Citation
MANAGEMENT AND NUTRITIVE EVALUATION OF MUCUNA PRURIENS AND LABLAB PURPUREUS-MAIZE INTERCROPS IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA

Material Information

Title:
MANAGEMENT AND NUTRITIVE EVALUATION OF MUCUNA PRURIENS AND LABLAB PURPUREUS-MAIZE INTERCROPS IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA
Copyright Date:
2008

Subjects

Subjects / Keywords:
Beans ( jstor )
Biomass ( jstor )
Corn ( jstor )
Crops ( jstor )
Defoliation ( jstor )
Farmers ( jstor )
Grains ( jstor )
Legumes ( jstor )
Nutrients ( jstor )
Soils ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright the author. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
2/8/2003
Resource Identifier:
70831855 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

MANAGEMENT AND NUTRITIVE EVALUATION OF MUCUNA PRURIENS AND LABLAB PURPUREUS -MAIZE INTERCROPS IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA By ELKANA M. NYAMBATI A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2002

PAGE 2

Copyright 2002 by Elkana M. Nyambati

PAGE 3

This work is dedicated to my mother whose constant encouragement and enduring love was a source of inspiration a nd strength to pursue knowledge to this day. May God give her the strength and courage to overcome her long-term ill health.

PAGE 4

iv ACKNOWLEDGMENTS I would like to sincerely tha nk Dr. Lynn E. Sollenberger, th e chair of my supervisory committee, for his patient support and scientific guida nce during the course of this study. I appreciate his interest and attention during the initial stag e of the research project and his diligence in reviewing the dissertation. Special gratitude is also extended to my supervisory committee members, Drs. C.K. Hiebsch, W.E. Kunkle, D.M. Sylvia, K.J. Boote, and A.T. Adesogon, for their advi ce and for reviewing the dissertation. The Rockefeller Foundation provided the fina ncial support for both my research work in Kenya and my studies at the University of Florida. I am particularly grateful to Dr. John Lynam of the Nairobi Office, for his advice and logistical arrangements that increased my interaction with other institu tions that provided valuable support. Acknowledgments are extended to the Inte rnational Center for Research in Agroforestry (ICRAF) and Tropical Soil Biol ogy and Fertility Programme (TSBF) for doing most of the soil and plant tissue chemical analysis. Specific ally, I would like to thank Dr. P. Smithson of ICRAF and Dr. C. Palm and Ms. C. Gachengo of TSBF and their staff for ensuring that the samples were handled and analyzed in a most efficient way. The support from NARC, Kitale center Direct or Dr. C. Mwendia and other staff is highly appreciated. I am partic ularly grateful to S. Rono a nd D. Shitandi, who supervised the field activities and data collection while I was here at UF. The contribution of other colleagues in KARI who made useful suggestio ns to the research project, notably Dr.

PAGE 5

v F.N. Muyekho, Dr. E.A. Mukisira, and Mr. J. Ngeny, is greatly appreciated. The support from the Soil Management Project team, particul arly that of Dr. J. Mureithi, the national coordinator, and Ms. T. Mwangi, the team leader, at Kitale is highly appreciated. The willing cooperation of the farmers, Jer ita Nasambu, Teresa Simiyu, Jafred Situma, Silvester Baraza, Raphael Simiyu, Wilson Khauka, and Lona Kufwafwa, is highly appreciated. Without their keen interest in taking care of the expe rimental plots, the collection of the on-farm data would have not be en possible. In addi tion the assistance of the frontline agricultural ex tension staff, Mr. Musuya, is highly appreciated. Lastly, but by no means least, I am grateful to my dear wife, Florence, and children Kelvin, Newton, and Nancy, for their love, enco uragement, support, and particularly for their patience for the long period that I was away in Kenya conducting this research.

PAGE 6

vi TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...............................................................................................................x LIST OF FIGURES.........................................................................................................xiii ABSTRACT.....................................................................................................................xi v CHAPTERS 1 INTRODUCTION...........................................................................................................1 2 LITERATURE REVIEW................................................................................................6 Cereal-Legume Cropping Systems.................................................................................6 Rotations Involving Forage Legumes......................................................................7 Intercropping and Adaptation Characte ristics of Mucuna and Lablab....................9 Replenishing Soil Fertility............................................................................................11 Effects of Mucuna and Lablab Green Manures on Soil and Crop.........................14 Introduction......................................................................................................14 Potential use of mucuna and lablab green manures.........................................15 Effect of Root and Shoot Stubbl e on Soil and Crop Responses............................18 Manure Effects on Soil Fertility Improvement......................................................22 Integrated Nutrient Management...........................................................................24 Residue Quality.............................................................................................................27 Synchrony Between Mineralized N and N Uptake.......................................................31 Mixed Crop-Livestock Farming...................................................................................34 Potential Use of Napiergrass Fodder............................................................................36 Background Information (Botany, Or igin, and Characteristics)............................36 Dry Matter Yield....................................................................................................37 Nutritive Value.......................................................................................................37 Legume Supplementation.............................................................................................38 Forage Yield and Nutritive Value of Mucuna and Lablab...........................................40 Antinutritive Factors..............................................................................................42 Dry Matter Intake and Animal Performance..........................................................43

PAGE 7

vii 3 NITROGEN CONTRIBUTION FROM RE LAY-CROPPED MUCUNA AND LABLAB TO MAIZE IN NORTHWESTERN KENYA ............................................47 Introduction................................................................................................................... 47 Materials and Methods..................................................................................................50 Experimental Site...................................................................................................50 Treatments and Cropping Systems........................................................................51 Green Manure Defoliation Management and Sampling........................................52 Soil Mineral N Sampling.......................................................................................53 Maize N Uptake Sampling.....................................................................................54 Grain and Stover/Straw Dry Matter Yield of Maize and Beans............................54 Chemical Analyses.................................................................................................54 Statistical Analyses................................................................................................55 Results and Discussion.................................................................................................55 Legume Residue Biomass......................................................................................55 Chemical Composition of Legume Biomass.........................................................58 Nutrient Mass Incorporated....................................................................................61 Soil Mineral N........................................................................................................64 Maize Yield and N Uptake.....................................................................................67 Relationships Between Residue Qualit y Parameters and Soil Mineral N..............71 Apparent N Recovery.............................................................................................74 Bean and Lablab Grain Yield.................................................................................77 Maize Grain Yield..................................................................................................79 Conclusions...................................................................................................................8 0 4 PRODUCTIVITY OF MAIZE-BEAN INTERCROP RELAY CROPPED WITH MUCUNA AND LABLAB GREEN MANURES........................................................82 Introduction................................................................................................................... 82 Materials and Methods..................................................................................................84 Experimental Site...................................................................................................84 Experimental Treatments and Layout....................................................................85 Herbage Yield and Chemical Composition............................................................86 Grain and Stover/Straw Dry Matter Yield of Maize and Beans............................87 Statistical Analyses................................................................................................87 Results and Discussion.................................................................................................89 Effect of One Year of Residue Application...........................................................89 Legume Biomass Yield....................................................................................89 Nitrogen Concentration....................................................................................90 Nitrogen Content..............................................................................................93 Bean Grain and Straw Yield............................................................................94 Maize Grain and Stover Yield.........................................................................94 Effect of One versus Two Years of Consecutive Residue Application.................98 Legume Biomass Yield....................................................................................98 Legume N Concentration...............................................................................101 Legume N Content.........................................................................................102

PAGE 8

viii Bean Grain and Straw Yield..........................................................................105 Maize Grain and Stover Yield.......................................................................107 Long-Term Residual Effects of Residue Application..........................................110 Bean Grain and Straw Yield..........................................................................110 Maize Grain and Stover Yield.......................................................................110 Conclusions.................................................................................................................111 5 ON-FARM PRODUCTIVITY OF RELAYCROPPED MUCUNA AND LABLAB IN SMALLHOLDER CROP-LIVESTOCK SYSTEMS IN NORTHWESTERN KENYA.....................................................................................116 Introduction.................................................................................................................11 6 Materials and Methods................................................................................................119 Experimental Site.................................................................................................119 Pre-experimental Activities..................................................................................119 Experimental Treatments and Layout..................................................................120 Green Manure Defoliation Management and Sampling......................................121 Chemical Analysis...............................................................................................123 Statistical Analysis...............................................................................................123 Farmer Evaluation................................................................................................124 Results and Discussion...............................................................................................125 FarmersÂ’ Resource Endowment and Priority Setting...........................................125 Soil and Cattle Manure Characteristics................................................................126 Herbage Mass and Nutritive Value of Top Canopy Herbage..............................128 Legume Residue Biomass....................................................................................128 Legume Residue Nutrient Concentration.............................................................132 Legume Residue Nutrient Content.......................................................................137 Yield Responses...................................................................................................143 Bean Grain and Straw Yield..........................................................................143 Maize Grain and Stover.................................................................................145 Stability Analysis...........................................................................................148 Farmer Evaluation..........................................................................................148 Conclusions.................................................................................................................151 6 NUTRITIVE VALUE OF TOP-CANOPY HERBAGE OF MUCUNA AND LABLAB RELAY CROPPED IN MAIZE IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA...............................................................................153 Introduction.................................................................................................................15 3 Materials and Methods................................................................................................155 Study Site and Treatments...................................................................................155 Top-Canopy Biomass Sampling..........................................................................155 Chemical and Statistical Analysis........................................................................156 Results and Discussion...............................................................................................157 Mass and Plant-part Proportions..........................................................................157

PAGE 9

ix Chemical Composition.........................................................................................158 Conclusion..................................................................................................................162 7 FEED INTAKE AND LACTATION PE RFORMANCE OF DAIRY COWS OFFERED NAPIERGRASS SUPPLEMEN TED WITH LEGUME HAY................164 Introduction.................................................................................................................16 4 Materials and Methods................................................................................................166 Production Environment......................................................................................166 Experimental Diets...............................................................................................166 Experimental Animals..........................................................................................167 Experimental Design............................................................................................168 Diets and Feeding Management...........................................................................168 Measurements......................................................................................................169 Sample Preparation and Chemical Analysis........................................................170 Statistical Analysis...............................................................................................171 Results........................................................................................................................ .172 Chemical Composition of Feeds..........................................................................172 Intake, Fecal Output, and Apparent Digestibility................................................172 Milk Yield............................................................................................................175 Milk Composition................................................................................................178 Live-weight Changes and Body Condition Score................................................178 Discussion...................................................................................................................17 8 Chemical Composition of Feeds..........................................................................178 Apparent Digestibility and Intake........................................................................180 Milk Yield and Composition................................................................................182 Conclusions.................................................................................................................183 8 CONCLUSIONS, SYNTHE SIS, AND RECOMMENDATIONS..............................185 APPENDICES A TOTAL MONTHLY RAINFALL AND MEAN MONTHLY TEMPERATURES RECORDED AT NARC–KITALE, KENYA, IN 2000.............................................191 B COMPOSITION OF MU LTIVITAMINS FED TO EXPERIMENTAL COW..........192 C GERBER METHOD FOR BUTTER-FAT DETERMINATION IN FRESH MILK..193 LIST OF REFERENCES.................................................................................................194 BIOGRAPHICAL SKETCH...........................................................................................219

PAGE 10

x LIST OF TABLES Table page 3-1. Residue biomass of various fractions of mucuna and labl ab, weeds, and whole residue when legumes were relay cropped in maize...................................................56 3-2. Nitrogen, lignin, and polyphenol concen tration of various residue fractions of mucuna and lablab re lay cropped in maize.................................................................59 3-3. Phosphorus and potassium concentratio ns of various residue fractions of mucuna and lablab relay cropped in maize...............................................................................62 3-4. Nitrogen contribution of various re sidue fractions of muc una and lablab relay cropped in maize.........................................................................................................63 3-5. Phosphorus and K content of various re sidue fractions of muc una and lablab relay cropped in maize.........................................................................................................65 3-6. Total above-ground maize bioma ss yield at different sampling dates........................68 3-7. Nitrogen taken up by ma ize at different sampling dates.............................................70 3-8. Nitrogen recovery by ma ize at various sampling dates...............................................76 3-9. Maize and bean grain and stover /straw yield for the 2000 growing season................78 4-1. Outline of treatment arrangement showing crop combinations, cropping system sequences, and legume defoliation regime..................................................................88 4-2. Effects of legume and defoliation on re sidue biomass of mucuna and lablab relay cropped in maize for 1 yr............................................................................................91 4-3. Residue N concentration of muc una and lablab relay cropped in maize.....................92 4-4. Residue N content of mucuna and lablab relay cropped in maize...............................95 4-5. Grain and straw yield of common bean intercropped in maize after one year of mucuna and lablab residue incorporati on. Data are means across two seasons (2000 and 2001)...........................................................................................................96

PAGE 11

xi 4-6. Grain and stover yield of maize afte r one year of mucuna and lablab residue application. Data are means across two seasons (2000 and 2001).............................97 4-7. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue biomass of mucuna and lablab relay cropped in maize during the 2000/2001 season.......................................................................................100 4-8. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue N concentration of mucuna and lablab relay cropped in maize.......................................................................................................................... .103 4-9. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue N content of muc una and lablab relay cropped in maize........104 4-10. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation of mucuna and lablab gr ain and straw yield of common bean intercropped in succeeding maize...............................................................................106 4-11. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation of mucuna and lablab on grai n and stover yield of succeeding maize......109 4-12. Residual effects of mucuna and labl ab residue application in March 2000 on grain and straw yield of common bean inte rcropped in succeeding maize in 2001.............112 4-13. Residual effects of mucuna and labl ab residue application in March 2000 on grain and stover yield of succeeding maize in November 2001...........................................113 5-1. Nutrient concentration of cattle manur es from smallholder farms and a large-scale farm used in the experiment........................................................................................127 5-2. Biomass and nutritive value of the top-canopy (above a 10-cm stubble) herbage of mucuna and lablab relay croppe d in maize on farmersÂ’ fields in Tumaini at Kitale, Kenya. Means are across farms (n = 6)......................................................................129 5-3. Nitrogen concentration of various muc una and lablab residue fractions at time of soil incorporation in 2 yr.............................................................................................133 5-4. Phosphorus and K concentrations of va rious mucuna and lablab residue fractions at time of soil incorporation during the 1999/2000 season.............................................135 5-5. Calcium and Mg concentration of vari ous mucuna and lablab residue fractions at time of soil incorporation during the 1999/2000 season.............................................136 5-6. Nitrogen content of mucuna and lablab residue fractions at time of soil incorporation in 2 yr....................................................................................................138 5-7. Phosphorus and K content of various fr actions of mucuna and lablab relay cropped in maize on farmersÂ’ fields in 1999/2000....................................................................140

PAGE 12

xii 5-8. Calcium and Mg content of various fr actions of mucuna and lablab relay cropped in maize on farmersÂ’ fields in 1999/2000....................................................................142 5-9. Mean grain and straw yield of common bean relay cropped in maize after mucuna and lablab residue incorporation on farmersÂ’ fields for 2 yr.......................................144 5-10. Mean grain and stover yield of subse quent maize after relay cropped mucuna and lablab residue incorporation on farmersÂ’ fields for 2 yr..............................................146 5-11. Farmer ranking (1 = highest) of the gr een manures for suitability in improving soil fertility and providing fodder......................................................................................150 5-12. Ranking by farmers of the performance of maize in various treatment plots.............151 6-1. Herbage dry matter mass and plant-pa rt proportions of various fractions of defoliated mucuna and lablab relay cropped in maize for 2 yr...................................159 6-2. Nutritive value of top-canopy bioma ss of mucuna and lablab relay cropped in maize during the 1999/2000 growing season at NARC-Kitale...................................160 6-3. Crude protein concentration of t op-canopy biomass leaf and stem fractions of mucuna and lablab relay cropped in ma ize during the 2000/2001 growing season at NARC, Kitale..............................................................................................................161 7-1. Chemical composition of f eeds fed to experimental cows..........................................173 7-2. Intake, fecal output, and apparent digestibility when Friesian cows were fed napiergrass alone or supplemented with legume hay or dairy meal............................174 7-3. Crude protein intake (CPI) and digestib le energy intake (DEI) of Friesian cows fed a basal diet of napiergrass alone or supplemented with legume hay or dairy meal....176 7-4. Milk production, body condition score (B CS), and body weight gain (BW Gain) of Friesian cows fed napiergr ass alone or supplemented w ith legume hay or dairy meal........................................................................................................................... ..177 7-5. Milk composition of Friesian cows fed napiergrass alone or supplemented with legume hay or dairy meal............................................................................................179

PAGE 13

xiii LIST OF FIGURES Figure page 3-1. Mean inorganic N in the soil at different time periods.................................................. 66 3-2. Relationships between soil mineral N and residue lignin, li gnin-to-N ratio, and (lignin + polyphenol)-to-N ratio at 4 WAP.................................................................. 72 3-3. Relationship between residue N concen tration and soil mineral N at 4 WAP.............. 73 5-1. Mass of various residue fractions of mucuna and lablab relay cropped in maize on farmersÂ’ fields during the 1999/2000 grow ing season. Treatments are undefoliated mucuna (UD-M), defoliated mucuna (D -M), undefoliated lablab (UD-L), and defoliated lablab (D-L). P values for legume, defoliation, and legume x defoliation effects, respectively are as follows: total residue, < 0.001, < 0.001, and 0.003; leaf + stem, < 0.001, < 0.001, and 0.003; litter, 0.879, 0.693, and 0.347; roots, < 0.001, 0.585, and 0.212............................................................................................................130 5-2. Mass of various residue fractions of mucuna and lablab relay cropped in maize on farmersÂ’ fields during the 2000/2001 grow ing season. Treatments are undefoliated mucuna (UD-M), defoliated mucuna (D -M), undefoliated lablab (UD-L), and defoliated lablab (D-L). P values for legume, defoliation, and legume x defoliation effects, respectively are as follows: total residue, < 0.001, 0.373, and 0.351; leaf + stem, 0.001, 0.015, and 0.111; litter, 0.004, 0.412, and 0.431; roots, < 0.001, 0.939, and 0.772...................................................................................................................... .131 5-3. Stability analysis of maize grain yield (t ha-1) after incorporati ng undefoliated or defoliated relay cropped mucuna and lablab on farmersÂ’ fields at Kitale, Kenya. A low environmental index is associated with low soil fertility. Treatments are undefoliated mucuna (UD-M), defoliated mu cuna (D-M), undefoliated lablab (UDL), and defoliated lablab (D-L).....................................................................................149

PAGE 14

xiv Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MANAGEMENT AND NUTRITIVE EVALUATION OF MUCUNA PRURIENS AND LABLAB PURPUREUS -MAIZE INTERCROPS IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA By Elkana M. Nyambati August 2002 Chair: Lynn E. Sollenberger Major Department: Agronomy Declining soil fertility and inadequate and low quality feed resources limit smallholder crop yields and dairy production in Kenya . Herbaceous legumes can provide an alternative to the use of commercial N sour ces for cereal crops and livestock production in these low external-input farming systems. Research reported in this dissertation was conducted to 1) determine the N contri bution from relay-cropped mucuna [ Mucuna pruriens (L.) DC. Var. utilis (Wright) Bruck] and lablab [ Lablab purpureus (L.) Sweet cv. Rongai] to succeeding maize ( Zea mays L.)-common bean ( Phaseolus vulgaris L.) intercrop when part of the legume biomass is removed for fodder and 2) evaluate the nutritive value of these legumes when used as supplements for lactating cows fed napiergrass [ Pennisetum purpureum (K. Schum) cv. Bana]. Relay-cropped mucuna and lablab survived the dry season producing a total biomass yield of 4 and 2.7 t ha-1 on the research station, and 2.3 and 0.75 t ha-1 on farmersÂ’ fields, respectively. The N contribution from muc una and lablab residues were 78 and 57 kg

PAGE 15

xv ha-1 on station, and 49 and 12 kg ha-1 on farmersÂ’ fields. Defoliation of the legumes to a 10-cm stubble removed 52 and 76% of the a bove-ground herbage of mucuna and lablab, and the residue quality of the remaining stubble was lower (higher lignin and lower N concentration) than that of whole above-ground biomass. Inclusion of the green manures in the maizebean intercrop increased subsequent bean and maize yields compared to the natural fall ow control, but crop yields were relatively low on farmersÂ’ fields where soil fertility was lower than on stati on. Defoliation reduced the nutrient contribution from the legume resi dues; the impact of defoliation was greater on lablab, which has a more upright growth ha bit than mucuna. The mean N recovery by maize across all sampling dates ranged from 21% for the cattle ma nure control to 52% following undefoliated lablab. The N rec overy after defoliated mucuna was 45% compared to 35% for undefoliated mucuna. Afte r 2 yr of consecutive residue application, yields of subsequent maize were greate st under defoliated mucuna and undefoliated lablab, possibly due to their lo wer quality (more stem and li gnin) residue which may have improved synchrony of N release wi th N requirement of the crops. Defoliation of relay-cropped mucuna and lablab to 10 cm provided 1 to 1.8 t ha-1 yr-1 of high quality livestock fodde r. Supplementing mucuna hay and lablab hay to dairy cows fed a basal diet of na piergrass increased total dry matter intake, apparent DM digestibility, and daily yield of 4% fat-corr ected milk compared to napiergrass alone. Utilization of the high nutritive value upper ca nopy of mucuna and lablab as livestock feed and soil incorporating the remaining stubble has potential fo r improving maize yield and performance of livestock in smallholder systems in Kenya.

PAGE 16

1 CHAPTER 1 INTRODUCTION Agriculture is a very important sector of the Kenyan economy, accounting for 70% of employment, 80% of export earnings, and c ontributing 25% of the total gross domestic product (Kenya, 2000). Most ag ricultural production in Ke nya is from smallholder farmers, and in northwestern Kenya, farmer s practice mixed farming where dairying is integrated with the production of maize ( Zea mays L.) intercropped with common bean ( Phaseolus vulgaris L.), in addition to other food crops (National Agricultural Research Center [NARC], 1995). Maize is the most important cereal crop in Kenya, with an average annual production of 3.3 million tons. It is estimated that 60% of the total national maize production occurs in western Kenya and 75 to 80% of this production is from smallholder farms (Ruto, 1992). Maize is ranked fourth out of Kenya Agricultural Research InstituteÂ’s (KARI) 53 commodity and commodity group research priorities (KARI, 1991). Maize in combination with beans is the staple diet for a majority of Kenyans. Maize is interplanted with beans, usually at the beginning of the rainy season in April; some farmers plant a second crop of beans in the maize fields in August. After harvesting maize in November, the land is left fallow during the dry season from November to March. The yields of maize are low and have been decr easing in the last few years on smallholder farms (Rees et al., 1997). The major cr op production constraint facing smallholder farmers is declining soil fertility (NARC, 1995).

PAGE 17

2 It is estimated that more than 18% of the total recorded, marketed agricultural production in Kenya is derived from livesto ck, out of which dairy products account for 33% (Kenya, 2000). Smallholder farmers pr oduce over 80% of the milk sold to the formal market in Kenya (Mbogoh, 1992). Dair y ranks first of the Kenya Agricultural Research InstituteÂ’s 53 commodity and comm odity groups research priorities (KARI, 1991). In northwestern Kenya, the major pr oduction constraint to dairy production is inadequate feed for livestock, particular ly lack of protein during the dry season (Wandera, 1996; Nyambati, 1997). Due to land limitations, farmers practice conti nuous cropping and grazi ng with little or no fertilizer application, which has led to d eclining soil fertility a nd productivity of both crops and livestock. In the intensively cultivat ed highlands of eastern Africa, losses of N, P, and K are estimated at 112, 3, and 70 kg ha-1 yr-1, respectively (Smaling, 1993; Stoorvogel et al., 1993; Van den Bosch et al., 1998). These losses are much higher than the average of 5 to 10 kg ha-1of inorganic fertilizers used in sub-Saharan Africa (FAO, 1995; Heisey and Mwangi, 1996; Larson a nd Frisvold, 1996), leading to nutrientdepleted cropping systems (Sanchez et al., 1997). Soil N and P are the major limiting nutrients (Smaling et al., 1997). Although cy cling of biomass through livestock and use of manure and urine to fertilize soil have b een an important link between livestock and soil fertility (Powell and Vale ntine, 1998), the quantities of manure available on farms are usually not enough to replenish nutrie nts harvested in grain and crop residues (Williams et al., 1995). Intercropping of the common bean with maize provides little or no N to concurrent or subsequent maize, as the majority of N fixed by bean is harvested in the grain (Giller et al., 1991; Gill er et al., 1994; Amijee and Giller, 1998).

PAGE 18

3 Intercropping of soil-improving legume green manures with cereal crops is a promising, low-cost, ecological means of impr oving soil fertility (Giller et al., 1997). Legumes managed as green manures have the po tential to furnish all or part of the N needed by a succeeding non-legume crop (Bowen et al., 1993). Legume cover crops and green manures may contribute from 30 to 60 kg N ha-1 (Utomo et al., 1992) to 110 kg N ha-1 (Hairiah and Van Noordwijk, 1989; Tian et al., 2000) to the subsequent cereal crop. Beneficial effects of legume cover crops for food crop production have been reported from West Africa (Carsky et al., 1999; Tian et al., 1999; 200 0), East Africa (Fischler et al., 1999; Fischler and Wortma nn, 1999; Wortmann et al., 2000) and elsewhere (Burle et al., 1992; Hairiah et al., 1992; Buckles, 1995). Although some adoption of green manures has been reported from West Af rica (Manyong et al., 1996) and Central Africa (Balasubramanian and Blaise, 1993), the adoption of herbaceous legumes in Africa is generally low (Thomas and Sumb erg, 1995; Drechsel et al., 1996). Although green manures often give greater yields of subsequent crops than intercropping with grain legumes, they suffer the handicap of occupying the land unproductively; thus additional benefits are ne cessary for farmers to adopt them (Becker et al., 1995; Versteeg et al., 1998a). It is unlikely that farmers with food security problems, like those of small-holder farmers in western Kenya, will adopt legume green manures purely for soil fertility improvement alone. Apart from improving soil nutrient status the legume herbage ha s high protein concentration, pa latability, and digestibility, and could be useful as a supplement to lives tock being fed mature tropical grasses and cereal crop residues that are of ten of low nutritive value (DÂ’M ello and Devendra, 1995). This would be particularly useful to small holder farmers who lack the financial resources

PAGE 19

4 to purchase commercial concentrates. Th ese two potential uses of green manure legumes, however, compete for the nutrients in the above-ground biomass. Thus, it is important to understand the tradeoff between incorporating all the herbage from green manures into soils or using part of the herbage as livestock feed. Mucuna pruriens var. Utilis (L) DC (Wright) Burck (mucuna or velvetbean) and Lablab purpureus L. (Sweet) cv. Rongai (lablab or dolichos) are promising legume green manures that have been successfully intercr opped with maize in different parts of the world. They have been shown to increase grai n yields of subsequent maize compared to continuously grown maize (Hairiah a nd Van Noordwijk, 1989; Buckles, 1995; Mandimba, 1995; Thomas and Sumberg, 1995; Ib ewiro et al., 1998; Versteeg et al., 1998a; Ibewiro et al., 2000a; Tian et al., 2000). Even incorporation of only the roots of mucuna or lablab had a positive effect on subsequent maize as compared to a control where no residue was applied (Ibewiro et al., 1998). One reason for the success of mucuna and la blab is that they have shown a greater competitive ability than many herbaceous fora ge and grain legumes under the shade of long-season maize cultivars when planted af ter maize (Maasdorp and Titterton, 1997). Mucuna and lablab could fit well in small-s cale farms where the cut and carry system of dairying is practiced. Although information on the contribution of whole-plant biomass incorporation on subsequent mai ze yields is available, info rmation on the contribution of roots or roots plus stem st ubble after removing some shoot biomass for feed is still limited. Also, lack of information on their poten tial as livestock feed and their effects on maize grain yield hamper the integration of these legumes into the farming systems of northwestern Kenya.

PAGE 20

5 The research reported in this dissertation was designed to answer the most critical management questions for use of these legum es among the financially constrained smallholder farmers in western Kenya, not only as dry-season forage but also for soil improvement purposes. The broad objectives of this research were to 1) determine if alternative cropping systems (l egume intercrop and green ma nure) to the current maizebean system affect maize and bean yields , 2) evaluate fodder production and nutritive value of legumes in alternative cropping syst ems and assess their value as a supplement to lactating cows during the dry season, and 3) determine the extent to which harvesting topgrowth of legumes for fodder reduces thei r beneficial impact on soil fertility as measured by yield of subsequent maize and beans.

PAGE 21

6 CHAPTER 2 LITERATURE REVIEW Cereal-Legume Cropping Systems The use of legumes in mixed cropping systems is one of the traditional soil-fertility maintenance strategies. The most common production systems of integrating legumes into cropping systems include the follo wing: simultaneous intercropping, relay intercropping, rotations , and improved fallows (Weber, 1996). The use of legumes in cropping systems offers considerable benefits because of their ability to fix atmospheric N2 (Weber, 1996; Giller et al ., 1997). Two mechanisms ha ve been postulated by which cereal crops benefit from legumes in these multiple cropping systems (Giller et al., 1991; Giller et al., 1994; Giller and Cadisch, 1995): (1) through immediate transfer in which N travels from the legume directly to the asso ciated crop, and (2) thr ough residual effects in which N2 fixed by the legumes is available afte r senescence of legume residue to an associated sequentially cropped non-legume. It is generally belie ved that the second mechanism is more important. According to Ledgard and Giller (1995), N benefits of these systems may accrue more to subsequent crops after root and nodule senescence and decomposition of fallen leaves. Cereal -grain legumes are the most common intercropping systems in mixe d farming systems of sub-Saharan Africa. However, these grain legumes contribute little or no N to associated cereal crops because a large proportion (60-70%) of the N is removed dur ing grain harvest (G iller et al., 1998). The use of forage legumes in many parts of the tropics is limited because they do not contribute directly to the human food supply. Forage le gume-cereal intercropping often

PAGE 22

7 increases the quantity and quality of resi dues, but may decrease the yield of the companion cereal crop (Mohamed-Saleem, 1984). Ibrahim (1994) showed that the dry matter (DM) yields and land equiva lence ratio (LER) of sudangrass ( Sorghum sudanense )-lablab ( Lablab purpureus ) intercrops were higher compared to those of monoculture stands, but the highest protein yiel d was obtained from sole lablab. Although the beneficial effects of intercropping forage legumes have been demonstrated by intercropping cereals with a few legumes such as lablab, the same cannot be said for all legumes. This is beca use intercropped legumes produ ce much lower above-ground yields and their root systems are probably le ss developed than roots of legumes cultivated as pure stands (Nnadi and Haque, 1988). Rotations Involving Forage Legumes The potential of forage legumes to increas e the productivity of crop-livestock systems has received increased attention in recent years because declining soil fertility and scarcity of livestock feeds ar e major constraints limiting agri cultural productivity in these systems. These legumes are grown as rota tion leys, cut and carry, or fodder banks. Studies conducted at the International Live stock Center for Afri ca (ILCA) (Nnadi and Haque, 1988) showed that rotations invol ving cut and carry forage legumes ( Trifolium steudneri , Vicia dasycarpa , Trifolium tembese ) preceding cereal crops [sorghum ( Sorghum bicolor ), maize ( Zea mays L.), and wheat ( Triticum aestivum )] increased the grain yields compared to cereal grown on plot s that were previous ly planted to oat ( Avena sativa ). An alternative system is to rotate cereal crops with short–term fallows of forage legumes (Tarawali and Mohamed-Saleem, 1995; Tarawali and Peters, 1996; Muhr et al., 1999a; 1999c). In reviewing the role of fo rage legume fallows in supplying improved

PAGE 23

8 feed and recycling N in sub-humid West Africa, Tarawali and Mohamed-Saleem (1995) showed that cattle with access to Stylosanthes ( S. hamata and S. guianensis ) fallow in the dry season produced more milk, lost less wei ght, and had shorter calving interval and better calf survival than thos e in non-supplemented herds. They also showed that maize following Stylosanthes had greater grain yields than following natural fallows, but the responses varied depending on species. These positive effects were attributed to improved soil properties such as soil bul k density, soil moisture retention, cation exchange capacity (CEC), organic C, and soil N. Thomas and Lascano (1995) showed that the potential rates of N mineralization were greater in soils under 5or 14-yr-old pastures containing legumes compared with sim ilar grass-only pastures. The yield of the first cereal crop was higher after grass-le gume pasture compared with the grass-only pasture when no N fertilizer was applied. Studying the rotational effects of forage le gumes, Muhr et al. (1999c) found that even though large amounts of N, P, and K (up to 120, 10, and 135 kg ha-1, respectively) were removed as dry season herbage, nutrient accumulation in the remaining green manure biomass increased grain yields of subseque nt maize grown on the legume plots. The nutrient export in legume fallow biomass remo ved in the preceding dry season apparently did not preclude the subsequent yield response of maize, but responses varied depending on the sitesÂ’ fertility status. Green manure legumes grown in rotation with cereal crops ha ve the capacity to provide high quality organic i nputs to meet N demands of subsequent crops (Carsky et al., 1999; Tian et al., 2000), but incorporat ing these non-food legumes in the farming system requires a sacrifice of land and labor that is norma lly devoted to crop production

PAGE 24

9 (Drechsel et al., 1996). Review ing studies on organic matter technologies for integrated nutrient management in smallholder farming systems of southern Africa, Snapp et al. (1998) concluded that green manures grown as relay intercrops have a lower N yield potential, but land and labor use may be more efficient and the system is flexible around farmer needs. The challenge is how to in tegrate these legume green manures into the current production systems. On-farm research in West Africa has shown that integration of these legumes into the farming systems and adoption by farmers could be improved if the legumes have multiple uses (Becker, 1995; Becker and Johnson, 1998; Versteeg et al., 1998a). In intensive ag ricultural production systems, rotations or simultaneous intercrops involving green ma nures may not be practical, thus: relay-cropping green manure legumes, in which partial rather than whole total biomass is harvested as fodder, could be the most feasible option of reducing competition for land between livestock and crops. Intercropping and Adaptation Charac teristics of Mucuna and Lablab Mucuna ( Mucuna pruriens ) and lablab in-row intercr opped at the same time with long-season maize for silage production show ed high competitive ability under the shaded conditions of the intercrop (Maasdorp and Titterton, 1997). These intercrops had a higher percent legume in total DM (30 and 15, respectively) compared to several other forage and grain legumes [silverleaf desmodium ( Desmodium uncinatum ), siratro ( Macroptilium atropurpureum ), archer ( Macrotyloma axillare ), lucerne ( Medicago sativa ), lupins ( Lupinus albus, L. angustifolius, L. luteus ), forage soya ( Glycine max ), semi-trailing cow pea ( Vigna unguiculata ), scarlet runner bean ( Phaseolus coccineus ), and grain soybean ( Glycine max )], which contributed 0.4 to 14.4%. The competitive ability can be attributed to the support offered by maize plants, which improves leaf

PAGE 25

10 display of the twining legumes to offset the shading e ffect of maize. However, the aggressive twining and trailing habit of the mucuna was observed to suppress the growth of intercropped maize in India (Singh and Relwani, 1978). Mucuna planted at the same time with maize reduced the DM yield of silage maize by 50% at 16 WAP (Maasdorp and Titterton, 1997). Versteeg et al. (1998a) obs erved that mucuna pl anted earlier than 5 wk after planting maize smothered the young mai ze plants by its aggressive development, resulting in serious grain yield losses. Planting of either muc una or lablab at 4, 6, or 8 wk after planting maize was shown to have no eff ect on maize grain yield in four sites in Zimbabwe having an annual rainfall range of 650 to 1000 mm (Muza, 1998). Although the DM yields of mu cuna and lablab when in-row intercropped with maize were lower (2.0 and 1.7 t ha-1) than when grown as sole crops (4 and 9 t ha-1), the proportion of leaf and pod under the shaded cond itions of the intercrops was shown to be higher (35.4 and 33.6%) than in sole crops (29.6 and 23.7%) for mucuna and lablab, respectively (Maasdorp and Titterton, 1997). The increased proportion of leaf and pods indicates the suitability of top shoot growth as a protein supplement. Lablab is fairly drought resistant and regrow s well even in the early part of the dry season following an earlier cu t (Mangawu et al., 1994, cited in Maasdorp and Titterton, 1997; Weber, 1996). Mucuna tolerates low so il fertility, acidic soils, and drought conditions (Hairiah et al., 1991; Burle et al., 1992; Weber, 1996), properties which indicates its potential for su rviving and producing biomass du ring the drier part of the year. In Brazil (Burle, 1992), when mucuna and lablab were grown at the end of the rainy season, they survived a dry season of 4 mo (Ca. 10 mm mean monthly rainfall). They produced an average of 2.4 and 1.1 t DM ha-1, respectively, and continued growing

