840 -.000 -.318 Zn - 0.206 -.619 Cu - 0.042 ADF% * Significant positive correlations between FLB and DM (r=0.785) , Ca% and CP% (r=0.736), Mg% and CP% (r=0.703), Mg% and Ca% (r=0.917)/ Cu and Ca% (r=0.745), Mg% and Cu (r=0.703), K% and P% (r=0.695), Zn and K% (r=0.728), and Zn and P% (r=0.840) . * Significant negative correlations existed between ADF% and DM (r=-0.821), CP% and DM% (r=-0.766), and DM% and P% (R=-0.667) . 129 APPENDIX 13. Spearman rank order coefficients between different leafy biomass and nutrients in P. purpureum FLB DM DM% CP% Ca% Mg% K% P% Zn Cu ADF% FLB - 0.964 -.107 -.120 0.071 -.405 0.429 0.096 -.214 -.275 0.262 DM - 0.036 -.198 -.036 -.286 0.393 0.198 0.179 -.072 -.000 DM% - -.072 0.238 0.333 -.667 -.443 -.024 0.036 -.952 CP% - -.795 0.050 0.485 0.555 0.159 0.752 -.033 Ca% - -.1000 -.717 -.795 -.533 -.845 -.150 Mg% - -.417 0.435 0.500 0.209 -.217 K% - 0.586 0.200 0.326 0.567 P% - 0.720 0.622 0.435 Zn - 0.603 0.067 Cu - -.159 ADF% * Significant positive correlations existed between FLB and DM (r=0.964), CP% and Cu (r=0.752), and P% and Zn (r=0.720). * Significant negative correlations existed between K% and DM% (r=-0.667), DM% and ADF% (r=-0.952), CP% and Ca% (r=-0.795), K% and Ca% (r=-0.717), P% and Ca% (r=-0.795), and Ca% and Cu (r=-0.845) 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items