PAGE 26

11 when the rains returned. In reviewing the challenges for research and development of legume-based technologies for the African savannas, Weber (1996) concluded that mucuna and lablab are among the species ad apted to cropping systems in sub-Saharan Africa. Further, they are shade and drought to lerant and can be relay cropped into maizebased cropping systems to provide, from residual moisture, additional benefits to the maize crops. Delayed planting, defoliating shoot growth for livestock feed, and planting these legumes between the maize rows may re duce the competitiveness of these twining legumes, while providing forage of good quality. When used as a cover crop, mucuna has a nematocidal effect (McSorley et al., 1994) as well as the ability to smother weeds (F ujii et al., 1992; Beck er and Johnson, 1998; Versteeg et al., 1998a), particul arly broad leaf weeds (Hepperl y et al., 1992). It is also reported that the use of lablab as a green manure/cover crop has the potential to control weeds (Weber, 1996; Becker and Johnson, 1998) including Striga hermonthica (Weber, 1996), a serious weed in cereal-based cropping systems in eastern Africa. In West Africa, the ability of mucuna to control a local weed, cogongrass ( Imperata cylindrica ), seemed to have a major influence on its a doption (Versteeg et al., 1998a), indicating that farmer adoption of cover crop/green ma nure technology may not only be based on agronomic yield, but other factors/uses ma y also be important (Becker et al., 1995). Replenishing Soil Fertility Low soil fertility, particularly N and P defici encies, is recognized as one of the major biophysical causes for declin ing per capita food production in sub-Saharan Africa (Sanchez et al., 1997). Nutrient balance studies in this regi on (Stoorvogel et al., 1993) have shown that on average 22 kg N, 2.5 kg P, and 15 kg K are lost annually and losses can be as high as 112 kg N, 3 kg P, and 70 kg K in the intensely cultivated highlands of

PAGE 27

12 eastern Africa (Van den Bosc h et al., 1998). These losses are much higher than the estimated inorganic fertilizer use in Af rica of 5 to 10 kg (FAO, 1995; Heisey and Mwangi, 1996), emphasizing the need for soil fer tility replenishment. Sustainable crop production in many soils of sub-Saharan Afri ca requires P inputs because the soils are either derived from parent material with low levels of P or have been depleted of plantavailable P through continuous cropping with in sufficient P inputs (Sanchez et al., 1997). The Oxisol soils that are widespread in this region have a major chemical constraint of high P fixation (Deckers et al., 1994). The low native soil P, high P fixation by soils with high Fe and Al concentration, and nutrient depleting effects of long-term cropping without additions of adequate external inputs have contribut ed to P deficiencies in many tropical soils (Jama et al., 1997). Phosphorus can be replenished eith er immediately with high, one-time P application in soils with high P-sorption capacity, or gradually with modera te seasonal applications at rates sufficient to increase P av ailability in soils with low to moderate P-sorption capacity (Buresh et al., 1997). The combination of P and N replenishment can have synergism. The elimination of P deficiency can enhance N2 fixation by legumes (G iller et al., 1997) and the integrated use of P fertilizers with organic materials to supply N can potentially enhance P availability through the addition of P and reduction of the P-sorption capacity of the soil by organic anions (Palm et al., 1997). Application of organic materials may increase crop-available P eith er directly by the process of decomposition and release of P from the biomass or indirectly by the production of organic acids (products of decomposition) that chelate Fe or Al, re ducing P fixation (Nzi guheba et al., 1998). However, the amount of P that can be added th rough organic materials is restricted by the

PAGE 28

13 limited supply of organic material at the farm level. Palm et al. (1997) showed that whereas lablab and mucuna contain sufficient N in 2 or 3 t of leafy material to match the requirement of a 2-t crop of maize, they cannot meet the P requirements and must be supplemented by inorganic P in areas where P is deficient. Jama et al. (1997), working in western Kenya, indicated th at it was economically attract ive to integrate triple superphosphate (TSP) with organic material s having high N to P ratios, the organic material provides the required N for the crop and TSP meets the additional requirement for P. Judicious application of inor ganic fertilizers is recognized as an indispensable means of overcoming soil fertility decline a nd decreasing food production per capita (Mokwunye and Hammond, 1992). Although inorgani c fertilizers are the most effective amendments to maintain soil fe rtility or alleviate nutrient deficiencies, their high cost, inaccessibility, and generalized recommendati ons resulting in low, erratic, and unprofitable crop responses limit their use, pa rticularly on smallholder farms in subSaharan Africa (Vlek, 1993; Nandwa and Bekunda, 1998). It has also been shown that on the poorly buffered Kaolinitic soils found in many areas in the tropics, including sub-Saharan Africa, continuous use of fertilizer alone cannot sustain crop yield and maintain soil fertility in the long-term because of soil acidification, loss of soil organic matter, and soil comp action (Juo et al., 1995a; 1995b). Kang (1993), working on an Alfisol soil in Nigeria, repor ted a soil pH decline from 6.2 to 5.1 during 10 yr of continuous cropping with maize, sweet potatoes ( Impomea batatas ), and cowpeas, manual tillage, and annual application of 160 to 200 kg N ha-1 as urea. On a similar soil in Nigeria, continuous cropping of maize wi th no stover return ed to the soil and

PAGE 29

14 application of 120 kg N as urea, 26 kg P, and 30 kg K ha-1, during 13 yr under no tillage resulted in a steady decrease in pH from a bout 6 to 4.5 in the 0to 15-cm surface soil (Juo et al., 1995b). Several other examples of acidification and the de cline of soil organic matter and exchangeable nutrients in subSaharan Africa are given in a review by Franzluebbers et al. (1998). Because the issue of low soil fe rtility is widespread in the tropics, many studies have been undertaken on the topic, pa rticularly related to the us e of organic manures as an alternative to high cost inorganic fertilizers. This review is confined to the use of mucuna and lablab with only a few studies on othe r species discussed for comparative purposes. Effects of Mucuna and Lablab Green Manures on Soil and Crop Introduction Recently there has been a resurgence of interest in leguminous green manure/cover crops in many parts of the tropics where the use of commercial inorganic N fertilizers is not economically feasible. It is estimated that N2 fixation ranging from 0 to 250 kg N ha-1 with a median of 110 kg N ha-1 can be achieved from annual legumes with growth periods of 100 to 150 d (Giller and Wilson, 199 1; Sanginga et al., 1996 a; Ibewiro et al., 2000b). The contributions of legume residu es to soil improvement and crop production depend largely on the amount of biomass produced (Sanginga et al., 1996a), chemical composition (Palm and Sanchez, 1991; Tian et al., 1992; and Constantinides and Fowness, 1994), and method of appli cation (Mafongoya and Nair 1997). The decomposition and nutrient release by these re sidues are also affected by both climatic and adaphic factors, including th e biological activity and availability of nutrients in the soil (Myers et al., 1994; Mugendi and Nair, 1997).

PAGE 30

15 Potential use of mucuna and lablab green manures Hairiah and Van Noordwijk ( 1989) reported that in a growth period of 14 wk on an acid soil in Onne, Nigeria, mu cuna contributed 110 kg N ha-1 compared to Desmodium , Centrosema , Pueraria, and Vigna, which contributed 26 to 60 kg N ha-1. In Brazzaville, Congo (Mandimba, 1995), mucuna green manure increased the grain yield of maize up to 56% (to a total of 3.6 t ha-1) compared to a control that did not receive any N fertilizer (2.3 t ha-1). This was comparable to the yields of maize fertilized with 100 kg N ha-1 (3.7 t ha-1) and maize grain yields following green manures of pigeon pea ( Cajanus cajan ) (3.3 t ha-1) and sun hemp ( Crotolaria juncea ) (3.4 t ha-1). Reviewing the potential of improved fallows and green manure in Rw anda, Drechsel et al. (1996) reported contrasting results in which they indicated that green manuring is a risky enterprise due to the highly variable biomass pr oduction that in some cases did not compensate for loss of crop yields and additional labor inputs during green manuring. They concluded that due to acute land shortage, farmers were reluct ant to allocate land to fallows with the exception of fields already out of producti on. Based on survey information, Mousolff and Farber (1995), working in Honduras, estimat ed that use of mucuna as a cover crop combined with a fifth of the recommended inorganic fertilizer increased maize grain yield from 0.7 to 2 t ha-1 and reduced cost per hectare by 22%. When mucuna was fertilized with P and K and grown in a derived savanna in West Africa, it accumulated about 310 kg N ha-1 in 12 wk as a sole crop and 166 kg N ha-1 when intercropped in maize (Sanginga et al ., 1996a). Sanginga et al. (1996a) also indicated that mucuna derived 70 % of its N from atmospheric N2, representing 167 kg N ha-1 per 12 wk in the field. Mucuna intercr opped in maize obtained a greater proportion

PAGE 31

16 of its N (74%) from fixation than did muc una planted alone (66%), suggesting that competition for soil N influences the proporti on of N fixed by mucuna. A preceding sole crop of mucuna resulted in a grain yield of s ubsequent maize equivalent to that obtained from 120 kg ha-1 of inorganic N fertilizer (Sanginga et al., 1996a). In an on-farm study in a derived savanna of West Africa, Versteeg et al. (1998a) indicated that when mucuna was used as an annual fallow cover crop, it produced 6 to 12 t ha-1 of DM, and improved subsequent maize grain yields by 70% comp ared to yields from continuously cropped maize. Relay-cropped mucuna in the derived savanna of West Africa increased the succeeding maize growth parameters (height, leaf area, dry matter production, ear-leaf N concentration) and grain yield compared to maize succeeding fallow regrowth (Ile et al., 1996), but these responses were lower than when mucuna fallow was combined with 40 kg N as urea, suggesting the need for additi onal N fertilization in combination with green-manure N. In field experiments on a moist savannah of West Africa, Akobundu et al. (2000) found that after one growin g season, mucuna reduced speargrass [ Imperata cylindrica (L.) Raeuschel] shoot density and dry matter, reduced labor required for subsequent maize, and signifi cantly increased grain yield of the subsequent maize crop by 65 to 129% compared to maize following natural fallow, depending on mucuna species. In a greenhouse experiment, incorporati on of lablab green manure significantly increased soil NH4 + concentration and sorghum DM production, but no residual effect was observed in a second crop sown 40 d afte r harvesting the first crop (Stamford et al., 1994). Incorporation of lablab green manure reduced the so il C:N ratio, and increased

PAGE 32

17 DM yield of subsequent forage maize (18.9 21.1 t ha-1) compared to when native vegetation was incorporated (10.5 t ha-1) (Crespo, 1996). When lablab was incorporated into a Vertisol in Australia, it improved both soil physical (aggregate stability) and chemical properties, reduced soil compac tion, and increased exchangeable cations, whereas surface mulching only improved friabili ty of surface soil (Hulugalle et al., 1996). Working in a derived savanna of West Afri ca, Carsky et al. (1999) evaluated improved fallows using mucuna and lablab in which th e residue was left on the surface at one site and most residue was burned early in the season at the second site. The mean N fertilizer replacement value from legume rotation was 14 kg N ha-1 in the site where all the residue was left on the ground compared to 6 kg N ha-1 where most of the residue was burned early in the dry season with no N applied to th e maize test crop. Maiz e grain yields were 365 (residue on surface) and 235 (residue burned) kg ha-1 higher than natural fallow, respectively. Mucuna and lablab grown as rotations with maize-b ean intercrop for two seasons in sub-humid highlands of East Africa (Wortmann et al., 2000) produced a legume biomass yield of 6.35 and 4.67 Mg ha-1 (means of 2 yr) and average P yield of 10.2 and 8.5 kg ha-1, respectively. These resulted in 2-yr averages for maize biomass yield of 7.66 and 9.78 Mg ha-1 and maize grain yield of 2.63 and 3.22 Mg ha-1, following mucuna and lablab respectively. The two-y ear average bean plant yields were 1.31 and 1.43 Mg ha-1, and bean grain yields were 0.51 and 0.49 Mg ha-1, respectively. In evaluating the P requirement and nodulati on of legumes in a derived savanna of West Africa, Sanginga et al. (1996b) found th at legumes grown in soil from degraded fields responded more to P application than those grown in soil from manured fields.

PAGE 33

18 This P response was greater for mucuna than lablab: however, the proportion of N derived from atmospheric N was greater for lablab than for mucuna. In a greenhouse experiment to assess the effect of seed size and P fertilization on gr owth of 12 herbaceous and shrub legume species, Kolawole and Kang (1997) found that P fertilization increased above-ground biomass, root DM biomass, nodula tion, and concentrations of N, P, K, Ca, and Mg of mucuna and lablab. Regardless of fertilization level, elem ent (N, P, K, Ca and Mg) concentrations were higher in mucuna than in lablab. In areas where land is scarce such as in smallholder farms in western Kenya, the intercropping/relay crop ping of green manures in cereals may be the only feasible means of generating organic inputs. In such cases , management of green manures to minimize competition with maize while maximizing the residual benefit in yield is critical. Further investigations are needed to evaluate the pe rformance of these legum es in relay-intercrop systems under the sub-humid tropical conditions of the highlands of eastern Africa. Research is also needed on the residual eff ect and whether incorpor ation of green manure for two consecutive years is better than 1 yr of application. This information will be useful for smallholder farmers who may want to alternate the growing of legume green manures with preferred grain legumes. Effect of Root and Shoot St ubble on Soil and Crop Responses The contribution of roots to the nutrient and organic matter of the soil is believed to be important for soil fertility maintenance a nd carbon sequestration, as the below-ground biomass forms a substantial proportion of the total biomass in an ecosystem. It is estimated that roots may be the source of 30 to 60% of the C in the soil organic pool (Heal et al., 1997). Root ti ssues are continuously sloughed off and replaced, and these sloughed-off tissues, along with senescent and de ad roots, constitute a significant avenue

PAGE 34

19 of organic matter (and nutrients) addition to the soil ecosystem. It is estimated that nodulated legume roots contain < 15 kg N ha-1 (Kumar Rao and Dart, 1987; Bergersen et al., 1989; Ibewiro et al., 1998) to between 30 and 50 kg N ha-1 (Chapman and Myers, 1987; Unkovich et al., 1994; Tian and King, 1998) . This amount of root N represents < 15% of total plant N (Peoples et al., 1995). On an acid soil, 6-wk-old mucuna had a shallow root system (within 15 cm) and a shoot:root ratio of 6.7:1 (Hairiah et al ., 1992). Roots contributed only 2 kg N ha-1 compared to above-ground biomass that contributed 21 kg N ha-1. In the same trial, perennial species Pueraria phaseoloides and Centrosema pubescens had shoot:root ratios of 2.4:1 and 19.5:1 and roots contributed 150 and 30 kg N ha-1, respectively. In a pot experiment to evaluate the biom ass and chemical composition of selected leguminous cover crops (Tian and Kang, 1998), r oots contributed about one third of total plant biomass. Fertilizer application favored the accumulation of biomass in shoots, and the effect was stronger in low than in high fe rtility soil. On averag e, the root and nodule contribution of mucuna and lablab were 31 a nd 37% of total plant biomass, respectively. Kiflewahid and Mosimanyana (1992) found that the contribution of lablab leaves, stems, and roots to the whole dry biomass was 47, 46.6, and 6.4%, respectively. Carsky et al. (1999) showed that mucuna and lablab grow n as improved fallows had root to shoot proportion of 2 to 3 and 8%, re spectively, but they cautioned th at their estimates of root DM may have been low because sampling was done only to a depth of 30 cm and not all the fine roots were recovered. Their data do indicate, however, that lablab had more root biomass than mucuna.

PAGE 35

20 Although Muza (1998) reported th at mucuna roots containe d lower N concentration (1.38%) compared to above-ground biomass (1.96%), Tian and Kang (1998) found that roots of mucuna and lablab contained hi gher N concentration (2.62 and 2.44%) compared to shoots (1.66 and 2.09%, respectively). Th e lignin concentrations in the roots of mucuna and lablab were higher (24.5 and 19.6% , respectively) than in the shoot (7.2 and 6.0%, respectively) (Tian and Kang, 1998), but the roots contained lower polyphenol concentrations (0.63 and 0.26%, respectively) than shoots (3.54 and 1.4%, respectively). The higher lignin concentration in roots suggests that, in comb ination with shoot stubble, the remaining biomass after removal of the top canopy may be of low quality. Quantities of soil N greater than 100 kg ha-1 have been reported under lucerne despite removal of large amounts of shoot by grazing animals or as hay (Gault et al., 1995). Russell and Fillery (1994), using 15N techniques with lupins, found that below-ground N may be almost three times greater than calcu lated from N contained in recoverable root material because the contribution of fine roots and slough off are not accounted for. Peoples et al . (1995) concluded that if this finding is applicable to a range of other legumes, then determinations of the amount of N returned to the soil, based on recoverable roots, have b een greatly underestimated. Smyth et al. (1991) observed lower yields and N accumulation by maize when mucuna roots were incorporated into the soil co mpared to whole biomass incorporation and attributed this reduction to th e removal of the N in above-ground biomass. In comparing N release from four legume speci es [alfalfa, arrowleaf clover ( Trifolium vesiculosum ), cowpea, and soybean], Frankenberger and Abde lmagid (1985) observed that the different plant components released N in the order fo liage > roots > stems and that roots with

PAGE 36

21 higher lignin concentration had slower N release than those with lower lignin concentration. Reid et al. (1987, in Smyth et al., 1991) found an increase in maize grain yield of 0.8 t ha-1 when mucuna roots were incorporated relative to a zero-N treatment. Oikeh et al. (1998) working in either low or high fertility sites in a moist tropical savanna, showed that independent of the differences in fertility, N uptake and N partitioned into grain, stover, and cob were 20% higher after legume-maize rotation than after maize intercropping even though the legume tops were removed from fields. Ibewiro et al. (1998) studied the N contri bution of mucuna, lablab, cogongrass and maize roots, shoots, and whole-plant bioma ss to succeeding maize. They showed that although N contribution from mucu na and lablab roots was only 3 and 4% of the total N, their incorporation resulted in maize grain yield that was 38 and 89%, respectively, of the yield obtained when whole residue was incorp orated. This was attributed to the low quality of roots that may have improved th e synchronization of N release with N uptake of the succeeding maize crop. These results suggest that incorpor ation of low quality root and stem stubble after defoliating t opgrowth for livestock feed may enhance N contribution to succeeding maize in low-exte rnal input, continuous cropping systems, as well as provide quality fodder for livestock. The dense cover formed by creeping legumes such as mucuna leads to self shading and senescence resulting in a mat of fallen leaves (Van Noordwijk and Purnomosidi, 1992). Van Noordwijk and Purnomosidi (1992 ) reported a total N input from mucuna litter in a 12-wk grow th period of 42 kg ha-1 from 1.5 t ha-1 DM. Litter quantity exceeded the live biomass measured at the end of the growth period.

PAGE 37

22 Significant increases in maize yields fo llowing mucuna even when mucuna was burned to ease land preparation (Vine, 1953) supports the hypothesi s that below-ground parts may contribute significant N to subseque nt maize. Despite the availability of data on the contribution of whole-plant biomass in corporation, more information is needed on the contribution of below-ground biomass pl us stem stubble and litter to succeeding maize. Manure Effects on Soil Fertility Improvement Livestock play a key role in farming sy stems where farmers rely on organic matter recycling for maintaining soil fertility. Cycling of bioma ss through livestock and the use of manure and urine to fertilize the soil is an important linkage between livestock and soil fertility maintenance (Powell and Valentine, 1998). Thus cattle manure is an integral component of soil fertility management in ma ny areas of the tropics and its importance as a source of nutrients for crop production is widely recognized (Bationo and Mokwunye, 1991; Powell and Williams, 1993). The quantit y and quality of manures available on smallholder farms are the major factors limiting its contribution. Crop responses to manure appli cation may be due to N, P, cations such as Ca and Mg or to physical effects of additional soil orga nic matter on water inf iltration and retention (Mugwira and Murwira, 1997). Powell and Willi ams (1995) showed that the use of 20 t ha-1 manure (as-is basis) increased soil pH, soil organic matter, total-N, and available P after one season of application. However th e responses to manure application are highly variable due to differences in the chemical composition of the manures and the rates and frequency of manure application. The chemi cal composition of manures differs because of variation in animal diet [i.e. lignin:neutral detergent fi ber (NDF), lignin:nitrogen (N), and polyphenol:N] (Somda et al., 1995) and ma nure storage. Poor storage conditions

PAGE 38

23 may result in ammonia losses through volat ilization (Murwira, 1995) and leaching of nitrate (Cahn et al., 1993). Proper management of crop residues, forage , and manure can have positive impact on the nutrient dynamics of low-input farming sy stems (Swift et al., 1989). Up to 95% of total nutrients consumed by livestock are excret ed. Nitrogen is excreted in both feces and urine, while P is excreted almost exclus ively in feces (Ternouth, 1995). The proportion and forms of fecal N are highly influenced by animal diet (Powell et al., 1994; Somda et al., 1995) and to a lesser extent on how manur e is collected, stored, or decomposed. Lower concentration of organic matter, N, and P in the feed, results in lower quantity and quality of manure (Powell and Saleem, 1987). High con centration of polyphenolic compounds in feed increases the amount of f ecal N and decreases the amount of urinary N (Powell et al., 1994; and Somda et al., 1995). Feeds with high lignin, tannins, and related phenolic compounds may improve N cy cling, by shifting N excretion from urine to feces and by lowering mineralization rates, th us increasing the availa bility potential of N to crops (Powell et al., 1994). It has also been shown that there is a high correlation between the concentrat ion of feed lignin (r2 = 0.97) and feed polyphenol (r2 = 0.85) with the concentration of neutral-detergent inso luble N in feces (Somda et al., 1995). The manure that is applied to the soil decom poses and releases nutrients faster than herbaceous biomass (Somda et al., 1995). Murwira and Kirchmann (1993) conducted a laboratory incubation study which showed that aerobically decomposed dry manure may be an unsatisfactory source of N because amount s of N mineralized after 8 wk were very low when communal-area and feedlot manur es were soil incubated (1.3 and 2.1%, respectively). They suggested that even though the C and N mineralization values were

PAGE 39

24 low, manures could have other beneficial effects such as in creasing the base status and improvement of water hold ing capacity of the soil. Despite the importance of animal manur e in sustaining crop productivity, the availability of sufficient amounts to perm it adequate food production and improvement of soil quality, and availability of feed to support the livestock that produce the manure, are important management issues that need to be addressed. In smallholder farming systems where pressure on land is high, the amounts of manure available may not be sufficient to permit adequate grain yield. Williams et al . (1995) estimated that 9 to 21 cattle are required per hectare of pearl millet ( Pennisetum glaucum ) to replenish the 22 to 38 kg N and 2 to 6 kg P ha-1 harvested in grain and crop residue s. The amount of natural pasture required to support livestock for the purpos e of manuring crop land, calculated as rangeland:crop land ratio, is estimated to ra nge between 10 and 40:1 (Fernandez et al., 1995; and McIntire and Powell, 1995). Manure ap plication rates need to be evaluated to estimate the amount of manure required to offset nutrient harvests from grain and cereal crop residues. Evaluation of the trade offs in applying plant materi al directly to soil versus feeding it to animals and applying their manure to soil is also required to assist farmers in managing green legume manures and other feeds and livestock to improve soil management. Integrated Nutrient Management Land intensification on diminishing farm sizes as a consequence of population pressure has resulted in the more sust ainable bush-fallow systems being replaced by nutrient-depleting, continuous cropping system s. Continuous cultivation, even when limited amounts of inorganic fertilizers are us ed, often leads to rapid decline in soil organic matter, pH, and nutrients (Juo et al., 1995a; 1995b), leading to negative nutrient

PAGE 40

25 balances (Stoorvogel et al., 1993; Smaling et al., 1996; 1997; Van den Bosch et al., 1998). Previous studies have shown that inorganic fertiliz ers alone cannot sustain crop yields on poorly buffered Kaolinitic soil in the tropics due to factors such as soil acidification, compaction, and loss of soil orga nic matter (Juo et al., 1995a). Thus the application of organic materials is needed not only to replenish soil nutrients but also to improve soil physical, chemical, and biological properties. Integrated Nutrient Management (INM), which seeks to maximize the complementary effects of mineral and organic nutrient source s is emerging as an important approach in improving soil productivity of smallholder fa rming systems (Jenssen, 1993; Smaling et al., 1996; Palm et al., 1997; Fanzlaebbers et al., 1998). The INM concept is based on the premise that the decline in soil productivity can be attributed in part to the negative nutrient budgets (the amount of nutrients re moved compared to the amount of nutrients being put into the system) in most agricultur al production systems in sub-Saharan Africa (Smaling and Braun, 1996). Thus, under contin uous cropping, recyc ling of nutrients from organic sources alone may not be su fficient to sustain cr op yield. Nutrients exported from the soil through harvested biom ass and lost from the soil through various processes such as soil erosion (Lal, 1989; 1995; Swift et al., 1994a ), leaching (Cahn et al., 1993), and denitrification (L ensi et al., 1992) must be re placed with nutrients from external sources. Organic residues such as crop stover and animal manures are often used as alternatives to chemical fertilizers. Long-term experi ments on a Kaolinitic Al fisol in West Africa have shown that with continuous maize croppi ng under a no-tillage sy stem, retention of residue as mulch resulted in a slower declin e in soil organic matter, and higher CEC, pH,

PAGE 41

26 exchangeable Mg, and crop yield compared to when residue had been removed (Juo et al., 1995b, 1996). These results contrast with those reported by Palm et al. (1997), who found that incorporati on of maize stover decreased so il available N with consequent reduction in maize grain yields , at least temporarily. Thes e temporary negative effects can be offset by combining low quality crop re sidues with inorganic fertilizers (Paustian et al., 1992) or high quality organic material s with N concentration more than 20 g kg-1 (Smith et al., 1993). In eastern Africa, a long-term experiment on a fine-textured Alfisol at Kabete, Kenya, showed that a combination of farmyard manure and inorganic fertilizer resulted in higher maize grain yield than either farmyard manure or inorganic fertilizer alone (Swift et al., 1994b). However, farmyard manure cannot meet nutrient demand over large areas because of limited availability, low nutrient concentration, and the high labor demand for processing and application. Other long-te rm experiments conducted in sub-Saharan Africa have shown that a combination of inorganic fertilizers and organic manures slowed the decline in soil organic matter after conti nuous cropping compared to when inorganic fertilizers were us ed alone or when no inputs we re used (Swift et al., 1994b; Kapkiyai et al., 1999). The co mbination of a high-quality organic material, tithonia ( Tithonia diversifolia ) leaves, and inorganic P in a pot experiment and subsequent field studies (Gachengo et al., 1999) re sulted in greater maize biom ass and P uptake than from equal amounts of nutrients added from inorganic fertilizers. Studies in Zimbabwe on N mineralization from poor quality manures have shown that decomposition of these manures can lead to N immobilization and th at N availability can be increased by supplementing with inorganic sources of N (Murwira and Kirchmann, 1993).

PAGE 42

27 Palm (1995) found that when organic mate rials (animal manures, leguminous tree biomass, and green manures) were applied in modest amounts, i.e., 2 to 5 t DM ha-1, and contained sufficient N to match nutrient demand for a 2-t crop of maize, the P requirement was not met and inorganic P was n eeded in areas where P is deficient. Onfarm research conducted in northwester n Kenya (KARI, 2000), showed that a combination of half the recommended ra te of inorganic fertilizer (30 kg P205 + 30 kg N ha-1) with half the recommended rate of farm yard manure or compost (5 t ha-1), produced maize grain yield comparable to that obtained when either in organic fertilizer or organic manures were used alone at recommended ra tes. Akobundu et al. (2000) showed that applying a low fertilizer rate (30 kg ha-1 N, P, and K) with mucuna residue, significantly increased maize grain yield in a moist savanna of West Africa. In reviewing results on the combined use of organic and inorganic nu trient sources in s ub-Saharan Africa, Palm et al. (1997) concluded that hi gh and sustained crop yields ca n be obtained with judicious use of organic residues combin ed with inorganic fertilizers. They attributed this advantage to enhanced sync hrony of nutrient release and demand by the recipient crop, increased nutrient-use efficiency, and residua l effects of soil organic matter associated with combined nutrient additions compared to inorganic fertilizers applied alone. Residue Quality Residue quality can be defined as the intr insic characteristics or the chemical constituents which affect residue decompositi on in the soil (Swift, 1989). The quality of organic residues has a significant effect on thei r ability to improve so il fertility both in the short term through their e ffects on nutrient availability and in the long-term by their effect on soil organic matter. The short-term effect on nutrient avai lability occurs via nutrient release during decomposition or indirectly through decomposition products

PAGE 43

28 (organic ions). Low quality residue mate rials have characteristics which inhibit decomposition including high lignin or polypheno l concentration or low N concentration. Low quality litter/residue limit th e transfer of organically bound N to the pools of mineral N available for plant growth (Constantinid es and Fowness, 1994; Handayanto et al., 1994). The N concentration and C/N ratio ar e major determinants of the ability of organic residues to supply nutrients (Kachak a et al., 1993; Myers et al., 1994), however, there are other modifying factors. Lignin, which is a recalcitrant substance (Spain and Le Fe uvre, 1987), results in slow mineralization of lignin-bound carbon and N (Van lauwe et al., 1996). This is because lignin intertwines with the cel l wall and physically protects cellulose and other cell wall constituents from degradation (Cheson, 1997). It is the lowest quality carbon constituent, providing little or no energy to decomposers until the late stages of decomposition. Polyphenol is a general term for compounds that have a hydroxyl group bonded to an aromatic ring. They include a range of compounds differing in size, complexity, and reactivity such as tannins, coumarins, and flavonoids. The condensed and hydrolyzable tannins are the most important in terms of decomposition and nutrient release dynamics. Although soluble polyphenols can serve as carbon substrate, many of them act as bactericides (Scalbert, 1991; Jones et al., 1994) and can therefore slow down decomposition by lowering the activity of mi croorganisms and enzymes. Phenolic compounds also bind mineralized N in nitroand nitrosoforms in soil humus, making it unavailable (Azhar et al., 1986) thus reducing the availabil ity of N for soil microbial population (Reed, 1995). Although the total so luble phenolics include a variety of compounds, consistent and signifi cant negative correlations have been obtained between

PAGE 44

29 total polyphenols obtained by the Folin-Denis method and N mineralization (Palm, 1995). The correlations may be improved if the extracts are analyzed for condensed tannins, the major group of polyphenols that bind proteins, or by doing assays that assess the protein-binding capacity (Waterman and Mole, 1994). The protein-binding capacity of polyphenols has been shown to greatly sl ow down decomposition (Palm and Sanchez, 1991; Handayanto et al., 1994). In estimating how release of mineral N was affected by chemical composition, Constantinides and Fowness (1994) , incubated soil with fresh l eaves and with litter from 12 species commonly used in agroforestry systems. Their results showed that decomposition of fresh legume leaves resulted in net N accumulation whereas litter had net N depletion. The difference between the N concentrations in gr een leaves and leaf litter can be explained by the process of nutrient resorption during senescence, when leaf proteins and other nitrogenous compounds ar e hydrolyzed and the products transported into perennial tissue before leaves fa ll off the plant (Norby and Cotrufo, 1998). Palm and Sanchez (1991) found a negative co rrelation between ne t N mineralization and polyphenol concentration or the polyphenol:N ratio. Kachaka et al. (1993) reported a significant linear relationship between N mineralization and lignin:N ratio (r2 = 0.996) or the (lignin + polyphenol):N ratio (r2 = 0.981) of prunings of four agroforestry species. According to Constantinides and Fowness ( 1994), the N concentra tion and C:N ratio are the most important residue quality characteri stics that influence N release patterns. Abdel-Magid (1992) found that th e N content in lablab plant tissue (foliage or foliage + stem) had no significant correlation with cu mulative amount of N mineralized in a laboratory incubated soil, but there was a significant corre lation between the C:N ratio

PAGE 45

30 and the percentage of N mineralized. Working with mucuna and lablab as cover crops in a derived savanna of West Afri ca, Ibewiro et al. (2000a) showed that the initial (lignin + lignin):N ratio, N concentration, lignin:N ra tio, polyphenol:N ratio, C:N ratio and lignin concentration of these cove r crop residues explained th e largest proportion of the variation in decomposition rate constants. Palm et al . (1997) reported that residue N con centration ranging from 18 to 22 g kg-1 is the critical range for transition from net im mobilization to net mineralization. Based on this criterion the commonly used cereal crop residues and cattle manure of low quality fall below the critical value (Palm et al., 1997) . However, not all residues with high N values exhibit net mineralization. Lignin concentration of > 150 g kg-1 and polyphenol concentration of > 30 to 40 g kg-1 could slow N release and result in net immobilization (Palm, 1995). However there is no agreement in the literature on wh ether C:N, lignin:N, polyphenol:N, or (lignin + polyphenol):N has more influence on N release patterns (Vanlauwe et al., 1997). Studying the impact of residue quality on C and N mineralization of leaf and root residues of three agroforestry species, Va nlauwe et al. (1996) found that leucaena ( Leucaena leucocephala ) and Dactyladenia barteri roots contained more lignin and less N than their leaves, whereas th e lignin and N concentration of Fleminga macrophylla leaves and roots were not different. Tian and King ( 1998) showed that roots of herbaceous legumes (including mucuna and la blab) contained higher lignin concentration (200 g kg-1) compared to shoots (60-100 g kg-1), but polyphenol concentration in shoots was higher than in roots. Although the nega tive relationship of lignin:N, polyphenol:N, (lignin + polyphenol):N ratios to mineralizati on of N have been shown (Palm and Sachez,

PAGE 46

31 1991; Tian et al., 1992; Consta ntinides and Fowness, 1994) , more information on the relationship between root and stem stubble qua lity and N mineralization of green manure legumes is needed. The ultimate aim of st udying residue quality indices is to identify robust indices that provide improved prediction of nutrient release. This will allow for cross-site comparisons and synthesis of results from a broad range of studies to model the effects of residue application on long-term soil fertility changes. Synchrony Between Mineralized N and N Uptake The efficiency of transferring N from a legume green manure to the succeeding crop depends on synchronizing the N release from the legume residue with the demand of the recipient crop. This is partic ularly important in high rainfa ll climates with high potential for N leaching. The plant species and manage ment practices have a great influence on the success of this synchroni zation, and N mineralization is also affected by moisture, temperature, and soil factors such as texture, mineralogy, acidity, biological activity, and the presence of other nutrients (Myers et al., 1994). Th e inorganic nutrients released by mineralization may be immobilized by the soil bi ota, taken up by plants, or lost as a result of leaching, denitrification, or vol atilization, or they may remain in the soil. The size of the inorganic pool of nutrients depends upon the balance of various processes that add to the pool (mineralization) a nd those that subtract (imm obilization, plant uptake, and losses). With respect to synchrony of N, the most important processes are mineralization-immobilization, plant uptake, denitrification, volatili zation, and leaching. Improved synchrony implies that there is less ex cess mineral nutrient in the soil, and thus the opportunity for loss is reduced. Uptake of N and other nutrients by maize cont inues until near maturity, but the highest demand for N is at the start of reproductive stage (R1) when grai n filling is initiated to R6

PAGE 47

32 at physiological maturity (Karle n et al., 1988). The fraction of total N added that is taken up by the crop is known as the N recovery va lue (NIV). The reported NIV values for most organic residues are in the range 10 to 30% by the fi rst crop (Giller and Cadisch, 1995; Palm, 1995; Mafongoya and Nair, 1997) and between 2 and 10% by the second crop (Mafongoya and Nair, 1997). Giller and Cadisch (1995) reported that approximately 20% of the N from high quality green manure residue is recovered by the first crop. Factors influencing the sync hrony and therefore NIV from organic manures by annual crops include type of species, biomass quali ty, and method and time of application (Tian et al., 1992; Mafongoya et al., 1997b). Incorpor ation of the residue improves N recovery compared to surface placement (Mafongoya and Nair, 1997), and this has been attributed to elimination of N losses through ammoni a volatilization (Glasner and Palm, 1995). An important question that needs to be answered is: can synchrony be enhanced through managing residue quality? In an attemp t to answer this question, Myers et al. (1994) concluded that the answ er is a cautious yes. This is based on the few studies reported in the literature wh ere results support the concept of using residue quality to achieve synchrony. Most of these studies have reported on the synchrony of aboveground leafy materials (Ha ndayanto et al., 1994; Myer s et al., 1994; Palm, 1995, Handayanto et al., 1997) and little information is available on the synchrony of N release and demand when root residue or root and lo w quality stem stubble is incorporated for the subsequent maize crop. Previous studies have indicat ed that residues of intermediate quality may result in nutrient release patterns in synchrony w ith crop nutrient demand (Handayanto et al.,

PAGE 48

33 1994; Mafongoya et al., 1997a). Myers et al . (1994) found that a mi xture of straw and Gliricidia sepium residue mineralized N at an intermediate rate, and that the rate gradually increased coinciding with the period when growth demand by maize was highest. In a laboratory in cubation experiment, Frankenbe rger and Abdel-Majid (1985) compared N release from residues of four legume species (alfalfa, Egyptian clover, cowpea, and soybean). Different plant com ponents released N in the order foliage > roots > stems. When these parts were combined, the N mineralization pattern was intermediate between that of foliage and stem s. Also an interesting result was obtained from soybean and Egyptian clover. Soybean initially released N more rapidly than Egyptian clover in the first 3 wk. After W eek 3, the rate of release was faster from clover, so that by 14 wk the amounts were equa l. Soybean, that ha d greater root lignin concentration, released nutrients more sl owly over the entire 20-wk study period. However contrasting results were obtained from a 70-d incubation period by Nnadi and Haque (1988) in which there was no evidence of N mineralization from several different legume roots. The mixing of residues of di fferent quality may re sult in significant interactions involving physical protection of recalcitrant subs trates at later stages of decomposition (Mafongoya et al., 1997b) or by the protein-binding capacity of polyphenols (Handayanto et al., 1997). Add itionally, the mechanisms that alter the physical environment, such as increasing moisture availability under mulching or increasing soil organic matter (Tian et al., 1993), may play a role in these interactions. Constantinides and Fowness (1994), Myers et al. (1994), and Palm et al. (1997) hypothesized that plant residues which initia lly immobilize and subsequently release N slowly could enhance N uptake by plants. Hairiah and Van Noordwijk (1989 reported in

PAGE 49

34 Van Noordwijk and Purnomosidi, 1992) found that N uptake by maize following mucuna was 147 kg ha-1 higher than the control crop, while the N content of live biomass incorporated was only 71 kg ha-1. They attributed this to the large amount of litter fall during the growth period, a quantity that exceed ed the live biomass measured at the end of the growth period (Van Noordwijk and Pu rnomosidi, 1992). The research reported in this dissertation hypothesizes th at utilization of the high qua lity top growth of green manures as livestock feed and incorporating the remaining intermediate quality stubble and root will increase NIV and enhance th e synchrony between N release and uptake and thus provide better yield than when no fertilizer is applied. Mixed Crop-Livestock Farming The integration of crops and livestock, partic ularly dairy cattle, is one of the best avenues in which both soil-fertility impr ovement and increased productivity from livestock can be achieved on smallholder farms. These mixed farming systems involving complementary interactions between crops and livestock, such as using traction and manure for cropping, and feeding crop residues and other cropland fo rages to livestock, are increasing in importance in many areas in the tropics (Powell and Williams, 1995). The animals are also a source of income, wh ich can be invested in crop production apart from meeting social and cultural obligations. It is estimated that 74% of the cattle in subSaharan Africa are associated with small holder mixed crop-livestock farming (Brumby, 1989). Under these systems, three dairy produ ction alternatives can be distinguished, namely zero grazing, free grazing, and a comb ination of the two (semi-zero grazing) (Nyambati, 1997). The first two systems are the most common due to small farm sizes. Decreasing farm sizes has led to an increased intensification of land use, with consequent

PAGE 50

35 need for efficient cycling of nutrients among crops, animals, and soil if the productivity of these low-input systems is to be sustained (Mohamed-Saleem, 1998). Although there could be competition for crop residues between livestock feeding and soil fertility improvement, returning crop residue s to the soil, particularly stover may not be the best option. Long-term studies in the sub-humid highlands of eastern Africa have shown that maize stover retention results in reduction in yields compared to a control where no external input was a pplied, whereas application of farm yard manure improved crop productivity (Kapkiyai et al ., 1998), suggesting that it coul d be advantageous to feed crop residues to livestock and a pply the resultant manure to th e soil. It has also been shown that nutrient release from manure is faster and appears to be more synchronous with crop nutrient demand than the nutrient release patterns of low quality plant biomass applied to soils (Somda et al., 1995). The main feeds in these systems include the purposely-grown fodders such as napiergrass ( Pennisetum purpureum K. Schum), crop residues, natural weed fallow, and roadside grazing. The full integration of crops and livestock is not without risk, however. Poor livestock performance can occur due to confinement to small grazing areas during the cropping season to avoid crop damage. This in turn increases th e risk of overgrazing and environmental degradation through the proc ess of soil erosion. Excessive grazing of stover during the dry season has also contributed to removal of nutrients as feed from cropland without adequate replacement, leading to negative nutrient balances (Smaling et al., 1997). Napiergrass is the major fodder preferred in these systems because of its high yield potential and drought tolerance, making it suitable as a cut-an d-carry fodder. Napiergrass

PAGE 51

36 alone cannot support high levels of livestoc k productivity, particul arly dairy cattle (Anindo and Potter, 1986). Often during the dr y season napiergrass is harvested at an advanced stage of maturity and it is low in nutritive value (Mui a at al., 2001). The conventional methods of improving napi ergrass quality through fertilization (Sollenberger and Jones, 1989) or supple mentation using concentrates (Anindo and Potter, 1986) are limited because most smallhol der farmers cannot afford these inputs. The growing of forage legumes as rotations or improved fallows to specifically increase livestock productivity is in creasing (Tarawali and Mohamed-Saleem, 1995; Weber, 1996). Potential Use of Napiergrass Fodder Background Information (Botany, Origin, and Characteristics) Napiergrass, also known as elephantgrass, is the most popular perennial fodder in the intensively managed smallholder crop-livesto ck mixed farming production systems. The grass, which is native to eastern and cen tral Africa (Boonman, 1993), was named napier after Colonel Napier who advocated its use for commercial livestock production (Boonman, 1997). The name elephantgrass is probably associated with the African elephant ( Loxodanta africana ), which has a special preference for napiergrass. There are both giant (tall) and dwarf types. The tall na piergrass is robust, grow ing to 4 m in height and having up to 20 nodes (Henderson and Preston, 1977). This type resembles sugarcane in habit and ad aptation and forms bamboo-lik e stems when mature. Napiergrass is propagated vegetatively becau se seeds have low genetic stability and viability (Humpreys, 1994). The dwarf ‘Mott’ napiergrass bred at the Coastal Plains Research Station in Tifton, Georgia, has maximum height of 1.5 m (Hanna and Monson, 1988). Unlike the tall napiergrass, the dwarf type is very leafy and non-flowering. In

PAGE 52

37 East Africa and particularly in Kenya, there are a number of tall cultivars which have been selected and tested over a wide range of environments (Goldson, 1977). The most common cultivars grown in Kenya are Bana, French Cameroon, and Clone 13. Napiergrass grows well from sea level at the hot humid coast to altitudes of over 2100 m, but frost appears to limit its cu ltivation above this altitude (Skerman and Riveros, 1990). Dry Matter Yield Napiergrass produces greater dry matter yi elds than the other tropical grasses (Skerman and Riveros, 1990). In Kenya, dr y matter yields vary between 10 and 40 t DM ha-1 depending on soil fertility, climate, and ma nagement (Schreuder et al., 1993). These DM yields contrast with those of rhodesgrass ( Chloris gayana ) and kikuyugrass ( Pennisetum clandestinum ), popular ley pasture grasses in Kenya, which yield between 5 and 15 t DM ha-1 (Boonman, 1993). Exceptionally high DM yield of up to 85 tons DM ha-1 have been reported elsewhere when high rates of N fertilizer were applied to napiergrass (Skerman and Riveros, 1990). Despite its impressive performance in the southeastern USA as a grazing fodder (Solle nberger and Jones, 1989), dwarf napiergrass has not been adopted by small holder dairy fa rmers due to its relatively low DM yields (Sotomayor et al., 1997) and its high suscepti bility to the fungal snow mold disease, Cowdria sphaenoides , under Kenyan conditions (Boonma n, 1997). Smallholder farmers therefore prefer the giant napiergrass, which is well suited for the intensive “cut and carry” feeding system. Nutritive Value Napiergrass has been a subject of extens ive continuing research, not only in East Africa but elsewhere. Most of the initial stud ies tended to concentrate on its agronomy, chemical composition, and in vitro digest ibility (Ware-Austin, 1963; Odhiambo, 1974;

PAGE 53

38 Mugerwa and Ogwang, 1976; Goldson, 1977; Sn ijders et al., 1992; Anindo and Potter, 1994). Animal performance on napiergrass has only been evaluated in recent years, and emphasis on these studies has been on protei n supplementation to dairy cows feeding on napiergrass as a basal feed (Anindo and Po tter, 1986; 1994; Muinga et al., 1992; 1993; 1995; Mukisira et al., 1995; Abdulrazak et al ., 1997; Kariuki et al ., 1998; Muia et al., 1999; 2000a; 2000b; 2001). The crude protein concentra tion of napiergrass is lowe r during the dry season, but it remains relatively constant throughout the year compared to ley pastures (Ware-Austin, 1963). The nutritive value of napiergrass is affected by the stage at which it is cut (Odhiambo, 1974; Skerman and Riveros, 1990; Anindo and Potter, 1994; Muia et al., 1999; 2001). Anindo and Potter (1994) showed that crude protein co ncentration varied from 70 to 148 g kg-1 and digestibility of 8-wk regrowth was 560 to 720 g kg-1 for dryseason and wetseason forages, respectively. They also found an intake of 2.5 and 2.65 kg DM per 100 kg-1 of animal live weight for the dry and wet season forages. Muia et al. (2001) found that the CP concen tration decreased from 82 g kg-1 in 10-wk-old 1.3-m-tall napiergrass to 53 g kg-1 in 15-wk-old 2 m tall napiergr ass herbage, while the neutral detergent fiber (NDF) and lignin concentrations incr eased from 542 and 37 g kg-1 to 634 and 52 g kg-1, respectively. The intake of CP (15 vs. 11 g kg-1 W0.75) was higher for Friesian steers fed the less mature forage. Likewise concentrations of rumen ammonia were higher for steers fed the 10-wk-old material. Legume Supplementation Use of legumes to supplement cattle diets has continued to receive attention from livestock scientists, particular ly in the tropics where low qua lity roughages form the bulk of the feed. Supplementation of low quality basal roughage diets with legume forage

PAGE 54

39 increases essential nutrients available to ru men microbes, increases rate of passage of both particulate and liquid matter phases, a nd may increase degradation, intake, and consequently animal performance (Norton and Poppi, 1995; Osuji and Odenyo, 1997; Nsahlai et al., 1998). Legume forage contains protein, minerals, and vitamins essential for growth of rumen microbes that degrade roughages prior to ga stric and intestinal digestion by animals. The effect of legume supplementation on digestibility of basal feed is based on the fact that digestion of cell wall constituents of the basal diet are enhanced by providing a readily fermentable source of N so that rumen ammonia levels are maintained at an optimum level for cellulose digestion. There is a wide variation in response to legume from no change in intake to 100% increase in intake (Poppi and Norton, 1995). Though the data indicate that a 20 to 50% inclusion of legume in the di et results in a 10 to 45% in crease in intake, the response varies with quality of the basal diet. The positive effects of legume/protein supplementation to low quality basal diets, including napiergrass are well documented in the literature (Muinga et al., 1992; 1993; 1995; Abdulrazak et al., 1997; Muia et al., 2000a; 2001). Studies done in sub-Saharan Africa at the In ternational Livestock Research Institute (ILRI) have shown that supplementation with gr aded levels of lablab increased microbial N supply, rumen degradation, particulate passa ge, and intake and liveweight gain of crossbred calves (Abule et al., 1995). Nsahlai et al. (1998) using sheep found that supplementation with forage legumes (lablab and Sesbania sesban ) to low quality teff ( Eragrotis tef ) diets increased rate of degradation of teff by 50 to 142%, increased the fractional rate of passage of pa rticles, intake, and increased f ecal N. Increasing levels of

PAGE 55

40 lablab in basal diet of maize stover (labla b w/w of 70:30 vs 50:50) increased the amount of readily degradable DM, organic matter, in take, and reproductive performance of goats (Makembe and Ndlovu, 1996), a response which was attributed to increased opportunity for lablab selection. Legume ( Calliandra calothyrsus and Macrotyloma axillare ) supplementation increased feed intake, N intake, fecal production, and fecal-N concentration of steers compared to thos e fed only on a basal diet of barley ( Hordeum vulgare ) straw (Delve et al., 2001). In an experiment to evaluate the effects of forage type and level of concentrate supplementation, cows fed on a maize-lablab, fo rage-based diet consumed more organic matter, and had a higher apparent digestibil ity of organic matter than those consuming oat-vetch, forage-based diet (Khalili et al., 1994 ). Forage type did not affect daily milk yield, milk fat concentration or total solids in milk, except fo r milk protein concentration. Forage Yield and Nutritive Value of Mucuna and Lablab Ravindran (1988), working in Sri Lanka, evalua ted mucuna at four growth stages and found the optimum harvesting time to be around 90 d after planting, at about the onset of flowering. At this time th e dry matter yield was 3.1 t ha-1, CP concentration was 206 g kg-1, and in vitro organic matter digestibility (IVOMD) was 554 g kg-1. Similar CP results have been reported by Singh and Re lwani (1978) and Adjo rlolo et al. (2001), indicating its potential as prot ein supplement to low quality roughages. In an experiment using sheep to determine the nutritive value of mucuna forage at flowering stage (3 mo after planting), Adjorlolo et al . (2001) showed that mucuna whole plant forage had CP, NDF, and lignin concentrations of 200, 544, and 96 g kg-1 respectively, and that mucuna forage supplementation increased the CP dige stibility, in sacco de gradation rate, and

PAGE 56

41 particle outflow rate of both DM a nd NDF of sodium hydroxide-treated rice ( Oryza sativa ) straw. There is a wide variation in the dry matter yi eld and CP concentration of lablab forage depending on location and stage of harvesti ng. In Australia, Wood (1983) reported that lablab produced herbage yield of 8.6 t ha-1 at flowering which was comprised of 3.6 t ha-1 leaf containing 231 g kg-1 CP and 5 t ha-1 of stem containing 69 g kg-1 CP. In the midsouth of the USA, lablab at full bloo m stage produced a DM yield of 5.2 t ha –1 and contained 146, 471, and 352 g kg-1 of CP, NDF, and acid de tergent fiber (ADF), respectively (Fribourg et al ., 1984). Kiflewahid and Mosimanyana (1987) from Botswana reported average yields of 1.23 to 1.44 t DM ha-1 on smallholder dairy farms, and CP concentration of 161 g CP kg-1 when lablab was planted and harvested along with maize and sorghum. In Zimbabwe, sun-dried lablab at 8 wk growth (pre-anthesis) contained 250, 370, 89, 7.2, and 1.1 g kg-1 of CP, NDF, acid detergent lignin (ADL), Ca, and P, respectively, with a DM degradation of 842 g kg-1 (Mupangwa et al., 1997). Abule et al. (1995) showed th at CP, NDF, Ca, and P concentr ations of sun-dried lablab hay were 186, 420, 14, and 1.9 g kg-1, respectively. Maasdorp and Titterton (1997) reported a dry matter digestibil ity (DMD) of mucuna and labl ab at early green pod stage when maximum biomass occurred (16 and 18 WAP, respectively) of 500 and 588 g kg-1 respectively. At this time the CP concentration was 182 and 164 g kg-1 for mucuna and lablab, respectively. Lablab hay at 50 % flowering stage had IVOMD of 600 g kg-1 (Nsahlai et al., 1998). Seeds of lablab had a CP concentration of 249 g kg-1, with well balanced amino-acids for humans (Chau et al., 1998). Mucuna a nd lablab sampled at 112 d after planting

PAGE 57

42 (DAP) had lower N concentration (25.4 and 24.5 g kg-1, respectively) than when sampled at 91 DAP when N concentration was 33 and 38.3 g kg-1, respectively (Carsky et al., 1999). In a more recent study to evaluate the effect of variable harvest dates on biomass and quality, Agyemang et al. (2000) found that lablab hay harvested at 100, 114, 128, and 142 DAP had biomass yield of 1.28, 1.97, 2.04, and 1.99 t DM ha-1, respectively, CP concentration of 138, 175, 144 and 131 g kg-1 DM, respectively, and NDF concentration of 503, 546, 572, and 576 g kg-1 DM, respectively. The hay harvested at 114 DAP had the highest CP concentration, fastest ra te of degradation, and about 50% rumen degradable protein. Antinutritive Factors Although mucuna seed has a high protein conc entration and its quali ty is comparable to that of soybean (Ravindran and Ravindra n, 1988), it contains a toxic chemical [3, 4dihydroxyphenyl alanine (Levodopa, or L-Dopa)] (Lorenzetti et al ., 1998; Siddhuraju and Becker, 2001). Evaluating 36 accessions, Lo renzetti et al. (1998) found L-Dopa concentration in the seeds to range from 22 to 62 g kg-1 of DM. Lorenzetti et al. (1998) found that the stippled seed accessions had si gnificantly lower L-Dopa concentration than the black, speckled, or white, but Siddhuraj u and Becker (2001) found contradicting results where the white variety had higher L-Dopa concentration (49.6 g kg-1) compared to the black variety (43.9 g kg-1). The concentration has also been shown to vary with latitude and seeds grown within 100 of the equator contai ned significantly higher concentrations. L-Dopa can cause toxic effects to humans (In fante et al., 1990) if consumed at levels above 1.5 g d-1 (Lorenzetti et al., 1998). It has been reported that the oxidation products of L-Dopa conjugate with SH group of pr oteins (cystein) forming a protein bound 5-S-

PAGE 58

43 cysteinnyldopa cross links, l eading to polymerization of proteins (Takasaki and Kawakishi, 1997). A procedure to prepare detoxified mucuna flour is available (Versteeg et al., 1998b), but this may demand extra labor. Mucuna seed was extensively used in the southern USA as part of a ration for cows at the beginning of the la st century (Tracy and Coe, 1918) and no toxic effects were observe d, suggesting that the L-Dopa in mucuna forage may have minimal detrimental effects to ruminant animals. Mucuna and lablab contain ot her anti-nutritional factors su ch as polyphenols, tannins, trypsin inhibitor activity, cyanogenic glycosid es, and hemaglutanati ng activities (Rajaram and Jonardhanan, 1991). Tannins are naturally occurring plant polyphenols that form strong complexes with proteins. They are usually subdivide d into two groups: hydrolysable tannins (water soluble) and proathocyanidins (condensed tannins) (Reed, 1995). Tannins in forage legumes have both negative and positive effects on their nutritive value. Tannins in high concentratio ns reduce intake, digest ibility of proteins and carbohydrates, and animal performance (Reed et al., 1990; Tanner et al., 1990). Tannins in low to moderate concentrations, especially condensed tannins, prevent bloat and increase the flow of non-ammonia N a nd essential amino acids from the rumen (Woodward and Reed, 1997). The presence of condensed tannins is reported to reduce N degradation in the rumen th rough the formation of tanninprotein complexes which are stable at neutral rumen pH, but do cleave at the low gastric pH (2.5-3.5) of the abomasum and relatively high pH (8-9) of the distal small intestines (Salawu et al., 1997). Dry Matter Intake and Animal Performance The dry matter intake (DMI) of feed eaten is an important factor controlling ruminant production. Obeid et al. (1992) observed that DMI of mai ze-mucuna silage was lower compared to that from pure maize silage. Pachauri and Upandhayaya (1982) found an

PAGE 59

44 increase in live-weight gain of goats fed hay from muc una fodder. A higher ADG of 0.38 kg was obtained from cattle feeding on mai ze-mucuna silage compared to 0.25 kg from cattle on pure maize silage, but the gain was lower than those for maize-soybean (0.68 kg) and maizeCrotalaria juncea (0.7 kg) mixtures (Obeid et al., 1992). A short-term observational study in Nigeria showed that milk production from cows consuming silage consisting of 75% maize forage and 25% mucuna was 8.22 kg head-1 d-1compared to 7.5 for those consuming silage made from elephantgrass alone (Umoh, 1975). Supplementation with mucuna forage harves ted at flowering to sheep feeding on rice straw was shown to marginally increase DMD and OMD, and significantly increase degradation rate of both DM a nd NDF, and also increase partic le outflow rate (Adjorlolo et al., 2001). Graham et al. (1986) indicated that steers gr azing lablab at a stoc king rate of 1.7 steers ha-1 gained 50 to 60 kg liveweight during 86 d of grazing. They also found that when buffelgrass ( Cenchrus ciliaris) was combined with lablab in varying proportions, liveweight gains were directly related to the amount of lablab available, but no further increase in gain resulted when lablab was more than 67% of the diet. McLeod et al. (1990) reported a higher voluntary intake of lablab hay by Hereford steers compared to that of grass hay and that the intake of the leaf fraction was higher th an that of the stem fraction. Lablab supplementation to calves or cows fed tef straw increased the rate of degradation of tef straw, decreased mean re tention time, and total DMI increased with increasing supplementation (Abule et al., 1995). A diet c ontaining 50:50 maize stover:lablab compared to a diet of 70:30 ma ize stover: lablab fed to goats had higher intake (55 vs. 53 g kg0.75), ADG (22.2 vs. 6.3 g d-1) and milk yield (884 vs. 434 g d-1),

PAGE 60

45 indicating the potential of la blab in improving animal perf ormance when supplemented to maize stover-based diets (Makembe and Ndlovu, 1996). In a field experiment (Milera et al., 1989), fresh lablab forage (30%) and gui neagrass silage (70 %) supplementation to crossbred cows grazed for 4 h d-1 significantly increased milk production (10.7 kg cow-1 d-1) compared to when guineagrass was supplemented alone (9.2 kg cow-1 d-1). Based on farmersÂ’ evaluation using forage yield, animal responses, and milk quality of cows fed lablab forages as ranking criteria (ranke d on a scale of 1-4), Agyemang et al. (2000) concluded that forage harvested at 114 and 128 DAP had the best utilization rank score. Although information on chemical composition is available, the use of mucuna and lablab as feed for cattle has not been studi ed in detail and information is scanty, particularly on their effect as supplement for lactating cows feeding on low quality fodder. Based on the soil fertility and fodder qua ntity and quality constraints facing smallholder farmers in western Kenya and th e previous research described in this literature review, a series of experiments was designed. The objectives were to 1) quantify the N contribution from mucuna and lablab residue incorporation to succeeding maize when all biomass is incorporated and when part of the residue is removed as livestock feed, 2) determine the effect of alternative cropping systems (legume relay intercrop), defoliation regime, and number of years (1 yr vs. 2 yr) of residue incorporation on soil fertility as measured th rough yields of a subs equent test crop of maize and bean, and 3) evaluate fodder pr oduction and nutritive value of legumes in alternative cropping systems and assess their va lue as a supplement to lactating cows

PAGE 61

46 during the dry season. The results of these st udies are described in the chapters that follow.

PAGE 62

47 CHAPTER 3 NITROGEN CONTRIBUTION FROM RELA Y-CROPPED MUCUNA AND LABLAB TO MAIZE IN NORTHWESTERN KENYA Introduction Low soil fertility is a major constraint to crop production in smallholder farming systems in northwestern Kenya. High population pressure leads to continuous cultivation and when little or no fertilizer is used soils become depleted of nutrients, with N and P the most limiting (Smaling et al., 1997). A lthough soil fertility can be improved with inorganic fertilizers, they ar e not widely used by smallholder farmers due to lack of financial resources. The integrated use of inorganic fertilizers and organic residues has been shown to be a useful al ternative for increasing crop productivity in these regions (Palm et al., 1997). Jama et al. (1997), work ing in western Kenya, indicated that it was economically attractive to integrate inorgani c P fertilizer (triple super phosphate) with organic materials having a high N-to-P ratio. Herbaceous legumes are examples of this type of material and they can be used in association or rotation with crops in green manure/cover cropping system s (Ibewiro et al., 2000a). When organic residues are incorporated into the soil, they release N during decomposition, however for effective decompos ition, the release of nutrients should be synchronized with the demand of the recipien t crop (Myers et al., 1994). Most studies evaluating synchrony of N release have used above-ground leafy materials (Handayanto et al., 1994; Myers et al., 1994; Palm, 1995). Little information is available on the synchrony of N release and demand when root residue or root and low quality stem

PAGE 63

48 stubble are incorporated for th e subsequent maize crop. Prev ious studies have indicated that residues of intermediate quality may result in nutrient release patterns in synchrony with crop nutrient demand (Handayanto et al., 1994; 1995; 1997; Mafongoya et al., 1997a; 1997b). Myers et al. (1994) f ound that a mixture of straw and Gliricidia sepium residue mineralized N at an intermediate ra te, and that the rate gradually increased coinciding with the period when growth dema nd by maize was highest , showing that the release of N from plant residues can be regulated through the manipulation of quality (Handayanto et al., 1997). Decomposition and N release of organic resi dues in the soil are influenced by biomass yield (Sanginga et al., 1996a), chemical co mposition (Palm and Sanchez, 1992; Tian et al., 1992), and by abiotic factors such as cl imate and soil conditions (Mugendi et al., 1997). Several studies have shown that re sidue concentrations of N, lignin, and polyphenols, and the ratios of these constituents are useful indices of residue quality and affect decomposition and N release (Vanla uwe et al., 1997; Maf ongoya et al., 1998). Mucuna ( Mucuna pruriens (L.) DC. var. Utilis (Wright) Bruck) and lablab ( Lablab purpureus (L.) Sweet cv. Rongai) are promising legume green manures that have been successfully used in cover cropping systems with maize ( Zea mays L.) in many parts of the tropics. They have been shown to cont ribute N (Ibewiro et al ., 2000a), increasing the yields of subsequent maize (Versteeg et al., 1998a; Ibewiro et al., 2000a; Tian et al., 2000). In a rotational cropping system in the sub-humid highlands of East Africa, Fischler (1996) showed that preceding mucuna or lablab increased the maize grain yield by 50 or 40%, respectively. Working in a deri ved savanna of West Africa, Ibewiro et al. (2000a) found that relative to the controls, mucuna and lablab residues decomposed

PAGE 64

49 rapidly, releasing 50 and 64%, respectively, of their N by 28 d after pl anting of maize and increased dry matter and N upt ake of subsequent maize. Although use of green manures has been re ported in Africa (Balasubramanian and Blaise, 1993; Drechsel et al., 1996; Manyong et al., 1996; Snapp et al., 1998), the adoption of herbaceous legumes there is generally low (Thomas and Sumberg, 1995; Drechsel et al., 1996). Adoption of green manures has been much more rapid when the legumes had uses in addition to soil fertility improvement (Versteeg et al., 1998a). One such use is as a livestock feed (Abule et al, 1995; Agyemang et al ., 2000, Adjorlolo et al., 2001), but there has been little research evaluating multiple uses of legumes for improvement of soil fertility and livestock diets. In evaluating the potential of herbaceous legumes for relay-cropping systems with maize and beans in the sub-humid highl ands of northwestern Kenya, a clearer understanding is needed of legume biomass yiel d and the chemical attributes of herbage that influence residue decomposition, N rel ease, and uptake by maize. The specific objectives of this study were to 1) characte rize the chemical composition of mucuna and lablab, 2) quantify the N contribution from in corporated mucuna and lablab residue to succeeding maize when either all biomass is incorporated or part of the residue is removed as livestock feed, and 3) relate residue chemical composition of mucuna and lablab with mineral N in the soil, N uptak e, and grain yield of the succeeding maize intercrop.

PAGE 65

50 Materials and Methods Experimental Site The experiment was conducted in western Ke nya at the Kenya Agricultural Research Institute’s National Agricultura l Research Centre, Kitale (10 N and 350 E, altitude 1860 m), Kenya during the 1999 and 2000 growing seasons. Located in the sub-humid highlands, the area has a unimodal rainfall pattern lasting from mid-March to midNovember. This pattern allows for one late maturing hybrid maize and two shortduration crops of common bean ( Phaseolus vulgaris L.) intercropped with maize. Maize and the first crop of beans are planted in April, and the second crop of beans is planted in August. In this study, maize and bean were planted in April 1999, but mucuna and lablab substituted for the second crop of beans in August 1999. The legumes grew through the dry season until soil incorporation in March 2000. Data reported for maize and bean are those for the season following that inco rporation, i.e., April to November 2000. The rainfall during the e xperimental period was 1100 mm and the mean monthly minimum and maximum temp eratures were 12 and 240C, respectively. The soils are classified as humic Ferrolsols based on the FAO/UNESCO system (FAO – UNESCO, 1994) and are equivalent to a Kandiudalfic Eutaudox in the USDA soil taxonomy system (Soil Survey Staff, 1991). These are deep, highly weathered, and well-drained loam soils that are dark red to dark reddish brown in color with low activity clay and moderately acid topsoil. The topsoil (0-20 cm) ha d the following properties; pH (1:2.5 H2O), 5.4; organic C, 14.2 g kg-1; total N, 1.3 g kg-1; extractable P (modified Olsen), 9.7 mg kg-1; and texture, clay loam with 39% clay, 41% sand, and 20% silt.

PAGE 66

51 Treatments and Cropping Systems Seven treatments were evaluated. Four treatments consisted of the two legumes (mucuna and lablab) intercropped with mai ze and the legumes were either undefoliated with their herbage incorporated into the soil before subsequent maize or defoliated to a 10-cm stubble with topgrowth removed for liv estock feed and the remaining residue incorporated. The three control treatments we re the maize-bean-bean system receiving 1) inorganic N fertilizer (30 kg N ha-1), 2) no N fertilizer, and 3) cattle manure at half the local recommended rate of 5 t ha-1 (supplying approximately 65 kg N and 18 kg P ha-1). The treatments were laid out in a randomi zed complete block design replicated three times. Experimental plots were 4.5 by 6 m with a 1-m border around each plot. The experiment was initiated at the begi nning of the growing season in April 1999. Maize was planted at an interand intra-ro w spacing of 75 by 30 cm, respectively, using two seeds per hill of Kitale hybrid 614D maize gr own in the region. The maize seedlings were thinned to one plant per hill 4 wk afte r planting (WAP) to give a plant population of 44,440 plants ha-1. At planting, triple super phosphate fertilizer (0-46-0) was applied in the same hill as for maize at a rate of 13 kg P ha-1 to all treatments except for cattle manure. The first crop of common bean was planted simultaneously with maize at onset of rains in April. An improved bean variety GLP2 (Rosecoco), commonly planted by farmers in the region, was used. The second crop of beans and the legume green manures were relay cropped in maize in August, 135 d after maize planting. Before planting of the second bean crop or the legume intercrop (aft er harvesting the first crop of beans), the maize was weeded and all leaves below the ear were removed to minimize shading. The common bean, mucuna, and lablab were plan ted between maize rows at an intra-row spacing of 30 cm using three seeds per hill. Th ey were thinned to two plants per hill 4

PAGE 67

52 WAP. All the plots were hand weeded twice before harvesti ng the first crop of beans. The plots were manually weeded once after th e August planting of le gumes. Stalk borer ( Chilo spp), a common pest of maize, was cont rolled by application of Beta-cyfluthrin (Bulldock). Beta-cyfluthrin, a synthetic pyrethroid insecticide, was applied in granular form into the whorl of each plant at a rate of 7 kg ha-1 at 4 and 8 wk after germination. Green Manure Defoliation Management and Sampling After harvesting maize in mid-November 1999, all stover was removed in accordance with the farmersÂ’ practice. Mucuna and labl ab were left to continue growing until land preparation for the 2000 growing season in mi d-March. The legumes were sampled in mid-March (30 WAP) before the residue wa s incorporated. The biomass production was assessed in terms of litter fall, above-ground biom ass (leaf and stem), and root. For plots of the defoliated treatment, legume herbage was cut to a stubble of 10 cm. Prior to clipping the entire plot, two representative 0.5-m2 quadrats were sampled. Herbage above 10 cm was composited across the two si tes per plot and separated into weeds (grass and broadleaf), legume leaf, and legume stem fracti ons. Fractions were dried, weighed, and ground for laboratory analysis. In these same quadrats, all material below 10 cm was removed at soil level, composited across the two sites per plot, and separated into weed, live legume leaf, live legume stem, and litter fractions. Fresh weights of each fraction were taken, fractions were subsampled, and the rema inder was returned to the two sampling sites. The subsample was we ighed fresh, dried, weighed again, and then ground for analysis. Undefoliated treatment plots were sampled at the same time. Again two representative 0.5-m2 quadrats were sampled. In this case, all material was removed at soil level, composited, and separated into weed, live legume leaf, live legu me stem, and litter

PAGE 68

53 fractions. Fresh weights were taken, frac tions were subsampled and the subsample weighed, and the remainder was returned to the quadrat. The subsample was dried, weighed, and ground for analysis. Within each quadrat of all legume treatments, a root sample was taken. A soil core from a 0.5-m2 area was removed to a depth of 20 cm. All visible roots were removed from the soil and washed with water on top of a 0.5-mm sieve to remove all soil and then rinsed with distilled water. The roots were dried at 700C for 48 h and weighed before grinding for analysis. Weed biomass yiel d was determined from two quadrats and samples for chemical analysis we re taken as described above. Soil Mineral N Sampling Soil samples were taken from all seven treatments to determine soil mineral N before residue incorporation (BI) and at six dates af ter planting maize and bean in April 2000 to monitor mineral-N release in relation to the maize growth phase. Sampling dates were 4, 8, 10, 15, 21, and 27 WAP (at maize harvest). At each sampling date eight 10-cm cores were taken per plot to a depth of 20 cm. Th e soil from the eight sampling points per plot was composited, thoroughly mixed, and then s ubsampled and packed into polyethylene bags. The samples were kept in a cooler box with ice and transported immediately to the laboratory for analysis. Th ey were extracted with 2 N KCl. The extracts were analyzed for NH4 + and NO3 following the method outlined by A nderson and Ingram (1993). The NH4 + and NO3 were calculated to elemental N and then summed to give soil mineral N. Soil nitrate was estimated in kg ha-1 using the measured bulk density of 1.25 g cm-3 for the 0 to 20-cm depth.

PAGE 69

54 Maize N Uptake Sampling For determination of maize DM yield and N uptake, six representative maize plants were harvested (cut at soil su rface without harvesting roots) from the two rows next to each outer border row at 2, 4, 8, 10, 15, 21, and 27 WAP (at harvest) in 2000. To minimize gap effects, care was exercised to en sure that two consecutive plants were not removed. At each sampling date, plant subsamples were taken for DM and N concentration determinations. The samples we re washed with wate r (where necessary) and oven dried at 600C to constant weight to determin e DM. The samples were ground to pass a 1-mm sieve and kept in airtight plastic bags in a cool, dry place until ready for chemical analysis. Grain and Stover/Straw Dry Matte r Yield of maize and Beans The first crop of common bean was harveste d at the end of July and the second crop of common bean and the maize were harvested in mid-November. Maize and bean were harvested from the two center rows of every plot . The first and last two hills of each row were not harvested. The size of the sampling unit was 1.5 by 5.4-m (8.1 m2). Grain yield at 13.5% moisture concentration and st over/straw DM yields were recorded. Chemical Analyses Legume fractions from below 10 cm of defoliated plots and all fractions from undefoliated plots were analyzed for N, P, K, lignin, and polyphenol. Total N, P, and K were determined using the Kj eldahl digestion with concen trated sulphuric acid followed by colorimetric determination of N and P and flame photometric determination of K (Anderson and Ingram, 1993). Ne utral detergent fiber and lignin were determined using the methods of Van Soest et al. (1991). Total polyphenols were extracted using 50% aqueous methanol with a plant to extract ratio of 0.1/50 mL and phenols analyzed

PAGE 70

55 colorimetrically using Folin-Ciocalteu reag ent (Anderson and Ingram, 1993). The soil samples were analyzed for soil pH in wate r (soil:water ratio of 1:2.5), total N, NO3 -, NH4 +, P, K, organic carbon, and soil particle size using procedures outlined by Anderson and Ingram (1993). Statistical Analyses The general linear models pr ocedure of SAS was used to test treatment effects on plant and soil responses (SAS, 2001). For res ponse variables pertaini ng only to the four legume treatments, the model included e ffects of legume, defoliation, and their interaction. For response variables pertaining to all seven treatments in the study, a first analysis tested the effects of legume, defoliation, and thei r interaction, and this was followed by single degree of freedom contra sts that compared the legume treatments (individually or in groups) w ith the controls. Relationshi ps between residue chemical composition and soil mineral N were determined using regression analysis in the regression procedure of SAS (SAS, 2001). Di fferences referred to in the text are significant at P < 0.10, unless otherwise indicated. Results and Discussion Legume Residue Biomass Defoliation reduced leaf, stem, and total le gume residue as well as the whole residue incorporated (Table 3-1). Legume leaf biomass was greater for undefoliated than defoliated treatments by a factor of five for mu cuna and 26 for lablab. Stem biomass also was greater for undefoliated than defoliated treatments, but only by a factor of three. There were no differences between legume sp ecies for either leaf or stem biomass although there was a trend toward greater leaf mass for mucuna than lablab (P = 0.147).

PAGE 71

56Table 3-1. Residue biomass of va rious fractions of mucuna and lablab, weeds, a nd whole residue when legumes were relay cropped in maize. Treatments† Effect Fraction UD-M D-M UD-L D-L Legume (L) Defoliation (D) L x DSE Legume ---------------------------t ha-1 -------------------------------------P Values -------------Leaf 1.07 0.22 0.78 0.03 0.147 0.002 0.741 0.14 Stem 0.75 0.27 0.91 0.31 0.475 0.006 0.663 0.10 Litter 0.12 0.13 0.17 0.19 0.188 0.723 0.859 0.02 Roots 0.11 0.07 0.15 0.17 0.108 0.776 0.406 0.02 Total legume 2.05 0.69 2.0 0.70 0.943 0.005 0.934 0.25 Weeds 0.60 1.10 0.59 0.80 0.621 0.275 0.636 0.13 Whole residue 2.65 1.79 2.60 1.50 0.662 0.043 0.765 0.23 Shoot : Root ratio 16.9 11.1 13.5 3.2 0.038 0.009 0.333 1.74 Leaf : Stem ratio 1.43 0.85 0.82 0.14 0.002 0.003 0.679 0.15 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated labla b, D-L = Defoliated lablab

PAGE 72

57 Legume litter and roots averaged 0.15 and 0.13 t ha-1, respectively, and were not affected by treatment. Total legume biomass incorporated into the soil prior to planting maize was nearly identical for the two species, but was approximately three times greater for undefoliated than for defoliated treatments (Table 3-1). Leaf:stem ratio and shoot:root ratio of in corporated biomass were greatest for undefoliated mucuna (UD-M) and least for defo liated lablab (D-L), the response of the latter treatment was due in large part to very low leaf biomass. The presence of legumes in the cropping system reduced weed growth from 2.2 t ha-1 in the unfertilized natural fallow control to an average of 0.8 t ha-1 in the legume intercr op treatments, a response similar to that described by Becker and Johnson (1998). The grass weeds comprised mainly couchgrass ( Digitaria scalaraum ) and the broadleaf weeds were black jack ( Bidens pilosa ), wondering jew ( Commelina bengalensis ), and Gallant soldier ( Galinsoga parviflora ). When grown as monocrops and planted at the beginning of the growing season (April) in a side experiment at this location, mucuna and lablab produced greater than 5 t DM ha-1 compared to about 2 t ha-1 when relay cropped in maize and planted in August. This level of yield reduction is consistent with pr evious reports when mucuna and lablab were grown in association with cereal crops (Fis chler, 1996; Sanginga et al., 1996a; Wortmann et al., 2000) and can be attributed primarily to a shift in the season of growth to include the dry season. In addition, germination of mucuna was slow and uneven and the lablab seedlings were susceptible to aphid ( Aphis craccivora ) damage and leaf rust (Anthracnose, caused by Colletotrichum spp .) when planting occurred during August when rainfall was quite high. The ability of these legumes to remain alive and grow

PAGE 73

58 during the dry season when intercropped with maize is important. Although biomass yields were only 2 t ha-1, the great majority of this material was living at time of incorporation resulting in residue that was less affected by weathering and leaching losses of nutrients than would be th e case if litter was the primary fraction being incorporated. Chemical Composition of Legume Biomass Leaf N concentration was affected by a legume species by defoliation regime interaction (P = 0.05; Table 3-2). Legume sp ecies had no effect on leaf N of undefoliated treatments (P = 0.444), but defoliated mucuna (D -M) leaf N was greater than that of D-L (P = 0.016). Likewise there was no effect of defoliation on muc una leaf N, but leaf N of undefoliated lablab (UD-L) was greater than D-L. Litter N concentration was not affected by treatment and was approximately 10 g kg-1 less than for leaves, except for DL where the difference was only 1.3 g kg-1. The lower N concentration of litter than leaf can be explained by the process of nutrient resorption during senescence, when leaf proteins are hydrolyzed and pr oducts transported to other pa rts before leaf fall (Norby and Contrufo, 1998). There also was interaction of legume species and defoliation regime for stem N concentration (P = 0.043). Stem N concentration of undefoliated treatments was not affected by legume species (P = 0.657), but D-M stem (16.2 g kg-1) contained higher (P = 0.003) N concentration than D-L (10.2 g kg-1). Defoliation reduced lablab stem N concentration (P = 0.034), but had no effect on mucuna stem (P = 0.227). Mucuna roots had greater (P = 0.039) N concen tration than lablab. The N concentration of mucuna and lablab fractions in the current study were generally lower than those reported by Ibewiro

PAGE 74

59Table 3-2. Nitrogen, lignin, and pol yphenol concentration of various residue fractio ns of mucuna and lablab relay cropped in m aize. Treatments† Effects Fraction UD-M D-MUD-LD-L Legume (L) Defoliation (D) L x D SE Nitrogen ---------------------------g kg-1 -----------------------------------------------P valu es --------------Leaf 22.1 25.124.616.7 0.224 0.308 0.050 1.32 Stem 14.5 16.213.810.2 0.016 0.344 0.043 0.79 Litter 12.1 13.213.715.4 0.121 0.234 0.758 0.52 Roots 13.4 13.311.611.3 0.039 0.754 0.858 0.54 Lignin Leaf 52 57 35 99 0.302 0.028 0.056 8.64 Stem 77 96 89 109 0.019 0.004 0.921 4.00 Roots 132 121108 108 0.018 0.454 0.489 3.90 Polyphenol Leaf 45 552715 < 0.001 0.778 0.01 5.15 Stem 36 37138 0.004 0.876 0.495 4.53 Roots 23 2765 0.010 0.913 0.607 3.47 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated labla b, D-L = Defoliated lablab

PAGE 75

60 et al. (2000a). This could be due to differe nces in environmental conditions prevailing during plant growth, which can result in biom ass of differing quality (Handayanto et al., 1995), and the stage at which the residue was sampled. Biomass in our study was sampled at the end of the dry season (30 WAP) when some leaves had been lost as litter and the above-ground herbage was compri sed of older leaves and stems. There was an interaction effect of legu me and defoliation regime on leaf lignin concentration (P = 0.056). Defoliation increase d lablab leaf lignin concentration (P = 0.007), but had no effect on mucuna leaf (P = 0.743). Defoliated lablab leaf had greater (P = 0.021) lignin concentration than D-M, but there was no difference (P = 0.386) between undefoliated treatments. Mucuna st em contained less (P = 0.019) lignin than lablab stem, and defoliation increased lignin concentration of the stem fraction of both legumes (P = 0.004) (Table 3-2). Mucuna roots contained higher lignin concentration than lablab (P = 0.018), but defoliation did not affect root lignin concentration. In agreement with literature (Tian and King, 1998), roots had higher lignin concentration than other residues. The ligni n concentration in mucuna a nd lablab residues was below 150 g kg-1, a concentration above which decompos ition and N release is thought to be impaired (Palm et al., 2001) due to lignin protection of cell wall constituents from microbial attack (Cheson, 1997). Relatively large differences in polyphenol concentration occurred between species (Table 3-2). All mucuna fractions had hi gher (P <0.01) polyphenol concentration than those of lablab. Polyphenol c oncentration ranged from 5 g kg-1 in D-L roots to 55 g kg-1 in D-M leaves and concentrations in the plants were generally in the order: leaf > stem > roots. There was legume by defoliation tr eatment interaction (P = 0.01) for leaf

PAGE 76

61 polyphenol concentration. Interaction occurr ed because the magnitude of the species difference was greater for defoliated than undefoliated leaf. The legumes contained higher concentrations of polyphenol than t hose reported by Ibewiro et al. (2000a), but these were within the range reported for ag roforestry species by Constantinides and Fowness (1994) and were above concentrations (> 4 g kg-1) at which decomposition is reduced (Palm et al., 2001). Mucuna stem contained higher (P = 0.033) P concentration than lablab stem, and defoliation reduced P concentration of stems of both legumes (P = 0.049) (Table 3-3). Root P concentration was affected only by le gume species and was lo wer for lablab (0.85 g kg-1) than mucuna (1.35). Legume species eff ects on K concentration were significant for leaf, litter, and root fractions. Lablab le af and stem had greater K concentration than mucuna, but the opposite was true for roots (T able 3-3). Defoliation had no effect on K concentration of any fraction. The P concen tration of mucuna a nd lablab above-ground residue was similar to that reported from the sub-humid highlands of eastern Africa (Wortmann et al., 2000) and was close to the critical range for net P mineralization of between 2 and 3 g kg-1 reported by Singh et al. (1992), s uggesting a net P mineralization due to the application of green manure residue. Nutrient Mass Incorporated Defoliation reduced the N content in legum e leaf (P = 0.002), stem (P = 0.009), and total legume (P = 0.001) fractions, and in the whole residue (P = 0.018) that was incorporated (Table 3-4). Nitrogen cont ribution from UD-M and UD-L treatments was similar (49 and 47 kg N ha-1, respectively), but that of D-M was higher (P=0.089) than

PAGE 77

62 Table 3-3. Phosphorus and potassium concentrations of various residue fractions of mucuna a nd lablab relay cropped in maize. Treatments† Effects Fraction UD-M D-MUD-LD-L Legume (L) Defoliation (D) L x D SE --------------------------g kg-1 ------------------------------------------P valu es ----------------Phosphorus Leaf 1.6 1.51.80.9 0.427 0.119 0.217 0.15 Stem 2.4 1.51.40.7 0.033 0.049 0.850 0.22 Litter 0.8 0.70.80.7 0.780 0.194 0.414 0.03 Roots 1.5 1.20.90.8 0.009 0.281 0.534 0.10 Potassium Leaf 6.7 8.415.912.5 0.007 0.674 0.138 1.23 Stem 13.4 14.516.515.0 0.170 0.474 0.312 0.52 Litter 5.6 5.710.613.0 0.001 0.305 0.318 1.04 Roots 10.8 11.79.49.2 0.046 0.728 0.506 0.43 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated labla b, D-L = Defoliated lablab

PAGE 78

63 Table 3-4. Nitrogen contribution of va rious residue fractions of mucuna and lablab relay cropped in maize. Treatments† Effects Fraction UD-M D-MUD-LD-L Legume (L) Defoliation (D) L x D SE Legume ---------------------------kg ha-1 -------------------------------------------P valu es ---------------Leaf 26.1 5.619.00.5 0.092 0.002 0.565 3.59 Stem 8.8 4.212.43.2 0.609 0.009 0.260 1.45 Litter 1.5 1.72.32.9 0.133 0.547 0.764 0.28 Roots 1.5 1.01.81.8 0.270 0.622 0.547 0.22 Total legume 37.9 12.535.58.4 0.499 0.001 0.845 4.68 Weeds 10.7 19.511.713.1 0.577 0.316 0.450 2.02 Whole residue 48.6 32.347.221.5 0.385 0.018 0.496 4.61 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 79

64 D-L. The natural weed fallow control contributed 40 kg N ha-1. Defoliated mucuna and lablab treatments contributed 66 and 45% as much N as the undefoliated treatments, respectively. For UD-M and UD-L, the leaf fraction contributed the greatest amount of N (68 and 54% of the total legume, respectively) and root contributed the least (4 and 5%, respectively). In the D-L treatment, stem contributed 38% compar ed to only 6% from leaves, while for D-M the contributions of st em and leaf were about the same (35 and 44%, respectively). Defoliation reduced the P and K contribution of leaf, stem, and total legume fractions, but nutrient contribution from the whole residue was reduced only for P (Table 3-5). The green manures contributed low amounts of P, ranging from 0.5 (D-L) to 3.5 kg ha-1 (UDM). The K contribution ranged from 7.4 (D-M) to 31.7 kg ha-1 (UD-L). Soil Mineral N The bulk of the inorganic N found in the soil at all sampling periods was in the form of NO3 --N. The NO3 --N fraction was three times that of NH4 +-N. Mineral N in the 0to 20-cm soil depth at the beginning of the seas on before residue inco rporation ranged from 16 kg ha-1 in the inorganic N plot to 20 kg ha-1 in UD-M plots, with a mean of 18 kg ha-1 (Figure 3-1). Throughout the period from planting through 27 WAP, there were no differences among treatments in soil mineral N at any sampling date. The mean amount of soil inorganic N across all treatments d ecreased progressively until 8 WAP. The soil inorganic N marginally increased between 8 and 21 WAP (grain filling stage) and decreased at harvest (27 WAP). Regressi on of soil inorganic N averaged across all treatments over time showed a cubic response (Figure 3-1).

PAGE 80

65Table 3-5. Phosphorus and K c ontent of various residue fractions of muc una and lablab relay cropped in maize. Treatments† Effects Fraction UD-M D-M UD-L D-L Legume (L) Defoliation (D)L x D SE Phosphorus ----------------------kg ha-1 --------------------------------------P values ----------------Leaf 1.8 0.3 1.5 0.1 0.262 0.006 0.712 0.28 Stem 1.5 0.4 1.4 0.2 0.830 0.040 0.796 0.26 Litter 0.1 0.1 0.1 0.1 0.180 0.884 0.907 0.01 Roots 0.2 0.1 0.1 0.1 0.969 0.327 0.483 0.02 Total legume 3.5 0.9 3.1 0.5 0.547 0.006 0.970 0.51 Weeds 1.1 1.6 1.1 1.1 0.568 0.501 0.572 0.17 Whole residue 4.6 2.5 4.2 1.6 0.479 0.036 0.766 0.55 Potassium Leaf 7.8 1.8 13.1 0.4 0.574 0.012 0.296 2.00 Stem 8.1 4.1 15.3 4.6 0.204 0.031 0.267 1.84 Litter 0.7 0.7 1.8 2.4 0.026 0.528 0.571 0.29 Roots 1.2 0.8 1.4 1.5 0.224 0.605 0.503 0.16 Total legume 17.8 7.4 31.7 8.9 0.156 0.013 0.238 3.61 Weeds 18.3 31.9 16.5 20.9 0.497 0.349 0.621 3.81 Whole residue 36.1 39.3 48.2 29.8 0.901 0.467 0.312 4.56 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated labla b, D-L = Defoliated lablab

PAGE 81

66 0 5 10 15 20 25 051015202530 WEEKS AFTER PLANTING (WAP)SOIL INORGANIC N (kg ha-1) Soil N = 19.38 3.35WAP + 0.25WAP2 0.005WAP3 R2 = 0.89; P < 0.014 Figure 3-1. Mean inorganic N in th e soil at different time periods.

PAGE 82

67 Maize Yield and N Uptake There was an interaction of legume by defoliation for above-ground maize biomass yield (P = 0.037) at 4 WAP (Table 3-6). U ndefoliated mucuna tended to yield greater (P = 0.11) maize biomass than UD-L, but there we re no differences (P = 0.492) between the defoliated treatments. Defoliation had no e ffect (P > 0.457) on maize biomass yield for either legume. Undefoliated mucuna yielde d (P = 0.093) more maize biomass than the natural fallow control. At 8 WAP, defoliati on of legumes resulted in lower (P = 0.015) maize biomass yield. All legume residue tr eatments (UD-M, D-M, UD-L, and D-L) yielded more (P < 0.001, P < 0.001, P = 0.024, and P = 0.092, respectively) maize biomass than the natural fallow control at 8 WAP. At 10 WAP, lablab plots yielded higher (P= 0.076) maize biomass than mucuna plots, and defoliation of both legumes reduced (P = 0.046) yield. The yield unde r UD-L was greater (P < 0.001) than the natural fallow control. By 15 WAP, th e legume and defoliation effects and their interaction were not signifi cant, however, maize biomass yield was higher under UD-M (P = 0.073), D-M (P = 0.081), and UD-L (P = 0.099) than the natural fallow control. Mucuna plots tended (P = 0.110) to outyield those of lablab and there were no defoliation effects at 21 WAP, but UD-M (0.034) and D-M (P = 0.10) outy ielded the natural fallow control. No treatment effects were si gnificant for maize DM yield at 27 WAP. Nitrogen uptake by maize following mucuna treatments was greater (P < 0.001) than lablab at 4 WAP, but defolia tion effects were not significan t (Table 3-7). At 8 WAP, legume defoliation reduced (P = 0.016) N uptake by maize, however, N uptake by

PAGE 83

68Table 3-6. Total above-ground maize bioma ss yield at different sampling dates. Weeks after planting Treatment† 4 8 10 15 21 27 -------------------------------------------------kg ha-1----------------------------------------------------UD-M 130 2030 3260 7180 15680 17310 D-M 110 1690 3060 7140 16990 16700 UD-L 100 1980 4100 7060 14600 17970 D-L 100 1610 3180 6450 15370 16800 N fertilizer 110 1600 2980 7550 16790 19800 Natural fallow 100 1400 2790 5830 13060 16310 Cattle manure 110 1750 3350 7890 15720 17840 Effects± -------------------------------------------------P values --------------------------------------------------Legume (L) 0.002 0.539 0.076 0.442 0.110 0.796 Defoliation (D) 0.530 0.011 0.046 0.533 0.842 0.550 L x D 0.037 0.849 0.159 0.587 0.484 0.851 SE 5.30 74.2 179.7 208.1 838.5 632.9 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated lablab, D-L = Defoliated lablab ± Significant contrasts of legume treatments vs . control treatments are reported in the text.

PAGE 84

69 maize in all legume residue treatments was higher (P < 0.02) than the natural fallow control. At 10 WAP, immediately after th e period when the N demand by maize is highest (Karlen et al., 1988) , the interaction of legume and defoliation effects was significant (P = 0.072). Defoliation of lablab reduced (P = 0.039) N uptake by maize, but did not affect (P = 0.478) N uptake under mucuna treatments. The N uptake was highest in the UD-L treatment and this was greater (P = 0.002) than UD-M. Relative to the natural fallow control, UD-L, and D-M resulted in higher (P = 0.001 and P = 0.036, respectively) N uptake. By 15 WAP, the effects of legume, defoliation, and their interaction were not signif icant, although analysis usi ng single degree of freedom contrasts revealed that the de foliation of lablab reduced (P = 0.085) N uptake by maize. Both UD-L (P = 0.028) and UD-M (P= 0.090) achi eved higher uptake than the natural fallow control. Treatment effects were not significant at 21 WAP and 27 WAP (at harvest), but UD-M achieved higher (P = 0.062) N uptake than the natural fallow control at 21 WAP. Defoliated lablab was associated with lo w N uptake and recovery rates at all time periods. This is likely due to two reasons. First, the defoliation of lablab at 10 cm considerably reduced the biomass left for in corporation of this upright growing legume (Table 3-1). Also, lablab biomass was compri sed of a higher proportion of stem than leaf (Table 3-1), and the stem had a higher lignin concentration than mucuna stem (Table 32), suggesting that D-L biomass was very lo w in quality. In contrast, defoliation of mucuna rarely had a significant effect on s ubsequent maize responses relative to UD-M.

PAGE 85

70Table 3-7. Nitrogen taken up by mai ze at different sampling dates. Weeks after planting Treatment† 4 8 10 15 21 27 ----------------------------------------------kg ha-1---------------------------------------------------UD-M 5 47 52 73 190 163 D-M 4 40 63 64 162 169 UD-L 3 44 82 81 132 171 D-L 4 38 57 60 111 151 N fertilizer 4 36 63 76 126 172 Weed fallow 4 32 40 49 113 143 Cattle manure 4 41 55 68 134 159 Effects± -------------------------------------------------P values ----------------------------------------------Legume (L) < 0.001 0.331 0.208 0.852 0.154 0.778 Defoliation (D) 0.608 0.016 0.432 0.211 0.496 0.701 L x D 0.088 0.622 0.072 0.572 0.915 0.499 SE 0.23 1.55 4.84 4.68 15.9 8.84 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated lablab, D-L = Defoliated lablab ± Significant contrasts of legume treatments vs . control treatments are reported in the text.

PAGE 86

71 Relationships Between Residue Qua lity Parameters and Soil Mineral N There were weak, but significant, negative correlations between soil mineral N and lignin (r = -0.61, P = 0.04), lignin-to-N ra tio (r = -0.66, P = 0.02) and (polyphenol + lignin)-to-N ratio (r = -0. 54, P = 0.07), and a positive correlation with residue N concentration (r = 0.6, P = 0.05) at 4 WAP (Fi gures 3-2 and 3-3). The significance of these relationships only at 4 WAP, and not in other time periods, could in part be explained by the fact that at 4 WAP N uptak e by maize was low and thus the soil mineral N reflected more closely the amount of N mine ralized from the resi due. In agreement with results of Meentemeyer ( 1978), these data indicate that lignin is an important quality component determining the rate of decompositi on and in part agrees with the observation by Melillo et al. (1989) that the lignin-to N ratio predicts the early stage of decomposition. The negative relationship is in agreement with previous studies that have demonstrated an inverse relationship between N release and lignin, lignin-to-N ratio, and (lignin + polyphenol)-to-N ratio (Palm and Sanchez, 1991; Oglesby and Fowness, 1992; Kachaka et al., 1993; Handayanto et al., 1994; Constantinides and Fowness, 1994; Vanlauwe et al., 1996; Ibewir o et al., 2000a) under laboratory incubations or litter bag decomposition studies. The positive relati onship between N concentration and soil mineral N at 4WAP was in agreement with previous report of Constantinides and Fowness (1994) in which the initial N explai ned most of the variation in N accumulation or depletion from the soil when fresh leaves or litter from commonly used agroforestry species were incorporated.

PAGE 87

72 y = -0.5632x + 15.226 R2 = 0.4374, P < 0.02 y = -0.5338x + 15.697 R2 = 0.2921, P < 0.07 y = -0.6871x + 17.807 R2 = 0.368, P < 0.04 6.0 7.0 8.0 9.0 10.0 11.0 12.0 13.0 14.0 2.03.04.05.06.07.08.09.010.011.012.0 Lignin, lignin-to-N and (ligni n + polyphenol)-to-N ratiosSoil mineral N (kg ha-1) Lignin to N (Lignin + polyphenol) to N Lignin Figure 3-2. Relationships between soil mi neral N and residue lignin, lignin-to-N ra tio, and (lignin + pol yphenol)-to-N ratio a t 4 WAP.

PAGE 88

73 Figure 3-3. Relationship between residue N concentration and soil mineral N at 4 WAP. y = 6.666x + 0.5 R2 = 0.6, P < 0.05 2 4 6 8 10 12 14 16 0.91.11.31.51.71.92.1 Residue N ConcentrationSoil Mineral N (kg ha-1)

PAGE 89

74 The material used in this study had a lignin: N ratio within a narrow range of 3.7 to 9.9. This limited range and the wide variability of mineral N in the soil under field conditions explains why we found only a weak correlati on between lignin:N and N release in the soil. Our results agree with those of Palm and Sanchez (1991) who used fresh legume leaves with a narrow range, but contrasts those of Constantinides and Fowness (1994) who used a wide range of materials. The low correlations could also be explained by losses of inorganic N in the soil due to l eaching (Cahn et al., 1993) and denitrification (Lensi et al., 1992) of nitrate under th e high rainfall of the subhumid tropics, by volatilization (Costa et al., 1990) , and to some extent from the uptake of N by maize. This relationship provides a practically useful function fo r estimating N release from residue quality under northwes tern Kenya conditions and co uld be used in modeling studies which evaluate the effects of organi c inputs on long-term so il fertility changes. Apparent N Recovery No treatment effects were detected on N recovery by maize (Table 3-8). Nitrogen recovery was lowest at 4 WAP (< 3%). At 8 WAP, when the demand for N by maize was greater, the recovery ranged from 12% fo r the inorganic N control to 27% for the UD-L treatment. The mean N recovery rates acr oss all time periods were in the range of 21% for cattle manure to 52% for the UD-L treatment (Table 8), but no significant differences were detected. The N recovery by D-M averaged 45% compared to 35% for the UD-M. Nitrogen recovery achieved by UD-L was similar to that from the inorganic N and was about twice that obt ained by D-L. The cattle manur e treatment had the lowest, but most consistent N recovery rate. The apparent N recovery by subsequent maize is similar to that reported by Tian et al. (2000) for legume cover crops in a derived savanna of West Africa.

PAGE 90

75 The high N accumulation in the natural weed fallow may have reduced the responses to legume biomass incorporation. Additiona lly across all treatmen ts, the weeding of maize during the period of maize growth and incorporating weeds could have contributed N to the system that was not quantified in our study. The relatively high N recovery found in our study could, in part be explained by unaccounted N from the fine roots. Roots slough off and litter decomposes during the dry season and this material was not accounted for in the DM quantified just before land preparation, but it did contribu te to the soil inorganic N ta ken up by the maize. Similar observations were made by Van Noordwijk and Purnomosidi (1992) who found a higher litter DM biomass during the gr owth period than live biomass measured at the end of the growth period. Also Hairiah and van Noor dwijk (1989), referenced in van Noordwijk and Purnomosidi (1992), found that N upt ake by maize following mucuna was 147 kg ha-1 higher than control crop while the N conten t of live biomass incorporated into the soil was only 71 kg ha-1. For lablab residue treatments, some seeds dehi sced and were incorporated into the soil but were not accounted for in the biomass record ed. Thus, these factors in addition to the time that elapsed between sampling/incorpor ating green manure residues and the planting of maize (21 d) as well as differences in ra infall amounts received prior to each sampling date, could partly explain the high variability observed in the N recovery data. On average, the N recovery was comparable to that obtained from inorganic fertilizer application (30-50%) reported from tropical cropping systems (Baligar and Bennet, 1986).

PAGE 91

76Table 3-8. Nitrogen recovery by ma ize at various sampling dates. Weeks after planting‡ Treatment† 4 8 10 15 21 27 Mean ---------------------------------------------------------%------------------------------------------------------------UD-M 2 26 21 41 84 34 35 D-M 2 24 73 54 51 66 45 UD-L -1 27 111 72 27 77 52 D-L 2 30 76 56 -12 16 28 N Fertilizer 1 12 60 92 45 98 51 Cattle manure 1 13 22 29 33 24 21 Effects ---------------------------------------------------------P Values ----------------------------------------------------Legume (L) 0.278 0.667 0.275 0.621 0.125 0.955 0.657 Defoliation (D) 0.199 0.981 0.848 0.960 0.668 0.813 0.759 L x D 0.405 0.742 0.305 0.661 0.879 0.456 0.355 SE 1.06 3.09 18.8 12.3 31.1 29.4 7.93 † UD-M = Undefoliated mucuna, D-M = Defoliated mucuna , UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 92

77 Bean and Lablab Grain Yield The effects of legume, defoliation regime, and their interactio n on bean grain and straw yields were not signifi cant (Table 3-9), however mucu na plots tended (P = 0.135) to yield greater bean grain yi eld than lablab. Yield of th e second crop of beans was low (102 kg ha-1) compared to both the first crop (648 kg ha-1) and the average grain yield of relay-cropped lablab (364 kg ha-1). The bean straw yield was not affected by treatment. Bean yields reported from the highlands of east Africa are generally between 0.2 to 1 t ha-1 without fertilization (Giller et al., 1998; Wortmann et al ., 2000), but yields above 1 t ha-1 have been reported for some varieties (W ortmann et al., 1996). Bean yields in our study were at the lower end of this range, proba bly due to low soil fer tility (Giller et al., 1998) and the occurrence of bean root rot, a major disease problem in western Kenya (Otsyula et al., 1999). The lower bean grain yield following lablab compared to mucuna could be due to greater competition from ma ize (under UD-L) and the susceptibility of lablab to leaf rust disease, a disease that c ould have also affected bean growth for this treatment. The lablab grain yield was cons istent with those (200-450 kg ha-1) reported in the literature (Weber, 1996) and was greater than the second bean crop grain yield, suggesting that the opportunity cost of foregoi ng the second crop of b eans to plant lablab is low. In addition there is an added adva ntage of producing biomass that could be used either as supplemental feed or for improving so il fertility. Thus re lay cropping of legume green manures has the advantage of growi ng the first crop of beans plus producing additional organic material instead of a sec ond bean crop, the yield of which is usually low.

PAGE 93

78 Table 3-9. Maize and bean grain and stover/straw yield for the 2000 growing season. Beans Maize Treatment† Grain Straw Grain Stover ------kg ha-1 ---------------t ha-1 --------UD-M 684 541 5.1 5.8 D-M 620 546 6.2 5.3 UD-L 546 498 7.4 6.7 D-L 481 455 5.7 4.3 Controls N Fertilizer 716 602 7.8 7.1 Natural fallow 715 621 4.8 4.6 Cattle manure 772 620 6.8 8.5 Effects ---------------------------------P Values ----------------------------Legume (L) 0.135 0.415 0.424 0.858 Defoliation (D) 0.456 0.812 0.753 0.334 L x D 0.998 0.769 0.215 0.587 SE 40.7 35.4 0.44 0.54 † UD-M = Undefoliated mucuna, D-M = De foliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab.

PAGE 94

79 Maize Grain Yield Despite relatively large numerical differences among legume treatment means, no differences in maize grain yield were dete cted (Table 3-9). For the green manure treatments, UD-L resulted in the maize grain yi elds closest to those of the N fertilizer control. Maize grain yield under UD-L was hi gher (P = 0.089) compared to the natural weed fallow where no inorganic N was applied. Yields following D-M were at least as great as for UD-M, whereas the trend was the opposite for lablab. The inorganic N control outyielded (P = 0.053) the natural fallo w, and cattle manure tended (P = 0.180) to yield higher than the natural fallow. Maize grain yield from the natural weed fallow in our study at NARC-Kitale, Kenya (Chapter 4) was higher than thos e reported from farmersÂ’ fields (Chapter 5). This is not unexpected because similar yields (average of 2 yr of 4.4 t ha-1) have been reported from the highlands of eastern Afri ca when no external inputs we re applied (Cheminingwa and Nyabundi, 1994). This could be attributed to th e higher fertility in the experiment station and the N contribution from the natural w eed biomass (Nguimbo and Balasubramanian, 1992; Carsky et al., 1999; Muhr et al., 1999c). The modest trend toward greater maize yield following incorporation of lablab resi due compared to mucuna concurs with the results of Wortmann et al. (2000) and is in agreement with our observation of a trend toward enhanced N uptake following lablab re sidue, which was of lower quality (lower N and higher lignin concentr ation) than mucuna. The high N-uptake by maize under the D-M treat ment was reflected in the N recovery and maize grain yield obtained from this treatment compared. Our hypothesis was that utilization of the high quality upper canopy of green manure legumes as livestock feed

PAGE 95

80 and soil incorporating the remaining intermed iate quality stubble and root will enhance synchrony between N release and uptake by succeeding maize. The reduction of the percentage DM contribution of leaves in th e D-M and the higher lignin concentration of mucuna roots and stems than leaves result ed in a reduction of biomass quality after defoliation. Thus the trend toward increased efficiency of N uptake and grain yield by maize under this treatment may be explained by this reduction in qual ity (Myers et al., 1994). This observation is in agreement with literature on the effect of residue quality on N release and uptake by the reci pient crop. Residue of intermediate quality has been shown to increase the efficiency of N uptake compared to very high or low quality residues (Handayanto et al., 1997). Biomass of very high quality (high N and low lignin/polyphenol concentration) releases N faster than it can be taken up by the recipient crop while very low quality biomass (low N and high lignin/polyphenol concentration) releases N too slowly to meet the demand by the recipient crop (Myers et al., 1994). Conclusions Under the conditions of northwestern Ke nya, undefoliated relay-cropped mucuna and lablab produced 2 t ha-1 of dry matter (Table 3-1) , contributing 38 and 36 kg N ha-1, respectively (Table 3-4). Green manure treat ments resulted in average maize grain yield advantages of 0.85 and 1.75 t ha-1, respectively, compared to the natural weed fallow (Table 3-9). Assuming that approximately 20% of the N from high quality residue is recovered by the first crop (G iller and Cadisch, 1995), the yield advantages are much higher than estimates of 500 kg ha-1 from similar biomass yields (Giller et al., 1997). There is also an added long-term advantag e of reducing the soil organic matter decline (Kapkiyai et al., 1999). Given that maize and bean production are not sacrificed to grow mucuna and lablab biomass, planting the interc rop and defoliating part of the biomass as

PAGE 96

81 livestock feed could be an attractive opti on for farmers. Therefore even though green manures grown as relay intercrops have lower N contribution potential than if grown as a sole crop, these systems are mo re likely to be adopted than sole-crop plantings because the green manure does not prohibit the main crop of beans and may also increase both livestock productivity and soil fertility over time. Even a low maize grain response may be of interest to farmers who have limited access to commercial fertilizers. These farm ers are more likely to adopt a legume plant that can produce some feed for their livesto ck or food for the household. Lablab may be preferred to mucuna because it can provide grai n before it is incorpor ated or part of its biomass can be used as livestock feed. Ba sed on our results of N uptake and maize and bean yield response, even after defoliation, it can be concluded th at both mucuna and lablab could be included in the design and development of sustainable green manure intercropping systems in the sub-humid highlands of eastern Africa, including those in northwestern Kenya.

PAGE 97

82 CHAPTER 4 PRODUCTIVITY OF MAIZE-BEAN INTERCROP RELAY CROPPED WITH MUCUNA AND LABLAB GREEN MANURES Introduction Declining soil fertility as a result of nutri ent depletion is a major constraint to crop production in many areas of the tropics, and in Africa it is recognized as the fundamental biophysical cause of declining per capita food production (Sanchez et al., 1996). Most agricultural production in western Kenya is fr om smallholder farmers who practice mixed crop-livestock farming where maize ( Zea mays L.) and dairy are the major enterprises. On average, these farmers cultivate less th an 2 ha, and 74% of them do not use any inorganic fertilizer. Those who use fert ilizer, apply it at rates far below those recommended for optimum crop production (M uriuki, 1998). Participatory rural appraisals (PRAs) carried out in the NARC-Kitale regional research mandate area indicated that crop yields and livestock produc tivity were low. This was attributed mainly to continuous cultivation of crops or grazing of livestock w ithout the addition of adequate external nutrient i nputs, leading to soil nutrient de pletion (Smaling et al., 1997). The use of legumes could play an impor tant role in improving productivity and sustainability of these smallholder production systems. Increased interest in the management of nutrients and organic matter in relation to sustainability of agricultural systems has resulte d in consistent research effort on the use of green manures. Herbaceous legume cover cr ops can provide an a lternative to the use of inorganic fertilizers for crops (Tian et al., 2000) a nd use of commercial feed

PAGE 98

83 supplements for livestock. Legume residues pr ovide organic matter to the soil (Hulugalle et al., 1996; Tian et al., 1999), nutrients, especially N (Carsky et al., 1999; Ibewiro et al,. 2000; Tian et al., 2000), and improvement in soil physical properties (Lal et al., 1995; Tian et al., 1999). Adoption of legume gr een manures on smallholder farms however, depends not only on their contribution to so il fertility improvement, but also their potential for other uses such as weed suppr ession, human food, and feed for livestock (Becker et al., 1995; Versteeg et al., 1998a). The mulch layer formed by legume residues suppresses weed growth (Weber, 1996; A kobundu et al., 2000), controls pests and diseases (Weber, 1996), and reduces labor required for the subsequent seasonÂ’s crop (Akobundu et al., 2000). A number of legumes such as mucuna [ Mucuna pruriens (L.) DC. var. Utilis (Wright) Bruck], lablab [ Lablab purpureus (L.) Sweet cv. Rongai], tropical kudzu ( Pueraria phaseoloides ), and centro ( Centrosema pubescens ) have shown promise in the tropics (Ibewiro et al., 2000a; Tian et al., 2000; Wortmann et al., 2000). Annual dry matter yields of these legumes range from 2 to 8 t ha-1 and N accumulation from 30 to 300 kg ha-1 (Hairiah and Van Noordwijk, 1989; Ibewiro et al., 2000a; Tian et al., 2000). Despite the availabi lity of data on the yield a nd contribution of above-ground legume biomass to succeeding cereal crops, information is still limited on the contribution of roots and stubbl e biomass when part of th e residue is harvested for fodder. Since 1995, the concept of green manure/ cover crops has been introduced to northwestern Kenya by the KARI/RF on-farm project whose broad objective was to develop organic manure-based, low cost tec hnologies for improved soil management on smallholder farms. Screening work conducted for 3 yr indicated that mucuna and lablab

PAGE 99

84 were the most promising green manure legumes that could be relay cropped into the maize-bean cropping system. The current study was part of an effort to integrate process research with participatory on-farm research to alleviate the nutrie nt depletion problem on smallholder farms. The objectives were to: 1) quantify the e ffects of relay-cropped legumes on yield of a subsequent maize-bean intercrop when part of the legume residue was harvested for fodder or all of it remained on site and was soil incorporated before maize planting, and 2) to determine the effect of number of years of residue use on soil fertility as measured through a subs equent test maize-bean intercrop. Materials and Methods Experimental Site The research was conducted from 1999 to 2001 at the National Ag ricultural Research Center (NARC) at Kitale in northwestern Kenya (10 01Â’N and 350 00Â’E; 1860 m). The center is in agro-ecological zone Upper Midl ands 4, as described by Jaetzold and Schmidt (1983). The experimental site was a field th at has been under continuous cultivation of maize for at least the last 10 yr. The soils are classified as humic Ferrolsols based on FAO/UNESCO system (FAO UNESC, 1994) equivalent to Kandiudalfic Eutaudox in the USDA soil Taxonomy system (Soil Survey Staff, 1994). These are deep, highly weathered and well-drained, clay loam soils that are dark re d to dark reddish brown in color with low activity clay and moderately acid topsoil (Jaetzold and Schmidt, 1983). Both N and P are limiting for crop growth (S maling et al., 1997). The top soil (0-20 cm) had the following properties; pH (1:2.5 H2O), 5.4; organic C, 14.2 g kg-1; total N, 1.3 g kg-1; extractable P (modified Olsen), 9.7 mg kg-1; and is clay loam with 39% clay, 41% sand, and 20% silt. Rainfall is distributed in one growing season with an annual total (30-yr average) of 1143 mm. The growing seas on is from mid-March to mid-November.

PAGE 100

85 The rainfall is relatively evenly distributed during April through N ovember with peaks in May and August. The dry seas on is from December to March. Experimental Treatments and Layout There were 15 treatments, including three controls, replicated three times in a randomized complete block design. Twelve treatments originated from a 2 x 3 x 2 factorial that included two legume cropping sy stems, three crop sequences (number of years of residue application), and two legum e defoliation treatments (Table 4-1). The two legume cropping systems were 1) maize + bean (both planted in April) + mucuna (planted in August), and 2) maize + bean (bot h planted in April) + lablab (planted in August). The three crop sequences were defined based on whet her the legume was planted in the first year onl y, in the second year only, or in both the first and second years. Plots planted to mu cuna or lablab only in the first year (August 1999) were planted to the maize-common bean intercrop in Year 2 (2000). Plots planted to mucuna or lablab in Year 2 (August 2000) only were planted to maize-common bean intercrop in Year 1 (1999). The two legume defoliation treatments were 1) herbage above 10 cm removed at season end, and 2) undefoliated. The three control treatments were 1) inorganic N fertilizer (30 kg N ha-1), 2) no N fertilization, and 3) 5 t ha-1 cattle manure (supplying approximately 65 kg N and 18 kg P ha-1). No inorganic N was applied to any plots other than the inorganic cont rol. At maize planting, 13 kg P ha-1 was applied in the same hill as for maize to all plots, except th e cattle manure treatment. In the third year of the experiment, all plots were pl anted to a maize-bean intercrop. Experimental plots were 4.5 by 6 m with a 1-m border on a ll sides. The experiment was planted at the beginning of the rainy s eason in April 1999. The maize was planted at an interand intra-row spacing of 75 by 30 cm , respectively, using two seeds per hill of

PAGE 101

86 hybrid 614D maize. The maize seedlings were thinned to one plant per hill 4 wk after planting (WAP) to give a plant population of 44,440 plants ha-1. An improved bean cultivar (GLP2; Rose coco), commonly plan ted by farmers in the region, was used. The first crop of common bean was planted simulta neously with maize at onset of rains in April. The second crop of beans or green le gume manures was relay cropped in maize in August, 135 d after the April planting. Befo re planting the second bean crop or the legume intercrop (after harvesting the first cr op of beans), the maize was weeded and all leaves below the ear were removed to mi nimize shading. The common bean, mucuna, and lablab were planted between the maize ro ws at an intra-row spacing of 30 cm using three seeds per hill, which were thinned to two plants per hill 4 WAP. All the plots were hand weeded twice before harvesting the firs t crop of beans. The plots were manually weeded once after the August plan ting of legumes. Stalk borer ( Chilo spp.), a common pest of maize, was controlled by application of Beta-cyfluthrin (Bulldock), a synthetic pyrethroid insecticide in granular form into th e whorl of each plant, at the rate of 7 kg ha-1 at 4 wk and 8 wk after germination. Herbage Yield and Chemical Composition After harvesting maize, maize stover was re moved in accordance with farmer practice, and mucuna and lablab were left to continue growing. For plots that were defoliated, the legumes were cut to a stubble of 10 cm in March before land preparation. The production of legume residue biomass was assessed in terms of litter fall, above-ground biomass (leaf and stem), a nd root mass from two 0.5-m 2 quadrats per plot at 30 WAP (mid-March). To measure root mass, soil from the 0.5-m2 areas was removed to a depth of 20 cm. All visible roots were separated fr om the soil, washed with water on top of a 0.5-mm sieve to remove remaining soil, and rins ed with distilled water. The samples of

PAGE 102

87 shoots and roots were oven dried at 600C for 48 h, weighed, and ground for determination of N concentration. Nitrogen concentration of the plant samples was determined using the Kjeldahl digestion with concentrated sulphuric acid followed by colorimetric determination of N (Anderson and Ingram, 1993) . Nitrogen accumulated in the harvested biomass and litter fall was calculated by multiplying biomass N concentration and quantity. Grain and Stover/Straw Dry Matte r Yield of Maize and Beans The first crop of common bean each year was harvested at the end of July and the second crop of common bean a nd the maize were harvested in mid-November. Maize and bean were harvested from the two center rows of every plot. The first and last two hills of each row were not harvested. The size of the sampling unit was 1.5 by 5.4-m (8.1 m2). Maize cobs were separated manually from the stover and threshed to separate grain. Maize grain yield at 13.5% mois ture concentration and stover DM yield were recorded. Common bean grain and straw DM yields were also measured from the same sampling unit. Statistical Analyses To assess the effect of legume croppi ng system, defoliation regime, and cropping sequence (number of years of residue application), and th eir interactions on biomass yield, chemical composition, and nutrien t accumulation, the above and below-ground legume fraction data were analyzed using th e general linear model (GLM) procedure of SAS (SAS, 2001). The maize and bean data were also analy zed using the same procedure. Single degree of freedom contrasts were used to compare controls with green manure treatments. Treatments were considered different at P < 0.10.

PAGE 103

88 Table 4-1. Outline of treatment arrange ment showing crop combinations, cropping system sequences, and legume defoliation regime. 1999 2000 2001 Cropping System† Legume Defoliation‡ Cropping System Legume Defoliation Z/B/M No Z/B/M No Z/B Z/B/M Yes Z/B/M Yes Z/B Z/B/B Z/B/M No Z/B Z/B/B Z/B/M Yes Z/B Z/B/M No Z/B/B Z/B Z/B/M Yes Z/B/B Z/B Z/B/L No Z/B/L No Z/B Z/B/L Yes Z/B/L Yes Z/B Z/B/B Z/B/L No Z/B Z/B/B Z/B/L Yes Z/B Z/B/L No Z/B/B Z/B Z/B/L Yes Z/B/B Z/B Z/B/B (IN)§ Z/B/B (IN) Z/B Z/B/B (No IN)§ Z/B/B (No IN) Z/B Z/B/B (CM)§ Z/B/B (CM) Z/B † Crops are represented as Z = maize, B = common bean, M = mucuna, and L = lablab ‡ Defoliation regime indicated as no = no defoliation occurring and yes = mucuna or lablab herbage above 10 cm removed at seas on end as fodder, and a dash (-) indicates absence of forage legume § Control treatments abbreviated as maize and beans IN = Inorganic N fertilized, No IN = No inorganic N applied, and CM = cattle manure applied

PAGE 104

89 Results and Discussion The results and discussion are organized to address three resear ch questions: 1) was there an effect of one year of residue application?, 2) was there a lesser effect of 1 yr versus 2 yr of residue applica tion?, and 3) were there long-te rm residual effects of residue application? In evaluating the effect of one year of residue app lication, legume, common bean, and maize data were used from plots where legume residue was incorporated in the first year only (March 2000) or in the s econd year only (March 2001). The comparison of 1 yr versus two consecutive years of residue application was made using 2001 maize and bean data from plots where legume residue was incorporated only in the second year (March 2001) and plots where legume residue was incorporated in both years (March 2000 and 2001). The long-term residual effects of residue application on common bean and maize yields were evaluated using 2001 maize and bean data from plots where legume residue was incorporated in March 2000, but not 2001. Effect of One Year of Residue Application Legume Biomass Yield Years of legume biomass data are those wh en the residue was in corporated (e.g. 2000 data are for legume planted in August 1999 and incorporated in March 2000). There was a significant interaction (P = 0.005) between le gume and year for le gume leaf, litter, and total residue (Table 4-2). Thus data were analyzed and presented by year. During the 2000 season, there was a trend (P = 0.157) towa rd greater legume le af mass for mucuna, but in the second year mucuna did produce higher (P < 0.001) leaf mass than lablab. Defoliation resulted in leaf mass re duction (P < 0.001) in both seasons. Type of legume had no effect (P = 0.598), but defoliation reduced (P < 0.001) stem mass. Stem biomass was higher (P = 0.008) during the 2001 season than 2000.

PAGE 105

90 Interaction of legume and year for litter bi omass occurred because the legume effect approached significance (P = 0.108) during the 2000 season, but mucuna produced higher (P = 0.013) litter mass than lablab during th e 2001 season. Defoliation had no effect (P = 0.858) on litter mass. Neither legume, defoliati on, nor their interaction had an effect on root mass, however, root mass during the 2001 season was greater (P = 0.024) than during 2000. The legume by year interacti on for total residue biomass was significant (P = 0.004). There was no difference (P = 0.944) between mucuna and lablab in residue mass during the 2000 season, but mucuna produced grea ter (P = 0.001) residue mass during 2001. Defoliation reduced (P < 0.001) total residue mass of both legumes in both seasons. Nitrogen Concentration Both the legume by defoliation and legum e by year interactions for leaf N concentration were significant (P = 0 .028, P = 0.031, respectively) (Table 4-3). Defoliation of lablab reduced (P = 0.066) leaf N concentration, but it had no effect (P = 0.821) on mucuna. During the 2001 season, lablab tended (P = 0.106) to contain higher leaf N concentration than mucuna, but there was no difference (P = 0.274) between the legumes during the 2000 season. The interaction of legume by defoliation for st em N concentration was significant (P = 0.005). Defoliation of lablab reduced (P = 0.001) stem N co ncentration, but it did not affect (P = 0.718) mucuna. Both UD-M a nd D-M stems contained higher (P = 0.061, P < 0.001, respectively) N concentration than UDL and DL, respectively. Stem N concentration was higher (P = 0.058) duri ng the 2001 season than the 2000 season. Lablab litter fraction contained higher (P = 0.004) N concentration than mucuna.

PAGE 106

91 Table 4-2. Effects of legume and defoliati on on residue biomass of mucuna and lablab relay cropped in maize for 1 yr. Leaf Stem Litter Roots Total Legume Residue Treatments -----------------------------t ha-1 ---------------------------2000 season UD-M† 1.070.750.120.11 2.05 D-M 0.220.270.130.07 0.69 UD-L 0.780.910.170.15 2.00 D-L 0.030.310.190.17 0.70 2001 season UD-M 1.691.521.530.27 5.01 D-M 0.730.881.330.27 3.21 UD-L 0.761.740.170.31 2.89 D-L 0.130.680.180.25 1.15 Effects -------------------------P values -------------------------Legume (L) < 0.0010.5980.0300.257 0.003 Defoliation regime (D) < 0.001< 0.0010.8580.528 < 0.001 L x D 0.1940.2090.8040.916 0.928 Year (Y) 0.034 0.0080.0210.024 0.003 L x Y 0.0050.6660.0200.404 0.004 D x Y 0.9660.1490.8100.751 0.467 L x D x Y 0.4850.4740.8280.404 0.999 SE 0.11 0.11 0.15 0.02 0.31 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 107

92 Table 4-3. Residue N concentration of mu cuna and lablab relay cropped in maize. Leaf Stem Litter Roots Total Legume Residue Treatments --------------------------g kg-1 --------------------------2000 season UD-M† 22.1 14.5 12.1 13.4 18.9 D-M 25.1 16.2 13.2 13.3 18.2 UD-L 24.6 13.8 13.7 11.6 17.5 D-L 16.7 10.2 15.4 11.3 12.3 2001 season UD-M 36.8 18.9 8.0 17.5 21.8 D-M 35.7 17.4 8.1 18.4 18.9 UD-L 43.2 15.0 13.0 13.7 21.5 D-L 39.1 11.0 15.4 12.4 15.3 Effects ----------------------P values -------------------------Legume (L) 0.624 < 0.001 0.004 < 0.001 0.104 Defoliation regime (D) 0.082 0.006 0.215 0.682 0.031 L x D 0.028 0.005 0.463 0.256 0.213 Year (Y) 0.020 0.058 0.197 0.095 0.205 L x Y 0.031 0.114 0.087 0.010 0.551 D x Y 0.754 0.143 0.947 0.961 0.581 L x D x Y 0.299 0.218 0.681 0.373 0.910 SE 2.10 0.66 0.74 0.57 0.87 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 108

93 The interaction of legume by year for root N concentration was si gnificant (P = 0.010) because mucuna N was affected by year to a gr eater extent than lablab. Mucuna root N concentration was greater (P < 0.10) than labl ab during both years. The N concentration of total residue tended to be greater for mu cuna than lablab (P = 0.104) and defoliation reduced (P = 0.031) N concentration of total re sidue. In summary, mucuna stem and root biomass generally had greater N concentrati ons than for lablab, while the reverse was true for litter. Trends toward greater N con centration in mucuna vs. lablab total residue and for undefoliated vs. defoliated plots were mainly a result of gr eater leaf mass for mucuna and for undefoliated residue and the hi gh N concentration of the leaf fraction. Nitrogen Content There was a significant interaction of le gume by year (P = 0.003) and defoliation by year (P = 0.059) for N content in the leaf fraction (Table 4-4). Mucuna accumulated greater (P < 0.001) N in the leaf than lablab during the 2001 season, but only tended (P = 0.183) to have greater N during the 2000 seas on. Interaction of defoliation and year occurred because the magnitude of the defolia tion effect was greater in 2001 than 2000. Nitrogen content in the stem fraction was a ffected by both the inte ractions of legume by year (P = 0.059) and defoliation by year (P = 0.031). Nitrogen content in mucuna stem was higher (P = 0.060) than lablab dur ing the 2001 season, but was not different (P = 0.665) during 2000. Defoliation reduced (P < 0 .004) N content in the stem fraction in both years. The interaction of legume by year for N yield in the litter was significant (P = 0.014). Lablab accumulated more (P = 0.067) N in the litter during 2000 due to the higher N concentration in this fraction but mucuna accumulated more N in the litter during the 2001 season, probably due to higher litter bioma ss as a result of a later sampling date (2 wk later). Root N c ontent was greater (P = 0.011) during the 2001

PAGE 109

94 season than 2000, but it was not affected by trea tment. The interaction between legume and year for total residue N yield was si gnificant (P = 0.001). Nitrogen content of mucuna total residue was greater (P < 0.001) than lablab during the 2001 season, but there was no difference (P = 0.553) in 2000. Defoliation reduced (P < 0.001) the residue N content of both legumes in both seasons. Bean Grain and Straw Yield There was no year by treatment interacti on, therefore the average bean and straw yields across the 2 yr are presented (Table 4-5). Bean grain yiel ds were higher (P = 0.015) after mucuna than lablab, but legume defoliation had no effect (P = 0.571). The year effect was significant (P = 0.014) for bean grain yield. The mean bean grain yield under legume residue treatments dur ing the 2000 growing season (583 kg ha-1) was higher than for the 2001 season (437). The y ear effect was also significant (P = 0.002) for bean straw yield. The mean straw yiel d under legume residue treatments during the 2000 season (510 kg ha-1) was higher than during the 2001 season (330). No differences (P > 0.10) were detected between grain and st raw yields of residue treatments compared to the controls. Maize Grain and Stover Yield The year by treatment interactions were not significant for both ma ize grain and stover yield (Table 4-6), therefore means across the two seasons are presented. The effect of legume, defoliation, and their interaction on both grain and stove r yields were not significant, but there was a year effect. The mean maize grain yields under residue treatments were higher during the 2000 (6.1 t ha-1) than 2001 season (5.1), but the reverse

PAGE 110

95 Table 4-4. Residue N content of mucuna and lablab relay cropped in maize. Leaf Stem Litter Roots Total Legume Residue Treatments -----------------------------kg ha-1 --------------------------2000 season UD-M† 26.18.81.51.5 37.9 D-M 5.64.41.71.0 12.7 UD-L 19.012.42.31.8 35.5 D-L 0.53.22.91.8 8.4 2001 season UD-M 61.628.712.24.8 107.3 D-M 26.215.311.14.9 57.5 UD-L 31.125.21.54.3 62.0 D-L 5.97.42.23.1 18.7 Effects -------------------------P values ------------------------Legume (L) < 0.0010.0670.0360.514 < 0.001 Defoliation regime (D) < 0.001 0.0010.9590.457 < 0.001 L x D 0.1730.4020.7600.710 0.415 Year (Y) 0.015 0.0110.028 0.011 0.030 L x Y 0.0030.0590.0140.096 0.001 D x Y 0.0590.0310.8740.782 0.096 L x D x Y 0.5840.9190.8390.348 0.307 SE 4.40 2.03 1.12 0.36 7.08 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 111

96 Table 4-5. Grain and straw yi eld of common bean intercropp ed in maize after one year of mucuna and lablab residue incorporati on. Data are means across two seasons (2000 and 2001). Treatments GrainStraw ---------------kg ha-1 --------------UD-M† 610460 D-M 560450 UD-L 440400 D-L 430380 Controls Inorganic N 605470 Natural fallow 547465 Cattle manure 658472 Effects ----------------P values -------------Legume (L) 0.0150.155 Defoliation regime (D) 0.5710.738 L x D 0.7550.885 Year (Y) 0.0140.002 L x Y 0.9010.978 D x Y 0.5150.937 L x D x Y 0.7520.714 SE 30.4127.30 † UD-M = Undefoliated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 112

97 Table 4-6. Grain and stover yield of maize af ter one year of mucuna and lablab residue application. Data are means across two seasons (2000 and 2001). Treatments GrainStover ---------------t ha-1 ----------------UD-M† 5.26.9 D-M 5.57.1 UD-L 6.57.5 D-L 5.26.3 Controls Inorganic N 7.59.4 Natural fallow 4.47.0 Cattle manure 6.69.2 Effects --------------P va lues --------------Legume (L) 0.4250.997 Defoliation regime (D) 0.4990.208 L x D 0.2010.486 Year (Y) 0.0270.006 L x Y 0.5680.783 D x Y 0.8580.916 L x D x Y 0.3760.901 SE 0.270.51 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 113

98 was true for stover yield. Single degree of freedom contrasts show ed that UD-L yielded more (P = 0.088) maize grain than UD-M, a nd defoliation of lablab reduced (P = 0.084) grain yield relative to UD-L. Defoliation of mucuna had no effect (P = 0.571) on maize grain yield. Average maize grain yield of the legume residue treatments was greater (P = 0.076) than the natural fallow control, and th ese differences were greater (P = 0.048) for undefoliated treatments versus the control. Defoliated mucuna tended (P = 0.166) to yield higher maize grain than the natural fall ow. Similarly, N fertilizer and cattle manure yielded higher (P < 0.01) maize grain than th e natural fallow control. The yields of undefoliated legume treatments were not different but tended to be lower than (P = 0.161) those obtained using cattle manure. Effect of One versus Two Years of Consecutive Residue Application Legume Biomass Yield The three-way interaction of legume by defoliation by cropping sequence was significant (P = 0.062) for leaf mass (Table 4.7 ). Undefoliated mucuna yielded higher leaf mass than lablab on one-y ear residue plots (P = 0.002) and tended to yield more (P = 0.105) on two-year residue plots. Defoliate d mucuna yielded higher (P < 0.001) leaf mass than D-L in both sequences. Defoliati on reduced (P < 0.01) leaf mass of both legumes in both oneand two-year residue plots. The interaction of legume, defoliation re gime, and cropping sequence on stem herbage yield was also significant (P = 0.018). Stem mass of UD-L was higher (P = 0.003) than UD-M on two-year residue plots but not (P = 0.551) on one-year resi due plots (Table 47). Stem mass of D-M was higher (P = 0.012) than D-L on two-year residue plots and tended to be higher (P = 0.140) on one-year residue plots. Defoliation of mucuna reduced (P = 0.003) stem mass on one-year resi due plots but not (P = 0.267) on two-year

PAGE 114

99 residue plots, however, defoliation of lablab reduced stem mass in both one(P = 0.036) and two(P < 0.001) year residue plots. Thus, unlike leaf mass, which was consistently greater for mucuna, stem mass tended to favor lablab for undefoliated plots and mucuna for defoliated plots. In agreement with the literature (Van N oordwijk and Purnomisidi, 1992) mucuna had greater amounts of litter biomass than lablab (P < 0.001; Table 4-7) . Root herbage mass was not affected by treatments, but lablab root comprised higher percentage of total biomass (10%) than mucuna (5%), in agreement with Tian and Kang (1998). The three-way interaction of legume by defoliation by cropping sequence for total residue mass was significant (P = 0.081). Undefoliated mucuna yielded higher (P = 0.058) total residue mass than UD-L on one-yea r plots and tended to yield more (P = 0.103) on two-year plots. Defoliated mucuna yielded higher residue mass than D-L on both one(P = 0.011) and two(P < 0.001) year residue plots. Defoliation of mucuna decreased (P = 0.073) residue mass on one-year residue plots, but it did not (P = 0.918) affect residue mass on two-year plots. De foliation of lablab decreased residue mass on both one(P = 0.033) and two(P = 0.004) year plots. Generally mucuna yielded more total resi due biomass than lablab, and defoliation reduced residue mass, except for D-M on two-y ear residue plots which yielded as much as UD-M. The generally higher residue biomass for mucuna vs. lablab plots was due to the greater leaf and litter biomass for mu cuna treatments. The greater amounts of mucuna biomass suggest that it was more ad apted to growth under shading by maize and to the dry conditions following maize harvest than lablab. The biomass yields under the

PAGE 115

100 Table 4-7. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue biomass of mucuna a nd lablab relay cropped in maize during the 2000/2001 season. Leaf Stem Litter Roots Total Legume Residue Treatments -----------------------------t ha-1 --------------------------1st-Yr legume UD-M† 1.691.521.530.27 5.01 D-M 0.730.881.330.27 3.21 UD-L 0.761.740.200.30 2.89 D-L 0.130.680.210.25 1.15 2nd-Yr legume UD-M 1.411.171.990.24 4.81 D-M 0.781.412.230.32 4.74 UD-L 0.961.940.060.34 3.31 D-L 0.120.570.150.22 0.96 Effects -------------------------P values ------------------------Legume (L) < 0.0010.919< 0.0010.898 < 0.001 Defoliation regime (D) < 0.001< 0.0010.8970.525 < 0.001 L x D 0.669< 0.0010.9610.131 0.095 Cropping sequence (CS) 0.9000.5520.3310.898 0.233 L x CS 0.1280.8500.1970.898 0.388 D x CS 0.6690.2270.6570.898 0.382 L x D x CS 0.0620.0180.7630.335 0.081 SE 0.11 0.11 0.15 0.02 0.33 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 116

101 intercrop were in the lower end of the range of biomass reported by Wortmann et al. (2000) from the highlands of east Africa when mucuna and lablab were grown in rotation with maize. The higher proportion of stem in lablab, and the higher lignin concentration in its stem fraction (Chapter s 3 and 7) suggests that lablab residue may be of lower quality than mucuna. Legume N Concentration Legume and defoliation regime effects on leaf N concentration were significant (Table 4-8). Undefoliated lablab c ontained higher (P < 0.004) leaf N concentration than UD-M. Defoliation reduced lablab leaf N concentra tion (P < 0.085), but it did not affect mucuna leaf N concentration (P > 0.552). Stem N concentration was affected by both legume (P < 0.001) and the interaction of defoliation regime by cropping sequence (P = 0.049). Mucuna contained higher N concentration in the stem than lablab (P < 0.001). Defoliation by sequence interaction effects occurred because defoliation tended to reduce stem N concentration on one-year residue plots (P = 0.118), but there was no eff ect (P = 0.870) on two-year plots. Mucuna litter contained lower (P < 0.003) N concentra tion than lablab, likely due to the greater initial N concentration in lablab leaf. The lo wer N concentration in the litter fraction than in the leaf may have been due to nutrient re sorption during senescence when leaf proteins and other nitrogenous compounds are hydrolyzed and the products are transported into perennial tissue before leaf fall (Norby a nd Contrufo, 1998). Nitrogen concentration in mucuna roots was higher (P < 0.001) than in lablab concurring with Tian and Kang (1998). The N concentration in the total legum e residue was not affected by legume (P = 0.823) and cropping sequence (P = 0.235), but defoliation reduced (P = 0.008) N concentration in the residue.

PAGE 117

102 Legume N Content The three-way interaction of legume by defoliation by cropping sequence for leaf N content was significant (P = 0.045) (Table 49). By virtue of greater leaf mass, undefoliated mucuna had greater (P = 0.003) N content in the leaf fraction than UD-L on one-year plots, but there was no difference (P = 0.628) on two-year plots. Defoliated mucuna had greater (P = 0.03) leaf N content than D-L on both oneand two-year plots. Defoliation reduced (P = 0.042) leaf N content of both legumes on both oneand twoyear plots. The N content of the leaf fracti on of D-M was on average 48% as great as that from UD-M, but the contribution from D-L was only 4% that of UD-L. Thus the impact of removing top canopy herbage for fodder on am ount of residue N incorporated is much greater for lablab than mucuna. The interaction of legume by defoliation by cropping sequence for stem N content was significant (P = 0.027). There was no differe nce (P > 0.127) in stem N content between UD-M and UD-L, but D-M had greater (P < 0. 009) stem N content than D-L on both oneand two-year plots. Defoliation of mucuna decreased (P = 0.004) stem N content on oneyear plots, but it did not affect (P = 0.349) stem N content on two-year plots. Defoliation of lablab decreased (P < 0.055) stem N conten t on both oneand two-ye ar plots. Mucuna accumulated more N in the litter than lablab (P = 0.001), but defoliation did not affect (P = 0.906) N content in the litter fraction. The three-way interaction of legume by defoliation by cropping sequence for total residue N content was significant (P = 0.054) . Undefoliated mucuna had higher (P = 0.033) N content in total legume residue than UD-L in one-year plots, but there were no differences (P = 0.317) on two-year plots. Defo liated mucuna had greater N content than

PAGE 118

103 Table 4-8. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue N concentration of mu cuna and lablab relay cropped in maize. Leaf Stem Litter Roots Total Legume Residue Treatments -------------------------g kg-1 -----------------------------1st-Yr legume UD-M† 36.818.98.017.5 21.8 D-M 35.717.48.118.4 18.9 UD-L 43.215.013.013.7 22.6 D-L 39.111.015.412.4 15.4 2nd-Yr legume UD-M 34.817.78.016.7 19.1 D-M 33.717.37.815.6 15.4 UD-L 45.413.015.113.2 22.3 D-L 36.912.812.611.2 13.4 Effects ---------------------------P values -------------------------Legume (L) 0.003< 0.0010.003< 0.001 0.823 Defoliation regime (D) 0.0360.0180.9640.104 0.008 L x D 0.1140.3280.9900.146 0.203 Cropping sequence (CS) 0.5200.5860.9040.021 0.235 L x CS 0.5600.6180.9830.304 0.335 D x CS 0.5770.0490.3990.191 0.717 L x D x CS 0.5570.2630.4410.578 0.894 SE 1.10 0.63 0.74 0.54 0.93 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 119

104 Table 4-9. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation on residue N content of muc una and lablab relay cropped in maize. Leaf Stem Litter Roots Total Legume Residue Treatments -----------------------------kg ha-1 --------------------------1st-Yr legume UD-M† 61.628.712.24.8 107.3 D-M 26.215.311.15.0 57.5 UD-L 31.125.21.54.3 62.0 D-L 5.97.42.23.1 18.7 2nd-Yr legume UD-M 49.220.716.23.9 89.9 D-M 26.224.417.45.0 73.0 UD-L 43.825.40.64.6 74.5 D-L 4.47.31.12.5 15.3 Effects -------------------------P values -------------------------Legume (L) < 0.0010.0050.0010.086 < 0.001 Defoliation regime (D) < 0.001< 0.0010.9060.403 < 0.001 L x D 0.3180.0040.9190.070 0.119 Cropping sequence (CS) 0.7630.8060.4750.630 0.806 L x CS 0.1090.9610.2900.867 0.696 D x CS 0.8460.0440.8520.994 0.583 L x D x CS 0.0450.0270.8260.404 0.054 SE 4.37 1.82 1.12 0.29 6.93 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 120

105 D-L on both one(P = 0.023) and two(P = 0.00 2) year plots. Defoliation of mucuna decreased residue N content relative to UDM on one(P = 0.005) and two(P = 0.083) year plots. Defoliation of lablab also re duced residue N content compared to UD-L on one(P = 0.094) and two(P = 0.028) year pl ots. In agreement with previous reports (Tian et al., 2000), mucuna generally accumulated more N in the total biomass than lablab. Bean Grain and Straw Yield Mucuna treatments resulted in higher (P = 0.004) subsequent bean grain yield than lablab (Table 4-10), but defo liation regime had no effect (P = 0.435) on bean grain yield. Mean bean grain yield on plots after 1 yr of residue application (437 kg ha-1) was not different (P > 0.830) from those where the residue had been applied for 2 yr (446), suggesting that there was no apparent adva ntage of 2 yr versus 1 yr of residue application. Single degree of freedom comparisons showed that mucuna treatments tended (P = 0.158) to yield higher bean grain yi eld than the natural fallow, and the yields under mucuna treatments were similar to those from inorganic N (P = 0.847) and cattle manure (P = 0.720). Legume species affected (P = 0.069) bean straw DM yield (Table 410), with mucuna plots outyielding lablab pl ots. Defoliation did not affect (P = 0.892) straw DM yield and neither did residue applica tion for 2 yr years versus 1 yr (P = 0.764). The yields of bean grain and straw yiel d show that mucuna residue application resulted in higher yields th an lablab. The D-M treatment performed particularly well after 2 yr of incorpor ation suggesting that it may have mo re residual effects than the other treatments. Overall, there was no appare nt advantage to appl ying residue for two consecutive years compared to one year.

PAGE 121

106 Table 4-10. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation of mucuna and lablab grain and straw yield of common bean intercropped in succeeding maize. Treatments GrainStraw ---------------kg ha-1 --------------After 1 year residue UD-M† 527379 D-M 499357 UD-L 342299 D-L 380297 After 2 years residue UD-M 428318 D-M 589397 UD-L 403359 D-L 362290 Controls N fertilizer 493338 Natural fallow 379310 Cattle manure 544328 Effects ----------------P values -------------Legume (L) 0.0040.069 Defoliation regime (D) 0.4350.892 L x D 0.4080.247 Cropping sequence (CS) 0.8300.764 L x CS 0.7520.495 D x CS 0.5010.757 L x D x CS 0.1160.132 SE 25.7413.91 † UD-M = Undefoliated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 122

107 Bean yield in this study was affected by both low fertility and root rot disease caused by a complex of fungal pathogens ( Fusarium solani , Rhizoctonia solani , and Pythium spp. ), which have been observed to be severe in western Kenya where bean production is intensive (Otsyula et al., 1998). The beans yields were higher, however than those reported from the highlands of east Africa (Wortmann et al., 2000) when maize-bean intercrops were preceded by mucuna and lablab green manures. Maize Grain and Stover Yield There was interaction of legume, defolia tion, and cropping sequence for maize grain yield (P < 0.002) (Table 411). After 1 yr of residue application, there were no differences (P > 0.553) between mucuna and lablab on maize grain yield. When the legume residue was applied for two consecuti ve years, D-M resulted in higher (P = 0.028) maize grain yield than D-L, but UD-L a ttained higher (P < 0.006) grain yield than UD-M, suggesting that residual effects may be greatest with D-M and UD-L. Defoliation of both mucuna and lablab te nded (P > 0.123) to decrease maize grain yield after 1 yr of residue app lication, but after 2 yr of re sidue application, defoliation of mucuna increased (P = 0.053) the grain yield compared to UD-M, but D-L resulted in a reduction (P < 0.001) in maize grain yield co mpared to UD-L. For maize grain yield, there tended to an advantage of applying the re sidue for 2 compared to 1 yr for D-M (P = 0.134), and UD-L (P = 0.059) but not for UD-M a nd D-L. The average maize grain yield for the legume treatments after the second year vs the first year of legume was 5.3 vs 5.1 t ha-1, which does not suggest much overall advantage. Single degree of freedom comparisons showed that the average legume treatment after 2 yr of incorporation (5.3 t ha-1) yielded higher (P < 0.055) ma ize grain yield than the

PAGE 123

108 natural fallow control (4.1 t ha-1). The differences were greater (P < 0.026) when only the undefoliated treatments were compared to the natural fallow. Defoliated mucuna and UD-L outyielded (P = 0.033, and P = 0.005, resp ectively) the natural fallow, but UD-M and D-L did not. Also the inorganic N a nd cattle manure treatments outyielded (P < 0.001, and P = 0.002, respectively) the natural fall ow. The yields from UD-L plots were similar (P = 0.444) to those from cattle ma nure, but both UD-L and D-M, which yielded the highest maize grain yields among the legume residue trea tments, achieved lower (P = 0.048, and P = 0.007, respectively) yields compared to inorganic N. The interaction of legume type, defoliation re gime and years of residue application on maize stover yield was also signi ficant (P = 0.035) (Table 4-11). After 2 yr of residue application, D-M resulted in higher (P < 0.001) maize stover yield than D-L, and there was trend (P = 0.131) for UD-L to outyield UD-M. Also after 2 yr of residue application, D-L reduced (P = 0.017) stover yield compar ed to UD-L, whereas D-M increased (P =0.095) stover DM yield rela tive to UD-M. There was no apparent advantage of applying residue for two consecutive years on maize stover yield, except for UD-L which achieved higher (P = 0.032) stover yield than after 1 yr of resi due application. Relative to the controls, D-M and UD-L after 2 yr of re sidue application and the inorganic N and cattle manure control treatments yielded the hi ghest stover yield, but only the inorganic N attained higher (P < 0.039) yi elds than the natural fallow.

PAGE 124

109 Table 4-11. Effects of cropping sequence (1 yr versus 2 consecutive yr), legume, and defoliation of mucuna and lablab on grai n and stover yield of succeeding maize. Treatments GrainStover ---------------t ha-1 ----------------After 1 year residue UD-M† 5.188.04 D-M 4.818.93 UD-L 5.598.29 D-L 4.698.23 After 2 years residue UD-M 4.468.15 D-M 6.0811.33 UD-L 6.7211.05 D-L 3.987.12 Controls N fertilizer 7.1811.63 Natural fallow 4.119.34 Cattle manure 6.489.85 Effects --------------P va lues --------------Legume (L) 0.7560.616 Defoliation regime (D) 0.0300.396 L x D < 0.0010.012 Cropping sequence (CS) 0.5000.202 L x CS 0.7400.403 D x CS 0.9760.771 L x D x CS 0.0020.035 SE 0.210.48 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 125

110 Maize grain yields were hi ghest for D-M and UD-L among th e residue treatments after 2 yr of residue application. These higher yiel ds may possibly be attributable to the lower quality of these residues. Defoliation of mu cuna reduced the proportion of leaves, which have the highest N concentration, whereas lablab residue was comprised of a higher proportion of stem which ha s low N and high lignin. Long-Term Residual Effects of Residue Application Bean Grain and Straw Yield The residual effect of muc una and lablab residue app lied during 2000 growing season on 2001 bean grain and straw yi elds are reported in Table 4-12. The interaction of legume and defoliation approached significan ce (P = 0.120) because defoliation had no effect (P = 0.966) on mucuna plots but tende d to affect (P = 0.107) lablab plots. Single degree of freedom contrasts show ed that the yields under UD-M were comparable (P = 0.254) to D-M, but UD-L yi elded higher (P = 0.098) bean grain yield than D-L. The bean yields from legume resi due treatments were not different (P > 0.10) than those from the three controls. Maize Grain and Stover Yield The effect of legume and defoliation on mai ze grain and stover yield from plots where the legume residue was applied the previous season were not significant (Table 13). Single degree contrasts showed that yields were higher unde r D-M (P = 0.052), UDL (P = 0.085), and D-L (P = 0.005) plots than the na tural fallow, but yields under UD-M only tended (P = 0.144) to be higher. Also inor ganic N and cattle manur e treatment yields were higher (P < 0.001, and P = 0.004, respectively) than the natural fa llow. The yields on D-L plot were comparable (P = 0.450) to the inorganic N, but D-M (P = 0.024), and UD-L (P = 0.013) plots were lower than in inor ganic N plots. The yi elds of maize grain

PAGE 126

111 suggest that D-M, UD-L, and D-L likely had higher residual effects, possibly because of the lower residue quality of th ese treatments (Chapter 3). These results suggest that farmers could intercrop legumes in alternati ng years and still reali ze some benefit from residue application in the y ear following no application, esp ecially for residues having lower leaf:stem ratio, lower N concentr ation, and higher lignin concentration. Conclusions Legume residue incorporation increased maize grain and stover DM yield compared to the natural fallow in agreemen t with the literature (Akobundu et al., 2000; Ibewiro et al., 2000a; Ile et al., 1996; Tian et al., 2000). Two years of residue appli cation resulted in a greater yield increases for D-M and UD-L treatme nts, suggesting that the residual effects of these treatments may be gr eater than those of UD-M. Pr evious studies (Myers et al., 1994; Handayanto et al., 1994; 1997; Mafongoya et al., 1997a; Vanlauwe et al., 1997) have shown that residues of very hi gh quality (i.e., high N and low lignin) release nutrients quickly while those of intermediate quality release nutrients at a slower rate perhaps in better synchrony with crop demand. Undefoliated lablab residue consisted of a higher proporti on of stem than UD-M and the stem was of lower N (Table 4-3) and hi gher lignin concentrations (Chapter 3) than that of mucuna. Defoliation of mucuna redu ced the proportion of leaves and increased the proportion of stems (Table 4-2) which contained lower N (Table 4-3) and higher lignin concentrations (Chapter 3). These differences sugge st that residues of both

PAGE 127

112 Table 4-12. Residual effects of mucuna and lablab residue application in March 2000 on grain and straw yield of common bean intercropped in succeeding maize in 2001. Treatments GrainStraw ---------------kg ha-1 --------------UD-M† 437376 D-M 433310 UD-L 553365 D-L 268267 Controls N fertilizer 493338 Natural fallow 379310 Cattle manure 544328 Effects ----------------P values -------------Legume (L) 0.7660.593 Defoliation regime (D) 0.1120.123 L x D 0.1200.753 SE 22.712.1 † UD-M = Undefoliated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 128

113 Table 4-13. Residual effects of mucuna and lablab residue application in March 2000 on grain and stover yield of succeed ing maize in November 2001. Treatments GrainStover ---------------t ha-1 -----------------UD-M† 5.29.0 D-M 5.68.9 UD-L 5.59.1 D-L 6.69.5 Controls N fertilizer 7.211.6 Natural fallow 4.18.8 Cattle manure 6.59.9 Effects --------------P va lues --------------Legume (L) 0.3490.675 Defoliation regime (D) 0.2290.824 L x D 0.5410.738 SE 0.170.36 † UD-M = Undefolated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 129

114 UD-L and D-M were of intermediate quality. The data on maize grain and stover yields suggest that UD-M may have re leased nutrients rapidly, thus some may have been lost before they were taken up by the long maturity maize hybrid, whereas UD-L and D-M could have released nutrients in greater synchrony with crop demand. Thus defoliation of mucuna is thought to ha ve increased the efficiency of N uptake (Chapter 3) and resulted in gr eater residual effects. Defoliati on of lablab resulted in such a large biomass reduction that the beneficial effects of lo wered quality could not be realized in the first year. However, D-L ha d the greatest second-year residual effect for maize grain yield. This may have occurred b ecause after defoliating lablab, which has an upright growth habit, the stubble left behind is comprised of mainly the stem fraction, which contained low N and high lignin concentr ation (Nyambati et al., 2001). Thus this low quality residue resulted in slower deco mposition and nutrient re lease (Handayanto et al., 1997; Vanlauwe et al., 1997; Palm et al., 2001), reduc ing losses and enhancing nutrient use efficiency and thus higher resi dual effects. Therefore our data tend to support the hypothesis that residues of interm ediate quality enhance the residual effects particularly in high rainfall areas that are prone to more nutrien t losses such as the highlands of northwestern Kenya. Considering these results together with thos e of the on-farm (Chapter 5) and mineral N study (Chapter 3) we conclude that inclusion of green manur e legumes as relay intercrop into the current maize-bean system increased subsequent maize yields. At the research station (this chapter) where th e legume residue biomasses were higher, the responses were greater for both legumes than on-farm (C hapter 5). On farmersÂ’ fields, where the soil fertility is much lower, mucuna was bette r adapted than lablab, resulting in greater

PAGE 130

115 responses to mucuna. Defoliation of mu cuna reduced the residue quality and subsequently enhanced the efficiency of nutri ent use and in some cases resulted in higher maize yields than undefoliated mucuna, wher eas defoliation of lablab, which was of lower quality than mucuna, resulted in le sser maize yield response due to greater proportion of biomass removed (because of its upright growth habit). Residues of intermediate quality (UD-L, and D-M) resulted in higher maize yields after two consecutive years of residue application, and D-M along with D-L had the greatest residual effect following a year with no residue incorporated.

PAGE 131

116 CHAPTER 5 ON-FARM PRODUCTIVITY OF RELAY-CRO PPED MUCUNA AND LABLAB IN SMALLHOLDER CROP-LIVESTOCK SYSTEMS IN NORTHWESTERN KENYA Introduction Declining soil fertility is a ma jor constraint to crop produc tion in smallholder farming systems in many regions of the tropics. Th is is due to a number of factors including continuous cultivation, removal of crop residue s, loss of nutrients through soil erosion, and overgrazing between cropping seasons (L al, 1995; Swift et al., 1994a). These practices are often associated with decreas ing farm size as a result of increasing population pressure. In north western Kenya, smallholder farmers practice mixed farming where livestock, particularly dairy production, and crop production are closely integrated. In these systems both low soil fertility and inadequate livestock feeds are the major constraints to production (Nyamb ati, 1997; Muyekho et al., 1998). Negative nutrient balance, particularly N a nd P, on smallholder farms is a main factor causing food production to decline in sub-Saha ran Africa despite an increasing trend in the rest of the world (Swift et al., 1994a). The potential to increase crop and livestock production in these systems is limited due to very minimal use of external nutrient sources in the form of inor ganic fertilizers and feed s upplements (FAO, 1995). It is estimated that the average inorganic fertilizer application in sub-Saharan Africa is less than 10 kg of fertilizer nutrients per h ectare (FAO, 1995; Heisey and Mwangi, 1996; Larson and Frisvold, 1996). These levels are well below crop and soil maintenance

PAGE 132

117 requirements and are likely to remain low because fertilizer is the most costly input used by smallholder farmers in Africa. Under continuous cropping, appropriate crop rotations and the in tegrated use of inorganic fertilizers and orga nic residues can sustain crop pr oductivity (Mitchell et al., 1991; Franzluebbers et al., 1998). Integrat ed nutrient management which seeks to maximize the complementary effects of inorgani c fertilizers and orga nic nutrient sources is emerging as one of the options that hold much promise in increasing crop productivity on smallholder farms (Smaling et al., 1996; Palm et al., 1997; Franzlue bbers et al., 1998). In a long-term experiment in the sub-humid highlands of eastern Af rica, Kapkiyai et al. (1999) showed that improved organic residue ma nagement practices resulted in a slower rate of soil organic carbon decline and a higher particulate soil organic carbon. The soil organic matter changes were also strongly correlated with nutrient availability. Green manure legumes, incorporated into the soil, have the potent ial to contribute N during decomposition (Tian et al., 2000), im prove soil organic matter and soil physical properties (Hulugalle et al., 1996; Tian et al., 1999), suppr ess the growth of weeds (Versteeg et al., 1998a; Akobundu et al., 2000), and be a so urce of high quality feed (Agyemang et al., 2000; Adjorlolo et al., 2001; Chapter 3). The adoption of these legumes on smallholder farms has been very low (Thomas and Sumberg, 1995), however. The main challenge is how best to integrate herbaceous legumes into the existing farming systems. The complementary effects of legumes on cropping have gained greater interest recently (ILCA, 1993) because of declining soil fertility. The integration of legumes into fallow periods (Muhr et al., 1999c) or as rela y crops could improve their acceptability to

PAGE 133

118 smallholder farmers who cannot afford to pur chase external inputs such as commercial concentrates and inorganic fertilizers. Mucuna ( Mucuna pruriens ) and lablab ( Lablab purpureus ) are some of the most promising gr een manure legumes for cropping systems in sub-humid regions of sub-Saharan Africa (Weber, 1996; Ibewiro et al., 2000a; Tian et al., 2000), however, the use of these organi c residues alone may not be sufficient to overcome both N and P deficiencies. The in tegration of small amounts of inorganic P and green manure legumes offers a strategy to meet both N and P requirements of crops (Jama et al., 1997; Palm et al., 1997). Combining the use of inorganic and organic fertilizers has been shown to increase crop yield compared to use of similar N am ounts from inorganic fertilizers (Murwira and Kirchmann, 1993). A better understanding of th e performance of relay-cropped legumes and their effect on the subsequent cereal crop may help in adapting these legumes to cereal-based cropping systems. The current study evaluated the potential of relay cropping mucuna and lablab into the current maize-bean intercrops in comparison with the traditional natural fallow system on farmer sÂ’ fields. The specific objectives were (i) to measure the agronomic effectiveness of relay-cropped mucuna and lablab residue combined with inorganic P compared to na tural fallow, cattle manure, and inorganic fertilizer for subsequent maize-bean test crops, and (ii) to determine the effect of removing part of the biomass for livestock fodder on maize yield unde r low soil fertility on farmersÂ’ fields.

PAGE 134

119 Materials and Methods Experimental Site A researcher-farmer managed on-farm trial was conducted at Tumaini village, 25 km north of Kitale on six farmer fields. Tumaini village is in agro-ecological zone classified as upper midland 4 with an average rainfa ll range of 1000 to 1200 mm (Jaetzold and Schmidt, 1983). The soils are classified as humic Ferralsols based on FAO classification (FAO-UNESCO, 1994) and are equivalent to Kandiudalfic Eustaudox in the USDA soil taxonomy system (Soil Survey Staff, 1994). These are deep, highly weathered and leached soils with low activity clay. Air-dry soil in the top 20 cm had the following characteristics; clay = 39%; sand = 41%; si lt = 20%; pH (1:2.5 soil water suspension) = 5.3; exchangeable acidity = 1.4 cmolc kg-1; exchangeable Ca = 3.1 cmolc kg-1; exchangeable Mg =1.2 cmolc kg-1; exchangeable K = 0.18 cmolc kg-1; extractable P = 5.0 mg P kg-1; soil organic carbon = 19 g kg-1; total soil nitrogen = 1.43 g kg-1; and total soil P = 0.46 g kg-1. The farms were chosen because their production constraints are representative of most sma llholder farms in the sub-humid highlands of northwestern Kenya. These constraints are low soil fertil ity and inadequacy of feeds, particularly during the dry season (Nyambati, 1997; Muyekh o et al., 1998). These constraints were identified using participat ory rural appraisa l (PRA) methods (Chambers, 1997). Participating farmers were selected by a farmer research group, which was formed in 1997 to undertake a farmer participatory eval uation of crop/animal husbandry activities. Pre-experimental Activities A PRA approach was used to diagnose th e production constraints, describe the farming system, and evaluate the results at th e end of the experiment. The participatory research was initiated in 1997, when resear chers from NARC-Kitale and local frontline

PAGE 135

120 extension staff conducted a preliminary char acterization and diagnosis of the farming system using PRA (Muyekho et al., 1998). This was followed by researcher-farmer managed trials focusing on animal health and inadequacy of dry season feeds. Production constraints related to soil fertility were not addressed due to the narrow focus of the initial project. The present study was a follow-up effort to address these constraints. A list of farms with severely low soil fertility were identified by a farmer research committee together with local key info rmants including the local extension staff. A research group comprising a multidisplinar y team of research ers and extension personnel visited the farms to ascertain the fert ility status and suitabi lity of the farms to participate in the researcher-farmer managed, on-farm experiments. The final list was decided after a community meeting where th e objectives of the present study were explained to the farmers. This study is part of a wider effort by the KARI/RF project to develop, with farmers, low cost manu re-based technologies for improved soil management with the goal of alleviating nutrient depletion problems in resource-poor farms. Experimental Treatments and Layout The experiment was laid out as a random ized complete block design with seven treatments on each of the six farms, and farm wa s considered a replicate. Four treatments originated from a 2 X 2 factorial combinat ion of two cropping systems and two legume defoliation treatments. The tw o legume cropping systems were 1) maize + bean (both planted in April) + mucuna (planted in A ugust) and 2) maize + bean (both planted in April) + lablab (planted in August). Th e two legume defoliation treatments were 1) herbage above 10 cm removed at season e nd, and 2) undefoliated. Control treatments included 1) a natural fallow with no inorga nic N, 2) inorganic N fertilized (30 kg N ha-1),

PAGE 136

121 and 3) cattle manure fertilized (5 t ha-1, supplying approximately 65 kg N and 18 kg P ha1). Cattle manure from a large-scale farm was used to avoid confounding manure quality among farms with experimental treatment eff ects. No inorganic N was applied to any plots other than the inorganic control. Th e plot size was 6 m by 3.75 m with an interand intra-row spacing of 75 cm and 30 cm, respectiv ely. The beans were planted between the rows of maize with an intra-row spacing of 30 cm. The experiment was started in March 1999 when all the plots on each farm were hand plowed. Maize and beans were interplanted in April, at the beginning of the gr owing season. Inorganic P was applied in the same hill as for maize at a rate of 13 kg P ha-1 in all the plots except for the cattle manure plot. At 8 wk after planting, the inorgani c treatment was top dressed with 30 kg N ha-1. All the plots were hand weeded twice befo re the beans were harvested in July. In August, the green manure treatment plot s were hand plowed between maize rows, before mucuna and lablab were planted. Af ter maize harvest in November, all the maize crop residues were removed from the plots in accordance with farmer practice, and the green manure legumes were left to continue growing during the dry season fallow period until mid–March 2000 when they were incorporated into the soil. The succeeding maize and beans were planted again in April. Th e same management operations were repeated before the second succeeding maize crop was planted in April 2001. Green Manure Defoliation Management and Sampling After harvesting maize at the end of November 1999 and 2000, mucuna and lablab were left to continue growing until land preparation time for the next growing season in mid-March 2000 and 2001 (30 wk after planting). Legumes were sampled and defoliation management instituted before th e residue was incorporated. The biomass production was assessed through litter fall, above-ground le gume biomass (leaf and

PAGE 137

122 stem), and root. For plots that were defo liated, the legumes were cut to a stubble of 10 cm. Prior to clipping the entire plot, two representative 0.5-m2 quadrats were sampled. The upper canopy herbage above 10 cm was co mposited across the two sites per plot, weighed fresh, and sub-samples were ta ken for dry matter yield and chemical composition determination to evaluate its potenti al for fodder. In these same quadrats, all material below 10 cm was removed at soil level and composited across the two sites per plot. Undefoliated plots were also sampled at two representative 0.5-m2 quadrats per plot. In this case, all material was remove d at soil level. The sampled herbage was separated into above-ground stem plus leaf a nd litter fractions. Fresh weights of each fraction were taken, fractions were sub-sampled, and the rema inder was returned to the two sampling sites in the respective plot. Th e sub-sample was weighed fresh, dried in the oven at 600C for 48 h to constant weight, weighe d again, and then ground for analysis. Within each sampled quadrat, a root sample was taken. A soil core from 0.5-m2 to a depth of 20 cm was carefully retrieved from the sampling area. All roots were removed, washed with clean water through a 0.5-mm siev e to remove all soil, and finally rinsed with distilled water. Th e roots were dried at 700 C for 48 h, weighed, and ground for analysis. To characterize the experimental site, soil samples were taken from all seven treatments at the start of the growing season of the initial year of the experiment in 1999. At each sampling period, soil samples were taken from six sampling points from the top 20 cm, composited, thoroughly mixed, and then sub-sampled. Cattle manure samples were taken from the large-scale farm that provided the treatment manure and from every participating farmer to characterize the quali ty of cattle manure available on smallholder farms. The cattle manure samples were air dried and ground for analysis.

PAGE 138

123 Chemical Analysis The top-canopy herbage above 10 cm was an alyzed for dry matter (DM), ash, crude protein (CP), neutral detergent fiber (NDF ), and in vitro dry matter digestibility (IVDMD). Residue biomass samples (below 10 cm in defoliated plots and all herbage to soil level in undefoliated plots) were analyzed for N, P, K, Ca, and Mg. The soil samples were analyzed for extractable P and K and fo r total N, P, and organic carbon, and for soil particle size. Legume biomass, N concentr ation, and N yield were determined for both the 1999/2000 and 2000/2001 seasons, whereas P, K, Ca, and Mg concentration and yield were determined during the 1999/2000 season only. The cattle manure samples were analyzed for N, P, K, Ca, and Mg. Total N, P, and K of plant samples were analyzed by Kjeldahl digestion with concentrated sulfuric acid (Anderson and Ingram, 1993), followed by colorimetric determination for N and P (Parkinson and Allen, 1975) and flame photometry for K (Anderson and I ngram, 1993). Calcium and Mg were determined following the procedure outlined by Anderson and Ingram (1993). The soil extractable P and K were determined using the modified Olsen method and the total soil N was determined by Kjeldahl digestion modi fied to include salicylic acid for the recovery of soil nitrate (Ande rson and Ingram, 1993). Total soil carbon was determined by wet oxidation with acidified dichromate a nd external heating followed by colorimetry (Anderson and Ingram, 1993). Ne utral detergent fiber was de termined according to the method of Van Soest et al. (1991). The I VDMD was determined using the Tilley and Terry (1963) method. Statistical Analysis The general linear models pr ocedure of SAS was used to test treatment effects on plant and soil responses (SAS, 2001). For res ponse variables pertaini ng only to the four

PAGE 139

124 legume treatments, the model included e ffects of legume, defoliation, and their interaction. For response variables pertaining to all seven treatments in the study, a first analysis tested the effect s of legume, defoliation, and their interaction, and it was followed by single degree of freedom contra sts that compared the legume treatments (individually or in groups) with the controls. To account fo r other variation in farmer resources and management, maize and bean grain yields were analyzed using the researcher and farmer oriented criteria of stability analys is (Hildebrand, 1984; 1996). In this approach the variations in physical or biological environments are expressed as an index based on yield of all treatments on each farm. The calculated environmental index reflects all the favorable and unfavorable factors found on farms in cluding soil fertility status and farmer management that affect re sponse of the technology being tested. Using regression analysis, the yield response of each treatment is related to different environments. Farmer Evaluation The potential acceptability of the legumes by farmers for enhancing soil fertility and the dry season feed supply was assessed in a participatory evaluation. The farmer evaluation was done by the six participating fa rmers using a ranking matrix in the third year of the experiment after 2 yr of residue incorporation. The farmers were asked to discuss as a group and decide the criteria to be used to evaluate the four legume treatments. The farmersÂ’ criteria were used to make a matrix which served as an instrument for farmers to decide which legume was better for each criterion. The matrix was also used to rank the seven experimental treatments on their farms for soil fertility improvement potential (effect on maize grain yield).

PAGE 140

125 Results and Discussion Farmers’ Resource Endowment and Priority Setting The average farm size of smallholder farms in the study area was less than 2 ha and most farmers have mixed systems in wh ich crops and dairy production are closely integrated. The most importa nt crops grown include: maize, beans, vegetables, bananas ( Musa cavendishii ), finger millet [ Eleusine coracana (L.) Gaertn], cassava ( Manihot esculenta Crantz), sweet potatoes ( Impomea batatas ), and fruit trees. The cropping system is characterized by continuous cropping as a result of land intensification due to population pressure. One maize crop is grow n per year and it is intercropped with common bean in April. In A ugust after harvesting the first crop of beans, a second crop of beans or sweet potatoes is sometimes rela y-cropped in maize. Vegetables (mainly Brassica spp. ) are planted in October near the rive rs for hand irrigation during the dry season from December to late March. Declini ng soil fertility is the major constraint to crop productivity. Maize, beans, milk, and sweet potatoes are the main foods consumed at home. Livestock are an important en terprise of the farming syst em. They are a source of income, food, manure, and traction. The lives tock feeding system may be described as semi-zero grazing where crop residues and hom estead and roadside grazing play an important role. The commonly used feeds are natural pasture grazing, maize stover, napiergrass [ Pennisetum purpureum (Schum.) cv. Bana], bean straw, sweet potato vines, weeds, and banana pseudostems. Maize stover is also used as a source of fuelwood and construction of farm structures. Inadequacy of feeds and livestock diseases are the major constraints to livestock production. The feed s hortage constraint is particularly critical

PAGE 141

126 during the dry season. There are few improved pastures and there is a general lack of knowledge concerning fodder legumes and fodder shrubs. Labor for farming activities is predomin antly provided by the household members. Labor for growing of finger millet, bean, and vegetable production is mainly provided by women. Men are involved in livestock feeding, maize production, and in cash crop production, such as the growing of vegetabl es during the dry season. Generally women and children provide most of the labor for farm activities. The use of inorganic fertilizers is limited. Manure is an importa nt and highly valued output of livestock. The manure is collected near the homestead where the cattle are enclosed at night. There is no deliberate effo rt to store the manure properly, and it is left in the open until needed. Apart from the re lay cropping of sweet potatoes in maize after harvesting a first crop of beans, there is no other form of improved fallows. Soil and Cattle Manure Characteristics Soil samples from various plots indicated that soils are moderately acidic, with fair amounts of exchangeable bases, and marginally available P. The soils were deficient in N. The N concentration of farmers’ cattle manures ranged from 6.2 to 13.8 g kg-1 (Table 5-1). Phosphorus concentra tion did not fluctuate much be tween farms, but there was a wide variation in the concentr ations of K, Ca, and Mg of the manures. On average the nutrient concentrations (N, P, K, Ca, and Mg ) in farmers’ manure were lower compared to the large-scale farm manure used in the experiment and were in the range reported from smallholder farms in the central highla nds of Kenya (Lekasi, 1998). The difference in organic C and N between the smallholde r farmer and commercial farm manure could be due to differences in diet (Somda et al., 1995), methods of collection and storage (Probert et al., 1995), degree of decompositi on, and handling conditions of the manures

PAGE 142

127 (Murwira, 1995). Farmers’ manures were contaminated with undecomposed crop residues and soil, because they were either left in the ‘boma’ (night enclosure for cattle) for several days before collecti on or were stored on bare ground. Table 5-1. Nutrient concentration of cattle manures from smallholder farms and a largescale farm used in the experiment. Nutrient Farmer N P Ca Mg K -----------------------------g kg-1 ------------------------------Teresa 11.1 2.08 7.0 1.89 14.2 Khaukha 13.5 2.93 10.9 4.01 13.4 Baraza 13.8 2.82 11.5 3.32 9.2 Simiyu 13.2 2.32 10.5 1.54 7.5 Nasambu 6.2 1.17 4.7 0.94 6.9 Kufwafwa 11.6 1.78 8.3 3.86 9.6 Small farm mean 11.6 2.18 8.8 2.59 7.8 Large-scale farm 17.2 3.69 17.3 6.22 14.6

PAGE 143

128 Herbage Mass and Nutritive Value of Top Canopy Herbage During the 1999/2000 growing season, top-ca nopy herbage mass of lablab was higher than mucuna, but in the 2000/2001 season mucuna yielded more biomass (Table 5-2). On average, defoliation of top canopy herbage provided 0.9 t DM ha-1 yr-1. The top canopy herbage (leaf + stem) of mucuna cont ained higher (P = 0.007) CP and lower (P = 0.001) NDF concentrations than lablab in 1999/2000 (Table 5-2), but there was no difference in IVDMD between the two legumes. Crude protein concentration of mucuna top canopy herbage tended (P = 0.217) to be greater than lablab in 2000/2001 season. The CP concentration of lablab topcanopy herbage was higher and the NDF concentration was lower than that repo rted for above-ground whole plant herbage harvested at 142 d after planting (DAP) (A gyemang et al., 2000). The CP and NDF concentrations of mucuna t op canopy herbage were comparable to above-ground mucuna herbage harvested at 90 DAP (Adjorlolo et al ., 2001). These results suggest that the topcanopy herbage is a potential dr y season protein supplement. Legume Residue Biomass Residue biomass during the 1999/2000 growi ng season was greater than during the 2000/2001 season (Figures 5-1 and 5-2). In both seasons, UD-M produced more (P < 0.001) biomass (2.3 t ha-1, mean of two seasons) than UD-L (0.8 t ha-1). Defoliation reduced (P < 0.01) total residue during 1999/2000 season, but it did not affect (P = 0.373) it during the 2000/2001 season. In the 999/2000 growing season, D-M and D-L supplied 39 and 57%, respectively, as much total residue as from undefoliated treatments. In the 2000/2001 growing season, D-M supplied 79% of the biomass supplied by UD-M, whereas both lablab treatments yi elded low but similar amounts of biomass, due to in part

PAGE 144

129 to greater litter production by the D-L treatment. The low biomass achieved by lablab in both seasons suggest that lablab may be less tole rant to low soil fertility in addition to the shading conditions under maize (Marsdorp and Titterton, 1997) and that lablab was less tolerant to drought than mucuna (Burle et al., 1992), which continued to grow throughout the dry season. In both seasons, mucuna produc ed more root biomass than lablab in agreement with Kolawole and Kang (1997). Mucuna shed more leaves during the 2000/2001 growing season than lablab probably due to sampling which occurred 2 wk later than during the 1999/2000 season. Table 5-2. Biomass and nutritive value of the top-canopy (above a 10-cm stubble) herbage of mucuna and lablab relay cropped in maize on farmers’ fields in Tumaini at Kitale, Kenya. Means are across farms (n = 6). Treatments1 Biomass constituent Mucuna Lablab SE P value Herbage mass --------------t ha-1-------------1999/2000 1.18 1.47 0.18 0.293 2000/2001 0.67 0.30 0.09 0.106 Nutritive value --------------g kg-1 -----------1999/2000 CP 175 118 11.5 0.007 Ash 94 73 4.3 0.022 NDF 374 520 43.4 0.001 IVDMD 659 691 18.4 0.442 2000/2001 CP 179 147 16.2 0.217

PAGE 145

130 0 0.5 1 1.5 2 2.5 3DM (t ha-1) UD-MD-MUD-LD-L TREATMENTS Total residue Leaf + stem Litter Roots Figure 5-1. Mass of various residue fractions of mucuna and lablab relay cropped in maize on farmers’ fields during the 1999/2 000 growing season. Treatments are undefoliated mucuna (UD-M), defoliated mucuna (D-M), undefoliated lablab (UD-L), and defoliated lablab (D-L). P values for legume, defoliation, and legume x de foliation effects, respectively ar e as follows: total residue, < 0.001, < 0.001, and 0.003; leaf + stem, < 0.001, < 0.001, and 0.003; l itter, 0.879, 0.693, and 0.347; roots, < 0.001, 0.585, and 0.212.

PAGE 146

131 0 0.5 1 1.5 2 2.5DM (t ha-1) UD-MD-MUD-LD-L TREATMENTS Total residue Leaf + stem Litter Roots Figure 5-2. Mass of various residue fractions of mucuna and lablab relay cropped in maize on farmers’ fields during the 2000/2 001 growing season. Treatments are undefoliated mucuna (UD-M), defoliated mucuna (D-M), undefoliated lablab (UD-L), and defoliated lablab (D-L). P values for legume, defoliation, and legume x de foliation effects, respectively ar e as follows: total residue, < 0.001, 0.373, and 0.351; leaf + stem, 0.001, 0.015, and 0.111; li tter, 0.004, 0.412, and 0.431; roots, < 0.001, 0.939, and 0.772.

PAGE 147

132 Legume Residue Nutrient Concentration There was no interaction between legume and defoliation effects on legume residue N concentration, except for the root fraction (T able 5-3). During the 1999/2000 season the leaf + stem fraction of mucuna contained hi gher (P < 0.001) N concentration than lablab, and defoliation reduced (P = 0.001) the N c oncentration of both legumes. Treatment effects were not significant fo r litter N concentration. Mucu na roots contained higher (P < 0.001) N concentration than lablab. In the same season, mucuna total residue contained higher (P < 0.001) N concentration than labla b, and defoliation reduced (P < 0.001) the N concentration. During the 2000/2001 season, mucuna leaf a nd stem fraction contained higher (P = 0.067) N concentration than lablab. As in th e first year, there were no treatment effects for litter N concentration during the 2000/ 2001 season. There was a significant (P = 0.018) legume by defoliation effect for root N concentration. The legume by defoliation interaction for P and K concentration of all residue fractions was not significant (T able 5-4). The leaf + stem and root fractions of mucuna contained higher (P < 0.001, and P = 0.022, respec tively) P concentration than lablab, but lablab litter was higher (P = 0.044) than muc una in P concentration. Defoliation resulted in a reduction in P concentration in both the leaf + stem (P = 0.012), and litter (P = 0.091) fractions. Mucuna roots cont ained higher (P = 0.022) P concen tration than lablab. The P concentration in the total residue was highe r (P < 0.001) in mucuna than lablab and defoliation reduced (P < 0.001) P concentration in both legumes.

PAGE 148

133Table 5-3. Nitrogen concentrati on of various mucuna and lablab residue fractions at time of soil incorporation in 2 yr. Treatments Effects Fraction UD-M D-M UD-L D-L SE Legume (L) Defoliation (D) L X D 1999/2000 --------------------------g kg-1 -----------------------------------------P values ----------------Leaf + stem 27.1 18.1 17.4 9.50 1.57 < 0.001 < 0.001 0.756 Litter 13.9 14.9 14.2 13.5 0.29 0.423 0.824 0.212 Roots 16.1 16.1 11.8 10.6 0.62 < 0.001 0.481 0.462 Total residue 25.0 18.1 16.5 10.7 1.20 < 0.001 < 0.001 0.647 2000/2001 Leaf + stem 24.5 20.9 19.2 18.0 1.24 0.067 0.244 0.578 Litter 10.9 10.8 13.5 12.2 1.20 0.322 0.713 0.732 Roots 15.1 12.9 11.5 18.5 1.10 0.583 0.192 0.018 Total residue 16.5 13.7 15.9 18.2 0.98 0.262 0.895 0.122 UD-M = Undefoliated mucuna, D-M = Defoliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 149

134 Legume, defoliation, and their interaction effect s were not significant for K concentration of all the fractions, except in the total residue where de foliation reduced (P = 0.047) K concentration (Table 5-4). The leaf + stem fraction and roots fraction of mucuna contained higher (P < 0.001) Ca concentration than lablab, and defoliation reduced (P = 0.043) the Ca concentration in the leaf + stem fraction (Table 55). The interaction of legume by defoliation was significant (P = 0.033) for Ca concentration in the to tal residue. The total residue of UD-M contained higher (P = 0.041) Ca concentration than UD-L, but there was no difference (P = 0.878) between D-M and D-L. Defoliati on did not affect (P > 0.249) the Ca concentration in the total residue. Magnesi um concentration was higher in the leaf + stem (P = 0.052), litter (P = 0.083), roots (P < 0.001), and total residue (P = 0.095) fractions of mucuna than lablab, but defolia tion had no effect (P > 0.181) (Table 5-5). The higher quality of mucuna residue could be attributed to the slower maturation of mucuna under the northwestern Kenya conditions. At the time of residue sampling/incorporation (30 WAP), lablab had set seed while mucuna had not flowered. The N concentration in mucuna was highe r than the critical level of 20 g kg-1 above which net mineralization of N would be expected (Palm et al., 1997) and the P concentration was within the critic al range of between 2 and 3 g kg-1 for net P mineralization (Singh et al., 1992; Palm et al., 1997; Mafongoya et al., 2000), suggesting a net N and P mineralization due to the appli cation of mucuna green manure residue. The P concentration of mucuna and lablab residue used in this study was similar to that reported from the sub-humid highlands of eastern Africa (Wortmann et al., 2000).

PAGE 150

135Table 5-4. Phosphorus and K concentrations of various mucuna and lablab resi due fractions at time of soil incorporation during the 1999/2000 season. Treatments Effects Fraction UD-M D-M UD-L D-L SE Legume (L) Defoliation (D) L x D -------------------------g kg-1 -------------------------------------------P values ---------------Phosphorus Leaf + stem 2.50 2.10 1.80 1.30 0.11 < 0.001 0.012 0.919 Litter 0.89 0.81 1.30 0.93 0.06 0.044 0.091 0.263 Roots 1.40 1.30 1.10 1.10 0.06 0.022 0.394 0.826 Total residue 2.28 1.72 1.65 1.12 0.10 < 0.001 < 0.001 0.915 Potassium Leaf + stem 11.5 10.6 10.4 11.0 0.87 0.709 0.874 0.426 Litter 5.0 4.2 5.5 5.6 0.57 0.248 0.713 0.534 Roots 5.80 5.20 5.60 5.90 0.46 0.596 0.722 0.350 Total residue 10.4 8.14 9.50 8.31 0.80 0.662 0.047 0.518 UD-M = Undefoliated mucuna, D-M = Defoliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 151

136Table 5-5. Calcium and Mg concentration of various mucuna and lablab residue fracti ons at time of soil incorporation during th e 1999/2000 season. Treatments Effects Fraction UD-M D-M UD-L D-L SE Legume (L) Defoliation (D) L x D Calcium -------------------------g kg-1 ---------------------------------------------P valu es ---------------Leaf + stem 19.5 15.6 12.2 10.3 1.13 < 0.001 0.043 0.449 Litter 26.2 25.0 21.2 28.9 1.77 0.808 0.236 0.096 Roots 5.90 5.10 11.2 12.8 0.78 < 0.001 0.572 0.133 Total legume 18.3 14.6 13.2 14.1 0.96 0.013 0.191 0.033 Magnesium Leaf + stem 5.70 5.30 4.80 4.10 0.35 0.052 0.261 0.727 Litter 5.84 6.09 7.21 8.02 0.71 0.083 0.570 0.749 Roots 5.80 4.80 2.60 2.40 0.37 < 0.001 0.181 0.310 Total legume 5.71 5.36 4.92 4.59 0.34 0.095 0.453 0.990 UD-M = Undefoliated Mucuna, D-M = Defoliated Mucuna, UD-L = Undefoliated Labla b, D-L = Defoliated Lablab

PAGE 152

137 Defoliation resulted in a reduc tion of N, P, and Ca concentration of the above-ground residue. This reduction of N results in a wi der lignin:N ratio, which lowered the residue quality of the remaining stubble (Nyambati et al., 2001). Legume Residue Nutrient Content During the 1999/2000 season, there was legum e by defoliation interaction for N content in the leaf + stem fr action (P < 0.001) (Table 5-6). The N content in mucuna leaf + stem fractions was greater (P < 0.01) than lablab, but much more so for undefoliated than defoliated plants, and defoliation reduced (P < 0.001) the N content of both legumes. Neither legume or defoliation affected (P > 0.684) N content of the litter. The N content of mucuna roots was greater (P < 0.001) than lablab. There was legume and defoliation interaction (P < 0.001) for total residue N content. The N in mucuna total residue was greater (P < 0.01) than in la blab regardless of defoliation treatment, and defoliation reduced the N content in total residue of both mucuna (P < 0.001) and lablab (P = 0.013), but to a much greater degree for mucuna. The interaction of legume by defoliation for N content of leaf + stem fraction was significant (P = 0.038) during the 2000/2001 seas on (Table 5-6). Undefoliated mucuna had greater (P = 0.007) N content than UD-L, but there was only a trend (P = 0.120) toward a difference between defoliated treatm ents. Defoliation of mucuna reduced (P = 0.006) N content of mucuna leaf + stem frac tion, but it did not aff ect (P = 0.308) lablab, probably because of low yields obtained from lablab treatments. Mucuna litter and roots had greater (P = 0.043, P = 0.001, respectively) N content than lablab. Legume by defoliation interaction for N content of total residue was significant (P = 0.044).

PAGE 153

138Table 5-6. Nitrogen content of mucuna and lablab resi due fractions at time of soil incorporation in 2 yr. Treatments Effect Fraction UD-M D-M UD-L D-L SE Legume (L) Defoliation (D) L x D ---------------------------kg ha-1 ----------------------------------------P Values ------------1999/2000 Leaf and stem 57.1 12.713.52.3 4.60 < 0.001 < 0.001 < 0.001 Litter 2.7 1.71.92.2 0.37 0.783 0.684 0.403 Roots 5.0 4.41.41.8 0.43 < 0.001 0.799 0.345 Total residue 64.8 18.816.86.3 4.90 < 0.001 < 0.001 < 0.001 2000/2001 Leaf and stem 20.1 7.05.23.4 1.74 0.002 0.007 0.038 Litter 10.0 9.92.35.3 1.30 0.043 0.457 0.150 Roots 2.6 2.10.61.3 0.22 0.001 0.848 0.104 Total residue 32.7 19.18.19.9 2.75 0.001 0.093 0.044 UD-M = Undefolated mucuna; D-M = Defoliated mucuna ; UD-L = Undefoliated lablab; D-L = Defoliated lablab

PAGE 154

139 The N content in UD-M residue was greater (0.002) than UD-L, but there was only a trend (P = 0.201) towards a difference betw een defoliated treatments. Defoliation of mucuna reduced (P = 0.044) N content of tota l residue, but it did not affect (P = 0.771) lablab. The interaction of legume by defoliation was significant (P < 0.001) for P content of leaf + stem fraction (Table 57). The leaf + stem fraction of mucuna had greater (P < 0.01) P content than lablab, and defoliation reduced (P = 0.001) the P content of both legumes but more so for mucuna. Neither le gume or defoliation effects were significant for litter P content. The P content in mucuna roots was greater (P < 0.001) than lablab. The P content in the total residue followed similar trends as in the leaf + stem fraction (Table 5-7). The interaction of legume by defoliation for K content in the leaf + stem fraction was significant (P = 0.018) (Table 5-7). Both UD-M and D-M had greater (P = 0.009, and P = 0.069, respectively) K content than UD-L and D-L. Defoliation reduced the K content of both mucuna (P = 0.008) and lablab (P = 0.035) leaf + stem. The K content in the litter fraction was not affected by treatmen t. Mucuna roots had greater (P < 0.001) K content than lablab. The inte raction of legume by defoliation was also significant (P = 0.019) for the total residue K content. Th e K content in total residue of UD-M was greater (P = 0.019) than UD-L, but there was only a trend toward a difference (P = 0.123) between defoliated treatments. Defoliation decreased (P = 0.024) the K content of mucuna total residue, but it only tended (P = 0.120) to affect lablab residue.

PAGE 155

140Table 5-7. Phosphorus and K cont ent of various fractions of mucuna and lablab relay cropped in maize on farmers’ fields in 1999/2000. Treatments Effects Fraction UD-MD-MUD-LD-L SE Legume (L) Defoliation (D) L x D ------------------------kg ha-1 ----------------------------------P valu es ------------Phosphorus Leaf + stem 5.21.31.40.30.40 < 0.001 < 0.001 < 0.001 Litter 0.20.10.20.10.03 0.494 0.385 0.817 Roots 0.40.40.10.20.04 < 0.001 0.888 0.434 Total residue 5.71.81.70.60.42 < 0.001 < 0.001 < 0.001 Potassium Leaf + stem 25.27.07.63.02.24 0.001 0.001 0.018 Litter 1.00.50.50.80.16 0.767 0.747 0.270 Roots 1.81.60.61.00.19 < 0.001 0.858 0.135 Total residue 28.19.08.84.82.35 0.001 0.001 0.019 UD-M = Undefoliated mucuna, D-M = Defoliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 156

141 The interaction of legume by defoliation eff ects for Ca content in the leaf + stem fraction was significant (P < 0.001) (Table 5-8), however, mucuna had greater (P < 0.030) Ca content than lablab, and defoliati on decreased (P < 0.009) Ca content in the leaf and stem fraction of both legumes. Th ere was an interaction between legume and defoliation for Ca content of the roots (P = 0.028). There was no difference (P = 0.181) between undefoliated treatments in Ca cont ent, but D-L tended to have greater (P = 0.092) Ca content in the root than D-M. Defoliation had no effect (P = 0.174) on Ca content of mucuna, but increased (P = 0.092) content in lablab. The interaction of legume by defoliation for Ca content in the total residue was significant (P < 0.001). Calcium content in the tota l residue was greater (P < 0.001) in UD-M than UD-L, but there was no difference (P = 0.242) between the defoliated treatments. Defoliation decreased Ca content in mucuna residue (P < 0.001), but had no effect (P = 0.347) on lablab residue. The treatment effects on Mg content followed the same trends as for Ca, except that Mg content in mucuna roots was greater (P < 0.01) than lablab. The leaf and stem fraction contributed the highest proportion of the total N, P, K, Ca and Mg supplied by the residues. During the 1999/2000 season, defoli ated treatments of mucuna and lablab contributed 29 and 36% the amount of N supplied by the undefoliated treatments, respectively. Under the UD-M and UD-L treatments, the above-ground fraction contributed the highest amount of N (88% and 80%, respecti vely) and root contributed 7.8 and 8.4%, respectively. During the 2000/20 01 growing season D-M contributed 58%

PAGE 157

142Table 5-8. Calcium and Mg content of vari ous fractions of mucuna and lablab relay cropped in maize on farmers’ fields in 1999/ 2000. Treatments Effects Fraction UD-M D-M UD-L D-L SE Legume (L) Defoliation (D) L x D ---------------------kg ha-1 --------------------------------------P values ----------------Calcium Leaf + stem 41.1 10.3 9.9 2.6 3.43 < 0.001 < 0.001 < 0.001 Litter 4.6 2.9 2.4 5.3 0.92 0.948 0.731 0.201 Roots 1.8 1.3 1.3 2.2 0.16 0.560 0.504 0.028 Total residue 47.5 14.5 13.6 10.1 3.87 < 0.001 <0.001 < 0.001 Magnesium Leaf + stem 11.8 3.3 3.8 1.1 0.92 < 0.001 < 0.001 0.002 Litter 1.0 0.6 0.9 1.3 0.21 0.453 0.888 0.328 Roots 1.6 1.3 0.3 0.4 0.12 < 0.001 0.289 0.049 Total residue 14.3 5.2 5.0 2.8 1.07 < 0.001 < 0.001 <0.001 UD-M = Undefoliated mucuna, D-M = Defoliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab

PAGE 158

143 of N supplied by UD-M, whereas D-L supplied about the same amounts of N as UD-L due to the higher proportion of litter fraction which constituted 65% under D-L vs 35% in UD-L. Generally the results show that mucuna residue supplie d greater amounts of nutrients than lablab, and that defoliation resulted in a decrea se of nutrient content in the residues, especially for mucuna. Yield Responses Bean Grain and Straw Yield There were no year by treatment interacti ons (P > 0.10) for bean grain and straw yields, therefore the data ar e presented averaged across the 2 yr (Table 5-9). The interaction of legume by defoliation for bean yield was not significant. Mucuna plots resulted in higher (P = 0.002) subsequent bean grain yield than labl ab, and defoliation did not affect (P = 0.485) bean grain yield. The low bean yield under lablab treatment could be due to the low nutrient contribution fr om lablab and the damage by aphids ( Aphis craccivora ) and leaf rust disease (Anthracnose, caused by Colletotrichum spp .) that affected both lablab and common bean. Si ngle degree of freedom comparisons showed that cattle manure yielded higher (P = 0.076) be an grain than the na tural fallow control, however, UD-M tended (P = 0.119) to yield high er bean grain than the natural fallow control. The bean grain yields in this study were at the lower end of the range (0.5 to 1.4 t ha-1) reported previously from the highlands of eastern Africa unde r maize intercrops (Giller et al., 1998; Wortmann et al., 2000). Thes e low yields could in part be due to the bean rot disease (caused by a complex of f ungal pathogens) and low soil fertility, which has been shown to cause low yields in the sub-humid highlands of western Kenya (Otsyula et al., 1999). There was a trend toward greater bean straw DM yield on mucuna

PAGE 159

144 Table 5-9. Mean grain and straw yield of common bean relay cropped in maize after mucuna and lablab residue incorpor ation on farmers’ fields for 2 yr. Treatments GrainStraw -----------------------kg ha-1 ---------------------UD-M 147132 D-M 138105 UD-L 95107 D-L 8081 Effects -----------------------P values ------------------Legume (L) 0.0020.111 Defoliation regime (D) 0.4850.067 L x D 0.8430.959 Year (Y) 0.0380.241 SE 10.58.16 Controls Inorganic N 10296 Natural fallow 11094 Cattle manure 156132 UD-M = Undefoliated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab No treatment by year interactio ns were significant (P > 0.10).

PAGE 160

145 plots (P = 0.111), and defoliation resulted in a decrease in straw yields. Undefoliated legume treatments tended (P = 0.135) to yield higher straw DM than the natural fallow control. Maize Grain and Stover Maize grain yield during 1999 (data not s hown), when legumes were first relay cropped in maize, was not different among treatme nts, suggesting that under this type of relay cropping competition from the legumes was not severe. Because the interaction of year by treatment was not significant (P > 0.10) , mean maize grain a nd stover yields are presented across years (Table 5-10). The in teraction of legume by defoliation for maize grain and stover yields wa s not significant (P > 0.194). Maize grain yield after mucuna residue was higher (P = 0.003) than after labla b. Defoliation of the legumes reduced (P = 0.028) maize grain yield. Stover yield was not affected by either legume or defoliation regime (Table 10). Relative to the controls, maize grain yi elds were higher under cattle manure (P = 0.005), inorganic N (P = 0.016), and UD-M (P = 0.023) than the natural fallow control. Defoliated mucuna tended (P = 0.155) to yield higher maize grain than the natural fallow. The maize grain yields under UD-M and DM were not different (P = 0.793, and P = 0.169, respectively) from the inorganic N contro l. The higher maize grain yields under cattle manure treatment is in agreement w ith previous reports (Bationo and Mokwunye, 1991; Probert et al., 1995; Murwira a nd Kirchmann, 1993). Undefoliated legume treatments yielded higher (P = 0.082) maize grai n than the natural fallow control. Only cattle manure and UD-M yielded higher stover DM than the natural fallow control. The UD-M, D-M, and UD-L residue treatments result ed in 97, 54, and 41% maize grain yield

PAGE 161

146 Table 5-10. Mean grain and st over yield of subsequent maize after relay cropped mucuna and lablab residue incorporati on on farmers’ fields for 2 yr. Treatments GrainStover -----------------------t ha-1 ---------------------UD-M 3.193.93 D-M 2.493.25 UD-L 2.283.30 D-L 1.923.38 Effects -----------------------P values ------------------Legume (L) 0.0030.334 Defoliation regime (D) 0.0280.272 L x D 0.4490.194 Year (Y) < 0.001< 0.001 SE 0.210.24 Controls --------------------t ha-1 --------------------Inorganic N 3.323.71 Natural fallow 1.622.91 Cattle manure 3.624.14 UD-M = Undefoliated mucuna; D-M = Defoliated mucuna; UD-L = Undefoliated lablab; D-L = Defoliated lablab No treatment by year interactio ns were significant (P > 0.10).

PAGE 162

147 increases, respectively, compared to the natural fallow control in agreement with previous reports (Versteeg et al., 1998a; Akobundu et al., 2 000; Tian et al., 2000), which have shown positive fertilization effects of green manures. The lower on-farm performance of maize after lablab is attributed to the low lablab biomass and N concentration. Lablab reached maturity and set seed earlier than mucuna which continued to grow during the dry s eason. Mucuna residue contained N and P concentration above the levels at which net mi neralization is expected (Singh et al., 1992; Palm et al., 1997; Mafongoya et al., 2000) a nd accumulated greater amounts of nutrients (Tables 5-6, 5-7, and 5-8). The hi gh maize yield response from 2 t ha-1 of green manure residue is not unexpected, given that this was combined with 13 kg P ha-1. This maize response to a combination of organic residue and P application was consistent with the results of previous studies in western Ke nya (Jama et al., 1997) which showed that a combination of Calliandra calothyrsus biomass and inorganic P was more economical than when organic residue alone was used to supply P. Other st udies in sub-Saharan Africa have indicated a substantial residua l effect of P fertilization at moderate application rates of 20 kg ha-1 (Warren, 1992). These results show that mucuna was adapte d to the on-farm rela y-intercrop conditions more than lablab. Maize yields after mucuna were not affected by defoliation despite the reduction in nutrient contents, suggesting that the nutrient recovery and residual effects under D-M were higher compared to UD-M. This is in agreement with previous reports (Ibewiro et al., 1998; Oikeh et al., 1998), which have showed that even legume roots alone had positive effect on subsequent mai ze grain yield. The defoliated treatments yielded proportionally higher yields than the estimated nutrient contribution, possibly

PAGE 163

148 because of underestimation of the belo w-ground contribution (P eoples et al., 1995), which can constitute as much as 39 to 49% of the total N accumulated by the legume (Ramos et al., 2001). Stability Analysis The stability analysis (Fi gure 5-3) showed that the magnitude of the difference between the undefoliated mucuna residue trea tment, which achieved the highest maize grain yield of all the residue treatments, and the natural fallow control treatment was greater under poor farm environments (i.e., fa rms with low soil fert ility) than under good farm environments. The analysis also show ed that the defoliation of mucuna in poor environments (i.e., farms with very low soil fertility), where the bi omass yield was very low, resulted in very low response of mai ze grain yield, whereas on farms with better soils, D-M resulted in yield approaching that of UD-M. This suggests that use of mucuna topgrowth for livestock feed and soil incorpor ation of the stubble is more likely to be a viable option on higher fertility soils. In an on-station experi ment where the fertility was higher (Chapters 3 and 4), defo liation of mucuna resulted in higher efficiency of N uptake and recovery and subsequent maize grai n yields were at leas t as great as those following UD-M. Farmer Evaluation Mucuna was preferred over lablab in overa ll ranking based on the five farmers’ criteria (Table 5-11), however, the opinions differed between gender. In the context of improving food security, the men preferred D-M because it provided feed for the animals, whereas women preferred lablab because it pr ovided edible grain for the family. The

PAGE 164

149 0 1 2 3 4 5 6 7 8 -2.5-2-1.5-1-0.500.511.522.5 ENVIRONMENTAL INDEXMAIZE YIELD (t ha-1) Cattle manure N fertilizer Natural fallow D-L D-M UD-L UD-M N fertilizer Cattle manure Natural fallow UD-L UD-M D-M Figure 5-3. Stability analysis of maize grain yield (t ha-1) after incorporating undefoliated or defoliated relay cropped mucuna and lablab on farmers’ fields at Kitale, Kenya. A low environmental index is associated with low soil fertility. Treatments are undefoliated mucuna (UD-M), defoliated mucuna (D-M), undefoliated lablab ( UD-L), and defoliated lablab (D-L).

PAGE 165

150 Table 5-11. Farmer ranking (1 = highest) of the green manures for suitability in improving soil fertility and providing fodder. Farmer ranking Criteria Mucuna Lablab Observations Human food 2 1 Lablab provides food for the family Soil fertility improvement 1 2 Plot of mucuna ha d higher maize grain yield, and soil texture of soil was like forest soil, and easy to plow Fodder supply 1 2 Mucuna produces more biomass Weed suppression 1 2 Mucuna spreads to suppress weeds Suitability as intercrop 1 2 Lablab is susceptible to leaf rust due to much rain at establishment in August Mean rank 1.2 1.8 Overall rank 1 2

PAGE 166

151 Table 5-12. Ranking by farmers of the performa nce of maize in various treatment plots. Farmer Treatment 1 2 3 4 5 6 Mean Overall Rank UD-M 4 2 4 6 3 2 3.50 3 D-M 3 3 5 5 7 3 4.33 4 UD-L 5 6 2 7 4 6 5.00 5 D-L 7 5 7 3 5 4 5.17 6 N fertilizer 2 1 3 2 2 1 1.83 1 Weed fallow 6 7 6 4 6 7 6.00 7 Cattle manure 1 4 1 1 1 5 2.17 2 UD-M = Undefoliated mucuna, D-M = De foliated mucuna, UD-L = Undefoliated lablab, D-L = Defoliated lablab 1 = highest and 7 = least. performance of the various treatments based on the farmers’ evaluation was in the order of: N fertilizer > cattle manure > UD-M > D-M > UD-L > D-L > natural fallow (Table 512). Conclusions Undefoliated mucuna was the most effectiv e in improving soil fertility on farm as indicated by the performan ce of succeeding maize and bean crops. In both seasons, mucuna produced more biomass of higher quali ty than lablab, indicating that mucuna was better adapted to low soil fertility on farm, the shaded conditions under maize, particularly in August during seedling establishment, and to drought conditions when it continued to grow. In both 1999/2000 a nd 2000/2001 growing seasons, the low maize grain yield response to the lablab treatment c ould be attributed to the low biomass yield

PAGE 167

152 and nutrient content of lablab. Defoli ation of relay-cropped green legume manure yielded on average 0.9 t ha-1 of top canopy herbage from either mucuna or lablab. Chemical analysis of this herbage showed that it was of high nutritive value and could form a protein supplement during the dry seas on when high quality f eeds are unavailable. Our results show that even after removi ng mucuna upper canopy biomass as feed, the incorporation to the soil of the remaining st ubble and root residues resulted in grain yields higher than those from the natural fall ow, with the exception of sites where soil fertility was extremely poor.

PAGE 168

153 CHAPTER 6 NUTRITIVE VALUE OF TOP-CANOPY HERBAGE OF MUCUNA AND LABLAB RELAY CROPPED IN MAIZE IN THE SUB-HUMID HIGHLANDS OF NORTHWESTERN KENYA Introduction Declining soil fertility and inadequacy of liv estock feeds, particularly the lack of protein during the dry season, are major produ ction constraints in many smallholder, mixed crop-livestock farming systems in the tr opics. These constrai nts arise due to land limitations, which cause farmers to practice continuous cropping and grazing, and to use little or no fertilizer. Although cycling of biomass through livestock and use of manure and urine to fertilize soil have been an important link between livestock and soil fertility (Powell and Valentine, 1998), the quanti ties of manure available on farm may not be enough to replenish nutrients removed in grain and crop residues (Williams et al., 1995). Intercropping of grain legume s such as the common bean ( Phaseolus vulgaris L.) with cereals, provides little or no N to the concur rent or subsequent crop as the majority of fixed N is harvested in the gr ain (Giller et al., 19 91; Giller et al., 1994; Amijee and Giller, 1998). In recent years, there has been a resurgence of interest from the scientific community in leguminous green manure/cover crops in ma ny parts of the tropic s where the use of commercial inorganic N fertiliz ers is not economically feasib le. Although green manures give greater yields of subseque nt crops than intercropping with grain legumes, their use is sometimes limited because they occupy th e land without providing human food or

PAGE 169

154 livestock feed. Thus benefits in addition to improving soil fertility are necessary in order for farmers to adopt them. Few studies have reported the testing of legumes for the combined purposes of enhanced soil fertility and feed supply fo r livestock (Tarawali and Mohamed-Saleem, 1995; Tarawali and Peters, 1996; Muhr et al., 199 9c). In a derived savanna of sub-humid west Africa, Muhr et al. (1999c), tested a ra nge of forage legumes in rotation with maize ( Zea mays L.) in short-term improved fallow syst ems. Studying the rotational effects of forage legumes, Muhr et al. (1999c) found th at even though large amounts of N, P, and K (up to 120, 10, and 135 kg ha-1, respectively) were removed in dry-season herbage, subsequent growth followed by soil incorporat ion of the green manur e biomass increased grain yields of maize gr own on the legume plots. Ibewiro et al. (1998) studying th e N contribution of mucuna ( Mucuna pruriens (L.) DC. var. Utilis (Wright) Bruck), lablab ( Lablab purpureus (L.) Sweet), cogongrass ( Imperata cylindrica ) and maize roots, shoots, and w hole-plant biomass to succeeding maize, showed that, although N in mucuna and lablab roots constituted only 3 and 4% of total legume N, their incorporation increased maize grain yield 38 and 89%, respectively, as much as when whole legume residue was incorporated. These results suggest that incorporation of root and stem stubble afte r removing topgrowth for livestock feed may enhance N contribution to succeeding maize in low-external input, continuous cropping systems as well as provide quality fodder. Weber (1996) concluded that mucuna and lablab are among the species adapted to cropping systems in sub-Saharan Africa that could be relay-cropped into maize-based cropping systems and provide additional benefits to the cropping system. This study is

PAGE 170

155 part of a research program evaluating use of mucuna and lablab for both soil fertility improvement and as livestock feed. The sp ecific objective of this experiment was to assess the nutritive value of the harvested top-canopy biomass of mucuna and lablab when relay cropped in maize in the subhumid highlands of northwestern Kenya. Materials and Methods Study Site and Treatments The study was conducted at the National Ag ricultural Research Center (NARC), Kitale in the sub-humid highlands of north western Kenya. The al titude was 1860 m and the soil a humic ferralsol (oxisol) (pH =5.5) . In the years of this trial, 1999/2000 and 2000/2001 growing seasons, the total annual rainfall received was 1115 and 1050 mm, respectively, falling between Ap ril and November. Other char acteristics of the site were described in Chapter 3. This study was part of the experiment repor ted in Chapters 3 and 4. The treatments were two legume species arranged in three replications of a randomized block design. The legumes were relay cropped in maize in August after harvesting a first crop of common bean intercropped with maize in April. Maize was planted at an interand intrarow spacing of 75 cm by 30 cm (target popul ation of 44,444 plants /ha), respectively, using two seeds per hill of hybr id 614D maize seed in April. The legumes were planted between the maize rows at an intra-row spaci ng of 30 cm using two plants per hill (target population of 88,888 plants/ha). At planting th e maize received a basal application of 13 kg P ha-1. Top-Canopy Biomass Sampling After harvesting maize in November, the legumes were hand weeded and continued growing into the dry season. The legumes were cut to stubble of 10 cm before land

PAGE 171

156 preparation in mid-March, 210 days after plan ting (DAP). Prior to clipping the entire plot, two representative 0.5-m2 quadrats were sampled. Herbage above 10 cm (top canopy) was composited across the two sites per plot and part of the sample was separated into leaf and stem fractions. Fractions were dried, weighed, and ground for laboratory analysis. In these same quadrats, all material below 10 cm was removed at soil level (lower canopy), composited across the two sites per plot, and separated into the same fractions described above. Fresh wei ghts of each fraction were taken, fractions were sub-sampled, and the remainder of the herbage was returned to the two sampling sites. The sub-sample was weighed fresh, dried at 600C for 48 h, weighed again and then ground for analysis. Chemical and Statistical Analysis First-year samples of live l eaf, stem, and leaf plus st em top-canopy fractions from defoliated plots were analyzed for crude prot ein (CP), ash, neutral detergent fiber (NDF), and in vitro dry matter (DM) digestibili ty (IVDMD). Second-year samples were analyzed for CP only. Total plant N wa s analyzed by Kjeldahl digestion with concentrated sulfuric acid, followed by colorimetric determination (Anderson and Ingram, 1993; AOAC, 1990). The NDF was de termined by the method of Goering and Van Soest (1970) as modified by Van Soest et al. (1991). In vitro DM digestibility was determined using the procedure of Tilley and Terry (1963). The general linear models pro cedure of SAS was used to test legume species effects on DM yield and nutritive value. The mode l for DM mass and plant-part proportions included the legume and year effects and their interaction, but only the legume effect was tested for chemical composition data because th ese data were collected only in the first year. Treatment effects were considered significant at P < 0.10.

PAGE 172

157 Results and Discussion Mass and Plant-part Proportions The interactions of the eff ects of legume by year were not significant (P > 0.10) for DM mass and plant-part pr oportions, except for lower canopy leaf and leaf + stem fractions (P = 0.002). Because the primary focus of this work was top-canopy herbage and to simplify presentation of the data, the m eans across the 2 yr are reported in Table 61. There was no difference (P = 0.461) betw een mucuna and lablab in the top-canopy leaf mass, however lablab yielded a greater mass both in the stem (P = 0.023) and leaf + stem fractions (P = 0.096). The top-canopy he rbage of mucuna had a higher (P = 0.001) leaf:stem ratio. Mucuna yielded greater (P = 0.020) leaf mass in the lower-canopy herbage than lablab, but there was no difference (P = 0.389) in lower-canopy stem mass. Also mucuna yielded greater (P = 0.064) mass for leaf + st em in the lower canopy. The greater mass in the top canopy for lablab than mucuna can be explained by growth morphology of the two legumes. Lablab has an upright growth habit, and defoliation to a 10 cm-stubble resulted in a greater (P = 0.011) proportion (0.76) of total DM removed than for mucuna (0.52) (Table 6-1). Defoliation of mucuna removed about the same quantity in the top canopy as that left on site in the lower canopy. The results show that relay cropping mucuna or lablab into maize in August afte r harvesting a first crop of common bean, and defoliating them to a 10 cm-stubble at 210 DAP could on average yield from 1 t ha-1 yr-1 (average of two seasons) of mucuna herbage to 1.8 t ha-1 yr-1 of lablab herbage, at least in soils like those on the research station.

PAGE 173

158 Chemical Composition There were no differences in CP (P = 0.461) or NDF (P = 0.582) in the leaf fraction of top-canopy herbage of mucuna and lablab, but lablab leaf containe d (P = 0.003) higher IVDMD (Table 6-2). The CP concentration of mucuna stem was greater (P = 0.022) than that of lablab, but lablab stem fraction contained higher (P =0.006) NDF concentration than mucuna. There was no difference (P = 0.681) in stem IVDMD between the legumes. The CP and IVDMD of total (leaf + stem) herbage were also not different between the two legumes (Table 6-2), although IVDMD tended to be greater for lablab. Lablab contained greater (P = 0.003) NDF concentration than mucuna. Top-canopy herbage harvested during the 2000/2001 growi ng season was analyzed only for CP, and there was no difference (P = 0.174) in leaf CP concentration, but the stem fraction of mucuna contained higher (P = 0.064) CP c oncentration than la blab (Table 6-3). The CP of top canopy herbage of lablab harvested after 210 DAP in this study was comparable to the above-ground herbage ha rvested at 140 DAP reported by Agyemang et al. (2000). The NDF concentrations of top canopy herbage were lower than those reported for above-ground, whole-plant he rbage (Mupangwa et al., 2000; Agyemang et al., 2000; Adjorlolo et al., 2001). The higher NDF concentration in lablab was expected because lablab had lower leaf:stem ratio than mucuna (Table 6-1).

PAGE 174

159 Table 6-1. Herbage dry matter mass and plantpart proportions of various fractions of defoliated mucuna and lablab relay cropped in maize for 2 yr. Treatment Fraction Mucuna Lablab SE Legume effect Top canopy (TC) --------------t ha-1 ---------------P value --TC-Leaf 0.72 0.87 0.08 0.461 TC-Stem 0.29 0.94 0.13 0.023 TC-Leaf + stem 1.01 1.81 0.20 0.096 TC-Leaf:stem 2.94 1.01 0.34 0.001 Lower Canopy (LC) LC-Leaf 0.48 0.08 0.09 0.021 LC-Stem 0.58 0.50 0.08 0.389 LC-Leaf + stem 1.05 0.58 0.16 0.064 LC-Leaf:stem 0.84 0.17 0.11 0.024 Total 2.06 2.39 0.27 0.485 TC:Total 0.52 0.76 0.04 0.011

PAGE 175

160 Table 6-2. Nutritive value of top-canopy biom ass of mucuna and lablab relay cropped in maize during the 1999/2000 growing season at NARC-Kitale. Treatment Fraction Mucuna Lablab SE Legume effect Leaf --------------g kg-1 --------------P value -CP 145 152 4.4 0.461 Ash 65 90 5.7 < 0.001 NDF 337 365 15.3 0.582 IVDMD 613 764 34.3 0.003 Stem CP 115 86 6.7 0.022 Ash 49 88 8.8 0.001 NDF 419 517 24.4 0.006 IVDMD 601 649 46.1 0.681 Leaf + stem CP 130 111 9.4 0.281 Ash 68 86.3 4.3 0.002 NDF 325 446 29.3 0.003 IVDMD 617 693 20.1 0.110 CP = Crude protein; NDF = Neutral dete rgent fiber; IVDMD = In vitro dry matter digestibility.

PAGE 176

161 Table 6-3. Crude protein con centration of top-canopy biomass leaf and stem fractions of mucuna and lablab relay cropped in ma ize during the 2000/2001 growing season at NARC, Kitale. Treatment Fraction Mucuna Lablab SE Legume effect --------------g kg-1 ---------------P values -Leaf 246 281 10.6 0.174 Stem 135 111 8.61 0.064 The IVDMD of top canopy biomass of mucuna and lablab in this study was higher compared to mucuna and lablab forage harv ested at maximum biomass yield (16 and 18 WAP, respectively) reported by Maasdorp and Titterton (1997). This could partly be attributed to the higher propor tion of leaves in the top-ca nopy biomass compared to total above-ground biomass (Chapter 4). Although lablab contained greater NDF concentration than mucuna, it tended to ha ve greater IVDMD than mucuna, concurring with the results of Maasdorp and Titterton (1997 ) and Chapter 7. This could be attributed in part to the higher lignin concentration in mucuna biomass compared to lablab (Chapter 7). Previous studies have shown that mucuna ha s an aggressive twining and trailing habit (Singh and Relwani, 1978) that could redu ce the yield of maize when the legume is planted together with maize earlier than 5 wk after planting (Maasdorp and Titterton, 1997; Versteeg et al., 1998a). Our results show that upper canopy herbage obtained after defoliating mucuna and lablab at 10 cm above the ground is of higher nutritive value than

PAGE 177

162 that of whole biomass hay. Thus defoliati on provides a management option that could provide high quality of fodder for livestock while reducing the competitive ability of the legumes. Moreover, the adoption of green manure has been shown to be higher when these legumes provide other uses (Becker et al., 1995; Versteeg et al., 1998a). A complementary study (Chapter 3), found that the reduction of quality of the mucuna stubble residue may enhance the efficiency of N uptake by succeeding maize. Also at the time of residue defoliation and incorporati on, relay-cropped lablab produced seed yield higher than the second crop of beans. Results from this study show that mucuna a nd lablab have potential to provide 1 to 1.8 t ha-1 yr-1 of livestock fodder respectively, when relay cropped in maize and defoliated to 10 cm at 210 DAP. A smallholder farmer keeping one dairy cow would need 0.225 or 0.125 ha yr-1 of relay-cropped mucuna or labl ab, respectively, to produce sufficient herbage for supplementation (2.5 kg cow-1 d-1) for at least the 3-mo dry season when feed scarcity is most severe. Thes e acreages are feasible given th at the same piece of land on which the legumes grow is used for maize and common bean production. In a feeding experiment to evaluate the potential of muc una or lablab hay as protein supplements for dairy cows (Chapter 7), supplementation resulted in an extra 0.41 kg milk cow-1 d-1 and an additional 208 kg cattle manure DM yr-1. The manure could supply 5 kg N, 0.6 kg P, 0.53 kg K, 1.4 kg Mg, and 5.5 kg Ca yr-1 (13 kg of nutrients yr-1). Conclusion The use of forage legumes in smallholder fa rms in the tropics has been generally low (Thomas and Sumberg, 1995). For increased adoption of legume-based forage technologies in smallholder farms, these legumes must fit into the overall farming strategy based on food production for the fam ily household. This study has shown that

PAGE 178

163 mucuna or lablab relay cropped into maize af ter harvesting a first crop of common beans, and defoliated to a 10-cm-stubble, provide fodder that could be a valuable protein supplement during the dry season, which is a major constraint to livestock production. This could reduce the trade-off associated wi th the introduction of green manure legumes for soil fertility improvement al one and increase their adoption.

PAGE 179

164 CHAPTER 7 FEED INTAKE AND LACTATION PERF ORMANCE OF DAIRY COWS OFFERED NAPIERGRASS SUPPLEMEN TED WITH LEGUME HAY Introduction A major constraint to smallholder dairy cattle production in the tropics is the scarcity and poor quality of on-farm f eed resources and the high co st and uncertain supply of purchased concentrates. In the highlands of northwestern Kenya, mixed farming based on high yielding fodder grasses has potential for improving both the quantity and quality of feed available throughout the year. Within these systems, the main feed constraint to dairy cattle production occurs during the dry se ason when the quality of available feed is low. Napiergrass [ Pennisetum purpureum (Schum) cv. Bana] is the most important fodder grown on smallholder farms practici ng intensive mixed crop-livestock production in northwestern Kenya (Nyambati, 1997). Napi ergrass, which can tolerate mild drought (Skerman and Riveros, 1990), is low in N a nd digestible nutrients when mature, reducing the efficiency with which it is utilized for milk production (Muia et al., 2000b). Previous research has shown that diet crude protein (CP) concentration of 60 to 80 g kg-1 dry matter (DM) is the minimum range for optimum rumen microbial activity (Minson and Milford, 1967). Working on smallholder farms in Kenya, Wouters (1987) indicated that napiergrass had a mean CP concentration of 76 g kg-1 DM. Later reviews (Shreuder et al., 1993; Muia et al., 2000b) reported a mean CP concentration ranging between 50 and 90 g kg-1 DM confirming that CP in napiergrass may limit milk production.

PAGE 180

165 Protein supplementation (c oncentrates, byproducts, and Leucaena leucocephala as sources) to napiergrass-based diets has been shown to increase feed dry DM intake (DMI) and milk production (Muinga et al., 199 2; 1993; 1995; Muia et al., 2000a). The benefits of using protein-rich forages as supplements include improved energy and protein intake, improved feed efficiency, increas ed availability of minerals and vitamins, improved rumen function, and generally, enha nced animal performance (Norton and Poppi, 1995). In addition to improving the nutritive valu e of livestock feeds, legumes have the potential to improve soil fertility through thei r ability to fix N (Sanginga et al., 1996a) or when incorporated into the soil as green manur es (Tian et al., 2000). Thus the integration of legume-based technologies into intensified farming sy stems has the potential to increase both crop and livestock productivity (Weber, 1996). Although napiergrass provides fodder throughout much of the year, its CP may be inadequate for production needs of lactati ng cows (Anindo and Potter, 1986). Mucuna [ Mucuna pruriens var. utilis (L.) DC (Wright) Burck] and lablab [ Lablab purpureus L. (Sweet) cv. Rongai] are legum es that can serve as both green manures and livestock feeds, but data are limited whic h assess their value as supplem ents to cattle diets. The objective of this study was to evaluate the e ffects of feeding mucuna and lablab hay as supplements to lactating dairy cows fed a basal diet of napiergrass on DMI, diet digestibility, cow body condition, body weight gain, and on the quantity and quality of milk produced.

PAGE 181

166 Materials and Methods Production Environment The experiment was conducted at the Kenya Agricultural Research Institute’s National Agricultural Research Center, Kitale (10 01” N, 350 00’’ E) in agroeco logical zone upper midlands 4 as described by J aetzold and Schmidt (1983). Th e center receives an average annual rainfall of 1100 mm and is on a clay loam soil with a pH (2.5:1 water soil ratio) of 5.5. During the period of the feeding experi ment from June to August mean monthly minimum and maximum temp eratures were 12 and 24.40C, respectively. The study was started at the beginning of June 2000 and lasted 12 wk. Experimental Diets The tall type of napiergrass was planted at NARC, Kitale in 1999 for the production of fodder. At planting, 10 kg ha-1 of P was applied as tr iple-super phosphate. The napiergrass was weeded twice and top-dressed with calcium ammonium nitrate fertilizer at the rate of 60 kg N ha-1 after cutting it back to a 5-cm stubble. The napiergrass was cut back in sequential blocks, to ensure unifor mity of maturity duri ng the feeding period. The napiergrass was harvested at a height of approximately 1 to 1.5 m using a machete, leaving a stubble of approximately 5 cm. The harvested material was chopped manually into 2to 4-cm pieces using a forage chopper and was fed fresh. Mucuna and lablab were grown in 1999 at NARC, Kitale prior to the start of the feeding experiment. One hectare each of mucu na and lablab were es tablished at an interrow spacing of 50 cm and within row spacing of 25 cm. They were weeded three times before flowering, and spot hand weeding was done just before harvesting. The legumes were harvested manually at the 50% flowering stage to a stubble height of 5 to 10 cm using hand shears. Lablab was harvested 16 wk after germination while mucuna was

PAGE 182

167 harvested 20 wk after germination because of delayed flowering. The harvested herbage was sun dried for 24 to 48 h before transfe rring it to a drying shade. The material was then spread in raised-wire stands inside a barn and turned several times until it was completely air dry. The material was choppe d into 2to 4-cm pieces using a manually operated forage chopper and stored in a wellventilated room ready for feeding. Dairy meal was purchased in bulk from a local feed miller. Experimental Animals Eight multiparous (with two to five previous lactations) lactating Holstein-Friesian cows (377 + 43 kg) were selected from the dairy he rd at NARC, Kitale, based on days in milk, lactation number, milk yield, and body we ight to minimize any carry over effects. The cows (261 + 23 days in milk) were grazing on low quality pastures prior to the start of the experiment and their average milk yield was 3.7 kg d-1. The cows were treated with 10 mL of 34% nitroxynil (Trodax)1 and drenched with 100 mL of 3% Oxyclozanide/1.5% Levamisole hydrochlor ide/0.385% Cobalt sulphate (Nilsan)2 3 wk prior to the start of the ex periment against endoparasites and sprayed weekly with a diamidide acaricide (Triatix)3 against ectoparasites. The cows were also injected with 10 mL of a multivitamin4 supplement at the beginning of the experiment, composition of which is given in Appendix B. The cows we re housed in an open-draft, free-stall barn, partitioned to facilitate individual feeding. Each stall was equipped with two feed 1 Rhone Merieux, Lyon-France 2 Cooper Kenya Ltd 3 Welcome Ltd, Nairobi Kenya 4 Alfasan, Woerden-Holland

PAGE 183

168 troughs and a water trough, and cows were able to comfortably move around. The stalls had concrete floors, and no bedding was provided; they we re cleaned every morning. Experimental Design The eight cows were divided into two groups each having similar milk yield based on milk production observed prior to commencem ent of the experiment. The two groups were used in a 4 x 4 cross-over design (C ochran and Cox, 1957). The change-over design was chosen to measure the direct effects of cows at post-peak l actation and to reduce variation among cows (Strickland, 1975). The four periods were 21 d each (14-d adjustment period and 7-d collection period). The treatments comprised a basal diet of fresh-cut napiergrass either fed alone or s upplemented with legume hay or a commercial dairy meal. The treatments were 1) napiergr ass basal diet fed alone (NG), 2) napiergrass supplemented with mucuna hay (MH), 3) na piergrass supplemented with lablab hay (LH), and 4) napiergrass supplemented with co mmercial dairy meal (CDM). Each of the four diets was allotted at random to one cow within a group. Diets and Feeding Management Napiergrass was used as basal diet becau se it is the major high yielding fodder available on smallholder croplivestock production farms in western Kenya (Nyambati, 1997). The supplements were offered at isonitrogenous levels calculated to meet the requirement of a 350-kg cow producing 8 to 10 kg d-1 of 4% fat corrected milk (FCM) and with DMI of 2.7% of live weight (9.45 kg d-1, necessary to fulfill nutrient requirement for maintenance, milk produc tion, and normal live weight gain) (NRC, 1989). The digestible energy requirement will be met (calculation not shown) from the diets based on the estimation of DE (Mcal kg-1 DM) from NRC (1989) for napiergrass, and from the equation of Heaney and Pigden (1963) for mucuna and lablab. The basal

PAGE 184

169 diet was fed ad libitum in two equal batches at 0800 h and 1600 h to ensure 10% refusal. The supplement portion was fed in a separate trough once at 0800 h every morning. Before morning feeding, refusals from each treatment were removed and weighed. A complete mineral mixture,5 at a level of 170 g head-1 d-1 was provided to ensure an adequate supply of all minerals. The an imals had ad libitum access to clean drinking water. Measurements Feed intake, body weight changes, milk yield and composition, and total fecal output were measured. Orts were weighed every morn ing and feed intake recorded. Samples of feeds offered and refused were taken every mo rning in the last 7 d of every period. Body weight (BW) change of cows during each expe rimental period was determined using an Avery (Birmingham, England) weighing scal e of 1000 x 1 kg precisi on attached to a weighing crate. Each estimate of BW was th e mean of two weight records taken on two consecutive days. The animals were weighe d prior to the morning feeding. The cows were milked twice daily at 0630 and 1430 h and individual cow milk yield was determined using a Waymaster (Precision Weighers, Reading, England) milk scale of 25 x 0.05 kg precision. Milk production is expr essed as 4% FCM yield. Body condition changes were measured using a body condition score ranging from 1 to 9 (Kunkle et al., 1994). The scoring was done by three evaluators in the morning before feeding at the beginning and end of every collection period. Sc ores were used from the two evaluators that were closest to each other. During the la st 5 d of the collection period the total daily 5Approximate elemental composition (% of DM): Ca = 15.2, P = 6.5, Na = 11.0, Chlorine = 17.0, Mg = 1.5, Fe = 0.4, Cu = 0.14, Mn = 0.2, Zn = 0.30, S = 0.2, Co = 0.016, I = 0.01, K = 0.003

PAGE 185

170 fecal output of the cows was collected by st ationing an attendant with buckets by each animal (Abdulrazak et al., 1997). Feces from every animal were weighed in the morning before feeding. In vivo apparent dry matter digestibility of diets was computed using feed intake and total feces (Osuji et al., 1993). Sample Preparation and Chemical Analysis Milk samples were taken on Tuesday and T hursday morning and afternoon during the collection period. Morning and afternoon milk samples were composited by cow for each day of sampling. Two subsamples were derived from the composite sample for each day. One subsample was analyzed for butterfat, density, solids-not-fat, and total solids. The other subsample was kept under re frigeration and later analyzed for CP. The Gerber method (Appendix C; Pearson, 1976) was used to determine milk butter fat on the same day the milk samples were taken. To determine total solids, the density of milk was measured by a lactometer calibrated at 200C. A correction for density of 0.24 for every 10C above 200C was made. The total solids were calculated using Fleischmann’s formula, total solids = 1.2F + 2.665 (100D )/D, where F = % butter fat, and D = density (Pearson, 1976). The solids-not-fat wa s derived by difference (total solids-butter fat). Samples of feeds offered and refuse d were weighed and dr ied in an oven at 600C to constant weight for about 48 h and weighed again. Dry samples were bulked per cow and per treatment in each period before be ing subsampled for grinding. The samples were ground in a Christy Norris hammer mill to pass 1-mm screen and stored in airtight plastic bags. Feed samples were analyzed fo r CP, neutral detergent fiber (NDF), in vitro dry matter digestibility (IVDMD), Ca, P, li gnin, and total soluble polyphenol. Each morning a sample of feces (approximately 300 g) from each cow was taken after mixing.

PAGE 186

171 The sample was weighed and dried at 650C to constant weight in forced air oven. After drying, it was weighed again to determine the DM. Total N, P, and Ca of feed and milk samples were determined following the procedures outlined by the Association of Offi cial Analytical Chemists (AOAC, 1990). The NDF and lignin were determined by th e method of Goering and Van Soest (1970) modified by Van Soest et al . (1991). In vitro dry matter digestibility was determined using the procedure of Tilley and Terry (1963). Total extractable polyphenols were determined using the Folin-Denis method as outlined by Anderson and Ingram (1993). The DE concentration was estimated from th e regression equation, DE concentration = – 0.559 + 0.056X; R2 =0 97 and SE = 0.083 (Heaney and Pigden, 1963) where X = digestible DM in g 100 g-1 DM. Statistical Analysis Analysis of variance using the general li near model procedure of SAS (SAS, 2001) was used to determine the effects of treatment diets on intake, apparent digestibility, milk yield and composition, body condi tion score, and BW changes. The analysis was done by the combined analysis of two Latin s quares according to Cochran and Cox (1957). The statistical model was ijkn = + i + j + k + n + ijkn,, where ijkn = observed value; = overall mean; i = effect of ith group (i = 1, 2); j = effect of jth period (j = 1, 2, 3, 4); k = effect of kth cow (k = 1, 2, 3, 4); n = effect of nth treatment (n = 1, 2, 3, 4); and ijkn = experimental error. Preplanned treatment comparisons were made using single degree of freedom contrasts (Montgomery, 1997). The comparis ons were NG vs. MH + LH, MH vs. LH,

PAGE 187

172 and MH + LH vs. CDM. The separation of means and pre-planned contrasts were done using the general linear mode l procedure in SAS (2001). Results Chemical Composition of Feeds The chemical composition of the feeds is shown in Table 7-1. Napiergrass CP concentration was 68 g kg-1 compared to over 160 g kg-1 for the supplements. Lignin and polyphenol concentrations were 114 and 23.7 g kg-1 for mucuna and 86 and 18.8 g kg-1 for lablab. The DE concentr ation ranged from 2.68 Kcal kg-1 DM for mucuna to 2.81 and 2.95 Kcal kg-1 DM for napiergrass and labl ab, respectively, to 3.76 Kcal kg-1 DM for dairy meal. Intake, Fecal Output, an d Apparent Digestibility There was no difference in legume hay DMI between MH and LH treatments (Table 7-2). The cows provided CDM ate all daily supplement offered while those provided legume hay did not (68 and 78% of hay offered was consumed for MH and LH, respectively). Supplementation did not affect (P>0.568) DMI of napi ergrass (Table 7-2), however, feeding MH, LH, and CDM increased (P<0.001) total dry matter intake (TDMI) by 21, 26, and 37% over the control, respectiv ely, and TDMI expressed per unit of metabolic body weight (TDMI g kg-1 BW0.75) by 20, 25, and 31%, respectively. Dairy meal supplementation had a greater positiv e effect on TDMI (P = 0.003) and TDMI expressed as g kg-1 BW0.75 (P = 0.005) than did legume hay supplementation, but there was no difference (P = 0. 173) between MH and LH.

PAGE 188

173 Table 7-1. Chemical composition of feeds fed to experimental cows. Feeds Constituents Napiergrass Mucuna Hay Lablab Hay Dairy Meal DM (g kg -1) 176 883 884 914 CP (g kg DM) 68.4 170 163 179 NDF (g kg DM) 712 564 565 485 Lignin (g kg DM) 44.2 114 85.8 38.3 Total polyphenol (g kg-1 DM) Nd 23.7 18.8 13.2 P (g kg DM) 1.5 2.4 2.8 7.5 Ca (g kg -1 DM) 3.2 14.7 14.7 10.5 IVDMD (g kg DM) 602 578 626 772 IVOMD (g kg OM) 579 529 591 756 DE (Kcal kg-1 DM) 2.81 2.68 2.95 3.76 DM = dry matter; CP = crude protein; NDF = neutral detergent fiber; IVDMD = in vitro dry matter digestibility; IVOMD = in vi tro organic matter digestibility; DE = Digestible energy estimated from the equation, DE = .559 + 0.056X; R2 = 0.97 and SE = 0.083 (Heaney and Pigden, 1963) where X = digestible DM in g 100 g-1 DM. Nd = not determined Tannic acid equivalent

PAGE 189

174Table 7-2. Intake, fecal output, a nd apparent digestibility when Friesian cows were fed napiergrass alone or supplemented with legume hay or dairy meal. Treatments Contrasts Constituents Napiergrass alone (NG) Napier + mucuna (MH) Napier + lablab (LH) Napier + dairy meal (CDM) SE MH + LH vs NG MH vs LH MH + LH vs CDM ---------------P va lue-------------NDMI (kg) 9.80 9.72 10.1 10.5 0.300.807 0.296 0.377 SDMI (kg) 2.06 2.20 2.90 0.07nd 0.891 0.011 TDMI (kg) 9.80 11.8 12.3 13.4 0.31< 0.001 0.173 0.003 NDMI (g kg-1 BW0.75) 108 107 111 110 3.260.823 0.334 0.790 TDMI (g kg-1 BW0.75) 108 130 136 142 3.27< 0.001 0.065 0.005 DMFO (kg) 3.96 4.60 4.46 4.87 0.08< 0.001 0.246 0.001 Fecal N (g kg-1 DM) 17.8 17.6 18.8 20.2 0.590.798 0.882 0.116 ADMD (g kg-1) 558 587 625 615 13.20.004 0.049 0.589 NDMI = napiergrass dry matter intake; SDMI = supplement dry ma tter intake; TDMI = total dry matter intake; BW = body weight; DMFO = dry matter fecal output; ADMD = apparent dry matter digestibility; nd = not determined

PAGE 190

175 Effect of supplementation on fecal output follo wed a similar trend to that for TDMI. Mucuna hay, LH, and CDM supplementation increased fecal output by 16, 13, and 23%, respectively, compared to napiergrass alone. However, supplementation had no effect (P = 0.384) on fecal N concentration. Compared to the control, supplementation increased apparent digestibility by 29, 67, and 57 g kg-1 for MH, LH, and CDM treatments, respectively. There was no difference (P= 0.589) in in vivo digestibility of the DM between the average of the legume hay treatm ents and CDM (Table 7-2), however cows fed LH had a greater apparent DM digestibi lity than cows fed MH (P=0.049). Effect of supplementation on DE intake followed a simila r trend to that of in vivo digestibility (Table 7-3). Legume hay supplementation did not affect na piergrass CP intake (P=0.759), but it increased (P=0.001) total CP intake (Table 7-3) over the control. There was no difference (P=0.680) in total CP intake betw een mucunaand lablab-supplemented cows, but the total CP intake of CDM-supplemented cows was higher than those of the legume hay-supplemented cows. Intake of DE follo wed trends similar to those for total CP intake except that cows fed LH diets had gr eater DE intake than cows fed MH diets. Milk Yield Cows supplemented with MH, LH, or CDM produced more raw milk and 4% FCM (P< 0.004) than those fed only napiergrass (Table 7-4). There was no difference (P>0.05) in milk yield between MH and LH treatment s. Cows supplemented with dairy meal produced more (P<0.001) raw milk and FC M (25%) than those supplemented with legume hay.

PAGE 191

176Table 7-3. Crude protein intake (CPI) and digestible energy intake (DEI) of Friesian cows fed a basal diet of napiergrass alon e or supplemented with legume hay or dairy meal. Treatments Contrasts Constituents Napiergrass alone (NG) Napier + mucuna (MH) Napier + lablab (LH) Napier + dairy meal (CDM) SE MH + LH vs NG MH vs LH MH + LH vs CDM -----------------P value--------------NCPI (g d-1) 695 680 757 791 58.0 0.759 0.770 0.418 SCPI (g d-1) 343 367 522 37.2 nd 0.943 0.003 TCPI (g d-1) 695 1060 1090 1310 57.9 0.001 0.680 0.02 TCPI (g kg-1BW0.75 d-1) 7.7 11.7 b 12.0 13.8 0.63 0.001 0.759 0.055 DEI (Kcal d-1) 25.2 32.1 37.4 39.0 1.76 < 0.001 0.033 0.053 NCPI = Napiergrass crude protein intake , SCPI = Supplement crude protein intake, TCPI = Total crude protein intake, DE = Digestible energy estimated from the equation DE = .559 + 0.056X; R2 = 0.97 and SE = 0.083 (Heaney and Pigden, 1963) where X = digestible DM in g 100 g-1 DM, nd = not determined

PAGE 192

177 Table 7-4. Milk production, body condition score (BCS), and body wei ght gain (BW Gain) of Friesian cows fed napiergrass alone or supplemented with legume hay or dairy meal. Treatments Contrasts Constituents Napiergrass alone (NG) Napier + mucuna (MH) Napier + lablab (LH) Napier + dairy meal (CDM) SE MH + LH VS NG MH vs LH MH + LH vs CDM ------------------P value---------------Milk (kg cow-1) 3.32 3.75 3.80 4.72 0.05 0.004 0.868 < 0.001 FCM (kg cow-1) 3.62 4.22 4.18 5.27 0.06 < 0.001 0.778 < 0.001 Mean BCS 4.6 5.2 5.0 4.9 0.13 0.013 0.254 0.299 BW Gain (kg/7 d) 5 10 12 17 4.89 0.304 0.804 0.347 FCM = 4% Fat corrected milk [FCM = (M ilk yield x 0.4) + (Milk fat yield x 15)].

PAGE 193

178 Milk Composition Legume hay supplementation had no effect (P<0.05) on milk composition (Table 7-5). Yield of milk fat and CP per cow were grea ter for legume hay-supplemented than control animals (Table 7-5) because of higher milk production of supplemented cows. Milk from cows supplemented with LH had greater (P<0.002) P concentration than those supplemented with MH. Dairy meal supplemen tation produced milk that was greater in solids-not-fat (P<0.001), Ca (P=0.046), and P (P =0.008) concentration th an that of cows supplemented with the legume hays. Live-weight Changes and Body Condition Score Supplemented animals tended to gain more weight (5, 7, and 12 kg for MH, LH, and CDM, respectively) than those fed napiergr ass alone, but these differences were not significant (Table 7-4). The mean body c ondition score of legume hay-supplemented animals was greater than that of cows fed napiergrass alone. Discussion Chemical Composition of Feeds The nutritive value of napiergrass was relative ly low (Table 7-1) but it was within the ranges reported by Muinga et al. (1995), Kariuki et al. (1999), and Muia et al. (1999). In this study, the growth period for napiergrass wa s longer than originally anticipated due to fewer showers than normal during the dry s eason and the later than usual onset of the rainy season. The longer growth period lik ely resulted in lower napiergrass nutritive value. This is a relatively typical occurrence on farms, so the napiergrass used in the trial is considered representative of that fed in the region during the dry season.

PAGE 194

179Table 7-5. Milk composition of Friesian cows fed napiergrass alone or supplem ented with legume hay or dairy meal. MF = Milk fat, TS = Total solids, SNF = Solids not fat, MCP = Milk crude protein Treatments Contrasts Constituents Napiergrass alone (NG) Napier + mucuna (MH) Napier + lablab (LH) Napier + dairy meal (CDM) SE MH + LH vs NG MH vs LH MH + LH vs CDM ---------------P value----------------MF (g kg-1) 46.0 48.4 46.7 47.8 0.37 0.328 0.350 0.888 MF (kg cow-1) 0.15 0.18 0.18 0.22 0.01 < 0.001 0.565 < 0.001 TS (g kg-1) 132 134 132 136 1.53 0.589 0.274 0.169 SNF (g kg-1) 86.0 85.9 85.1 87.9 0.51 0.414 0.330 < 0.001 MCP (g kg-1) 34.8 33.8 34.8 33.7 0.58 0.508 0.224 0.418 MCP (kg cow-1) 0.13 0.14 0.14 0.17 0.01 0.008 0.853 < 0.001 Ca (g kg-1) 11.1 11.1 10.8 11.3 0.13 0.431 0.078 0.046 P (g kg-1) 7.7 7.6 7.9 7.9 0.06 0.739 0.002 0.008

PAGE 195

180 Chemical composition of the legume hays wa s similar except for lignin concentration which was greater for mucuna. This may par tially explain th e greater IVDMD for lablab. In addition these responses are likely due to the 4-wk longer regrowth period for mucuna, associated with its delayed flowering. Lablab CP was similar to that reported in Nigeria when hay was harvested at a similar number of days after planting (Agyemang et al., 2000). Mucuna, harvested at th e same flowering stage in Sri Lanka (Ravindran, 1988), had higher CP (206 g kg-1) than in the current study (170 g kg-1), likely due to the 50-d longer growth period required to reach 50% flowering in northwestern Kenya. Apparent Digestibility and Intake Apparent DMD was greater for the cows fed the supplement treatments than those fed the napiergrass control diet and was greater for cows fed LH than for those fed MH. Apparent DMD of napiergrass was 44 g kg-1 less than IVDMD, perhaps indicative of a CP deficiency of cattle fed the control diet , whereas apparent DMD of MH and LH diets were similar to IVDMD of mucuna and lablab hays. An increase in apparent DMD can be caused by longer retention times in the rumen associated with a decrease in intake. In this case, feeding legume hays increased a pparent DMD while also increasing intake, supporting the argument that apparent DMD of the napiergrass control diet was limited by availability of CP to rumen microbes. It seems likely that the additional N from the legumes increased activity of the microbial population and increased the rate of digesta breakdown (Abule et al., 1995) to such an exte nt that even shorter retention time did not result in lower apparent DMD. The feeding value of forage depends not only on the digestibility and chemical composition of the feed but also on the quant ity voluntarily eaten. Of the variation in digestible DMI or DEI, 60 to 90% is thought to be due to differences in intake while 10

PAGE 196

181 to 40% is due to differences in digestibil ity (Mertens, 1994). Inta ke can be expressed simply as the quantity of DM or digestible nut rients consumed, or to minimize the effects of body weight differences, intake can be e xpressed per unit of me tabolic body weight. Napiergrass DMI was lower than previously reported (Muinga et al., 1995; and Muia et al., 2000a), likely due to its maturity (Tab le 7-2). There was no observed substitution (i.e., negative associative effects on intake) of supplement for napiergrass in the diet, therefore feeding 2 to 3 kg DM d-1 of the three supplements served to increase total DMI. Similar increases in total DM I resulting from legume fodder supplementation to roughage diets have been reported by other workers (Muinga et al., 1992; 1995; and Muia et al., 2000b). Increased total DMI due to hay and dairy meal supplementation in the current study seems to be linked at least in part to N nutrition. Napiergrass CP concentration (68 g kg-1) was lower than the requirement for main tenance of many classes of animals and was very likely deficient for lactating cows in this study (NRC, 1989). The occurance of CP deficiency was supported by the non-significant supplemen tal effect on fecal N. Assuming rumen degradability of CP in napi ergrass of about 0.8, the rumen degradable protein concentration in th e napiergrass control would be approximately 55 g kg-1 and the ratio of digestible energy (2.61 Kcal kg-1 DM) to rumen degradable protein of 1:21 would be well below 1:28 that is recommended by NRC (1989). The supplements provided additional CP, minimizing a deficiency which likely was limiting growth and cellulose digestion by rumen microbes (Poppi and Norton, 1995). The supplemental CP likely increased the rate of digesta breakdown, passage rate, and intake (Poppi and Norton, 1995; Nsahlai et al., 1998). These data s upport the hypothesis that rumen degradable

PAGE 197

182 protein concentration was limiti ng in the control diet and pr ovide evidence that protein supplementation of cattle fed napiergrass may of ten be needed in this production system. Cattle ate all of the dairy meal offered resu lting in higher intake of the CDM treatment than of MH and LH. Consumption of mucuna and lablab hays was 68 and 78% of that offered, respectively, and this resulted in lo wer intake of legume hay supplement and total CPI compared to the CDM treatment (Tab le 7-3). Greater CPI may have increased DMI of cattle on the CDM treatment, but it is not possible to sepa rate the effects of additional CP and DE intake from the supplement. Refusal of legume hays likely was due in part to loss of leaf during hay harves t and leaf shattering dur ing processing. Orts had a relatively high proportion of legume st em (not quantified) compared to the hay offered. Low palatability of the stem fr action or perhaps some other plant factor affecting palatability may have limited inta ke of legume hays. Mucuna hay had greater concentrations of lignin and lower concentrations of IVDM D than lablab hay and these factors contributed to lower apparent DMD of the MH diet . There also was a trend (P=0.173) toward lower DMI of the MH vs. LH diet, and DEI was lower for MH than for LH. Milk Yield and Composition Supplementation increased yield of raw a nd 4% FCM. The increase in 4% FCM relative to the contro l was 17, 15, and 46% for cows fed the MH, LH, and CDM diets, respectively. Milk producti on trends generally followed those of DMI and CPI, being greatest for the CDM treatment, intermediate and similar for the legume hay treatments, and least for the control. Digestible energy intake was greater for cows fed LH than MH, but this did not result in great er milk production for cows fed the LH treatment. The CP and DE intakes of supplemented cows met the requirements of a 350-kg Friesian cow

PAGE 198

183 producing 10 kg of milk (NRC, 1989). All co ws gained weight during the experiment, thus the increase in milk yiel d of cattle fed supplements can be attributed to greater DE intake and CPI rather than mobilization of body tissues. Supplemented cows also had higher body condition scores than control co ws. Increases in milk yield due to supplementation were similar to those reporte d by Muinga et al. (1992; 1993; 1995) and Muia et al. (2000a). The cows produced on average 0.23 and 0.52 kg of additional milk (above the control) per kg of legume hay and dairy meal fed, respectively, and these data are comparable to the mean value of 0.34 reported by Combellas et al. (1979) across several experiments. Milk fat, solids-not-fat, and protein concen trations of milk are at a maximum at the start of lactation, fall to a mi nimum for solids-not-fat and prot ein after 6 wk and for milk fat about 10 wk after parturition, and finally increase until the end of lactation (Pearson, 1976). Cows used in this experiment were in the latter part of lactation, and concentrations of these components were gr eater than those reported by Muia et al. (2000a) who used cows in early lactation. Although yield of fat and protein was increased by supplementation, their concentr ations in milk were not affected by treatment. These results are consistent with those of Muinga et al. (1992) who reported that when dairy cows were fed a basal diet of napiergrass supple mented with leucaena forage there was no effect of supplement on milk composition. In the current study, solids-not-fat, Ca, and P concentrations were greater for cows fed the CDM than for those fed MH and LH, but these differences were small. Conclusions Mucuna and lablab hay supplementation did not affect napiergrass intake of lactating cows fed a napiergrass basal diet but did in crease total diet DMI, apparent DMD, DE

PAGE 199

184 intake, and milk production. Concentration of ruminally degradable CP in napiergrass was likely limiting microbial growth and fo rage digestion, and th ese data support the hypothesis that greater intake of CP from feeding the legume hays was important in increasing digestibility, intake, and animal performance. Both mucuna and lablab are being tested for use as soil improving, green manures in smallholder, mixed-farming systems in wester n Kenya. These farmers also need feed for their livestock, and our results show that these legumes can provide good quality hay that can be used as a supplement to relatively low quality basal diets such as mature napiergrass. Despite widespread specul ation that L-DOPA [3 (3, 4-dihydroxyphenyl) alanine] in mucuna may be de trimental to livestock, results from this study do not support that conclusion, at least in diets where muc una constitutes only about 20% or less of the dietary DM. Incorporation of these legumes into farming systems could improve milk yield and also result in increased maize yield following the relay cropping of legume green manures with previous maize (Chapters 4 and 5).

PAGE 200

185 CHAPTER 8 CONCLUSIONS, SYNTHESIS, AND RECOMMENDATIONS The overall objectives of this study were 1) to assess the productiv ity of relay-cropped legume green manures and their effect on maize ( Zea may L.)-bean ( Phaseolus vulgaris L.) intercrop performance, 2) to gain great er understanding of th e N-release from soilincorporated legume residue and its uptake by succeeding maize, 3) and to determine the effect of harvesting top canopy biomass of legumes as fodder fo r livestock on their contributions to soil enhancement and livesto ck nutrition. The overa ll study is presented in five parts as found in Chapte rs 3 through 7 of the dissertation. Chapter 3 reports on the N contri bution of relay-cropped mucuna [ Mucuna pruriens var. Utilis (L) DC (Wright) Burck] and lablab [ Lablab purpureus L. (Sweet) cv. Rongai] to the succeeding maize-bean intercrop when part of the legume biomass was harvested for fodder. Mucuna had a higher leaf-to-stem ratio (1.43) than lablab (0.82). Defoliation to a 10-cm stubble reduced the high N-containing leaf frac tion in mucuna to a lesser extent (40%) than lablab (80 %). The effect of legume residue incorporation on maize growth was greatest between 8 and 21 wk after planting (WAP) when maize biomass yield was higher in legume treatments than the natural fallow. The greater maize biomass under legume treatment was a result of N uptake by maize, which was greater between 8 and 10 WAP in the legume residue plots than the natural fallow control. Maize growth and N uptake was highest under undefoliated lablab (UD-L), and defoliation of lablab resulted in reduced performance while for mucuna it had no effect. At 8 WAP, when the demand for N by mai ze was high, the N recovery was greatest

PAGE 201

186 under UD-L (27%). The mean recovery across all time periods was in the range of 21% for cattle manure to 52% for UD-L. The N r ecovery for defoliated mucuna (D-M) plots averaged 45% compared to 35% for undefolia ted mucuna (UD-M). Maize grain yield followed similar trends as N uptake, and was in the order of inorganic N = UD-L > cattle manure = D-M > UD-M = defoliated lablab (D-L) > natural fallow. In Chapter 4 the effect of legume cropping system, defoliation regime, and cropping sequence (number of years of residue application) on subseq uent maize and beans were evaluated on the resear ch station. Relay-cropped mucuna and lablab survived the dry season producing a total biomass yield of 4 and 2.7 t ha-1, respectively, and contributing 78 and 57 kg N ha-1. After 1 yr of residue application, UD-L yielded more (6.5 t ha-1) maize grain than UD-M (5.2 t ha-1), and defoliation reduced gr ain yield following lablab, but not following mucuna. Legume residue treatments yielded more (5.6 t ha-1) than the natural fallow control (4.4 t ha-1). Both after one and two consecutive years of residue application, mucuna plots yielded higher (510 kg ha-1) bean grain than lablab (372 kg ha-1), and mean bean grain yield on plots af ter 1 yr of residue application (437 kg ha-1) were not different from those where residue had been applied for two consecutive years (446). Maize grain yield after 2 yr of resi due application showed that D-M outyielded DL, whereas UD-L outyielded UDM, suggesting that the re sidual effects were higher under UD-L and D-M treatments. Defoliation of mucuna resulted in grain yield increase after 2 yr of residue applicat ion, but defoliation of lablab re duced grain yield. Generally, there was no apparent advantage of applyi ng residue for two consecutive years (5.3 t ha-1) compared to 1 yr (5.1 t ha-1), but UD-L after 2 yr (6.7 t ha-1) yielded higher maize grain than after 1 yr (5.6 t ha-1). Across the 2 yr, legume resi due treatments resulted in higher

PAGE 202

187 yields (5.2 t ha-1) than the natural fallow control ( 4.1). Maize grain yields from plots where the residue was applied the previous year but not in the current year, showed that the treatments with lower quality residue (D-M , UD-L, and D-L) resulted in the highest residual effects, and the yields under these treatments were higher than the natural fallow that had not received any N the previous year. The results from Chapter 4 and those repor ted in Chapter 3 showed that mucuna residue had a greater proportion of leaf fraction in total residue biomass, whereas lablab had a greater proportion of low quality stem, suggesting that la blab residue was likely of lower quality than mucuna. Defoliation of mu cuna resulted in residue of intermediate quality due to the reduced proportion of high quality leaves. Defoliation of lablab, which has an upright growth morphology, removed a greater proportion of the above-ground biomass (0.76) than for mucuna (0.52). After two years of consecutive residue application, D-M and UD-L treatments resu lted in subsequent maize grain yield increases, suggesting that the efficiency of N uptake by subsequent maize and residual effects were higher under residue s of intermediate than hi gh quality. Our hypothesis was that the incorporation of intermediate quality (low N and P, high lignin) residue at the onset of rain extends the time period of nutrient availability to succeeding maize, resulting in increased efficiency of N recove ry due to reduced losses. The results show that in the sub-humid highlands of nor thwestern Kenya, the use of green manure technology could be enhanced by relay cropping the legumes in maize and removing part of the biomass for livestock feed before incorporating the remaining intermediate quality stubble and roots to succeed ing maize-bean intercrop.

PAGE 203

188 Chapter 5 describes the evaluation of rela y-cropped mucuna and lablab on low soil fertility, smallholder farms in the sub-hum id highlands of northwestern Kenya. On farmers’ fields where nutrient depletion is mo re severe than on the experiment station, mucuna yielded more biomass (2.3 t ha-1) than lablab (0.75) and contributed more nutrients. This resulted in greater maize growth and grain yield after mucuna residue application than after lablab. Green manure treatments (UD-M, D-M, and UD-L) resulted in 97, 54, and 41%, respectively, grea ter grain yields than the natural fallow control where no N was applied. This study re vealed that when bi omass yield of green manures are very low the positive effects of de foliation on residue quality did not result in better maize performance. Stability analys is of maize grain yiel d showed that on farms where the legume herbage yields were higher, the defoliation of mucuna did not reduce the grain yield achieved by the undefoliated mucuna, suggesting that farmers could utilize the top-canopy biomass for fodder when the legume biomass yields are higher. The plant-part proportions and nutritive va lue of top canopy herbage of mucuna and lablab defoliated to 10-cm stubble are repor ted in Chapter 6. The results show that defoliation at 10 cm above ground provides an average of 1 to and 1.8 t ha-1 of mucuna and lablab fodder, respectively, th at is of high CP (130 and 111 g kg-1, respectively) and digestibility (617 and 693 g kg-1, respectively). These yiel ds were 52 and 76% of the above-ground leaf + stem herbage from mucuna and lablab, respecti vely, suggesting that the upper canopy herbage has pot ential as dry-season protein supplement, but this practice significantly reduces the quantity of nutrients that are soil incorporated. The objective of the feeding experiment repo rted in Chapter 7 was to evaluate the effects of feeding mucuna and lablab hay as supplements to lactating dairy cows fed a

PAGE 204

189 basal diet of napiergrass on dry matter in take and digestibility, body condition and body weight gain, and on the quantity and quality of milk and manure produced. The CP of mucuna and lablab hay (170 and 163 g kg-1, respectively) was comparable to that of commercial dairy meal (179 g kg-1), and their feeding as supplements to low quality napiergrass increased total dr y matter intake (130 and 136 g kg-1 BW0.75, respectively) compared to napiergrass alone (108 g kg-1 BW0.75). Legume hay supplementation also increased apparent dige stibility (587 and 625 g kg-1, respectively) compared to napiergrass (558 g kg-1), and daily yield of 4% fat corrected milk (4.20 and 4.12 kg) compared to napiergrass (3.69 kg). The re sults show that milk production can be increased in smallholder mixed farming sy stems by supplementing a napiergrass-based diet with mucuna or lablab hay, highlighti ng the importance of integrating legumes into the low-input, mixed cropping systems. Based on the studies presented here, the fo llowing recommendati ons can be drawn: 1. Relay cropping mucuna or lablab in the current maize-common bean intercrop has potential to increase grain yields of s ubsequent maize and beans when P is not limiting. 2. The impact of defoliating upper canopy herbag e of legumes as fodder on yields of subsequent maize and beans depend s on the legume growth morphology. Defoliation of lablab, which has upright gr owth habit, decreases grain yields of subsequent maize, but it does not elimin ate its longer term residual effects on soil fertility. Defoliation of mucuna enhances the synchrony of N release with maize demand for the current crop, and appears to increase the longer term residual effects on soil fertility.

PAGE 205

190 3. There is no apparent advantage of inco rporating undefoliated mucuna residue for two consecutive years compared to one year , but application of lablab residue for two consecutive years has cu mulative residual effects. 4. Farmers can skip one year of residue a pplication and still realize some residual benefit of residue app lied the previous year. 5. Feeding part of mucuna or lablab herb age as dry season protein supplement to lactating cows has potential to increase milk yield and cattle manure output. The fodder quality of mucuna is comparable to that of lablab, th erefore harvesting the top-canopy biomass of either mucuna or lablab as fodder can improve the productivity of smallholder low-external input mixed crop-livestock systems by providing benefits in addition to soil fer tility improvement, such as increased milk and manure yields. This could increa se the adoption of the green manure technology. 6. Future research should focus on the effi ciency of N uptake and recovery in relation to different times of residue app lication in relation to onset of rains and planting of maize. Also on-farm evaluati on of mucuna and lablab as supplements for dairy cows feeding on low quality feed s such as maize stover and napiergrass should be undertaken to el ucidate the potential of th ese legumes in providing additional benefits to the smallholder farmers.

PAGE 206

191 APPENDIX A TOTAL MONTHLY RAINFALL AND MEAN MONTHLY TEMPERATURES RECORDED AT NARC–KITALE, KENYA, IN 2000 0 40 80 120 160 200 JFMAMJJASONDRainfall (mm ) 10 12 14 16 18 20 22Temperature 0C Total Rainfall Mean Temperature

PAGE 207

192 APPENDIX B COMPOSITION OF MULTIVITAMINS FED TO EXPERIMENTAL COW INGREDIENT COMPOSITION Vitamin A 15,000 I.U Vitamin D3 1,000 I.U Vitamin E 20 ug Vitamin B1 10 mg Vitamin B2 6.85 mg Vitamin B6 3 mg Nicotinamide 35 mg Vitamin B12 50 mg Dexpanthenol 25 mg

PAGE 208

193 APPENDIX C GERBER METHOD FOR BUTTER-FAT DETERMINATION IN FRESH MILK 1. Put 10 ml sulphuric acid in the butyrometer. 2. Mix the milk and add 11 ml into the but yrometer. Avoid mixing with the acid. 3. Add 1 ml amyl-alcoho l (density 0.815). 4. Fix the stopper, wrap in a cloth and shake while holding the stopper. 5. Turn the butyrometer upside down a few times. 6. Place it in a water bath at 600 – 700C for 10 min. 7. Centrifuge at 1100 rpm for 5 min. 8. Put it back to the water bath for 5 min before reading. 9. Results are expressed as percentage or g butter fat per 100 ml of milk.

PAGE 209

194 LIST OF REFERENCES Abdel-Magid, H.M. 1992. Evaluation of fora ge and nitrogen supplying potential of hyacinth bean in Saudi Arabia. J. Arid Environ. 23:225-228. Abdulrazak, S.A., R.W. Muinga, W. Thorp, and E.R. Orskov. 1997. The effects of supplementation with Gliricidia sepium or Leucaena leucocephala forage on intake, digestion and live-weight gains of Bos taurus x Bos indicus steers offered napiergrass. Anim. Sci. 63:381-388. Abule, E., N.N. Umunna, I.V. Nsahlai, P.O. Osuji, and Y. Alemu. 1995. The effect of supplementing tef ( Eragrostis tef ) straw with graded levels of cowpea ( Vigna unguiculata ) and lablab ( Lablab purpureus ) hays on degradation, rumen particulate passage and intake by crossbre d [Friesian x Boran (zebu)] calves. Liv. Prod. Sci. 44:221-228. Adjorlolo, L.K., K. Amaning-Kwarteng, and F. K. Fianu. 2001. In vivo digestibility and effect of supplemental mucuna forage on treated rice straw degradation. Small Rumin. Res. 41:239-245. Agyemang, K., I. Oikike, J.H. Makun, and S. O. Magaji. 2000. Trade-offs between forage yields and feed quality of Lablab purpureus , and milk yields in relation to planting and harvesting schedules. Exp. Agric. 36:435-451. Akobundu, I.O., U.E. Udensi, and D. Chikoye. 2000. Velvetbean ( Mucuna spp.) suppresses speargrass [ Imperata cylindrica (L.) Raeuschel] and increases maize yield. Int. J. Pest Manage. 46:103-108. Amijee, F., and K.E. Giller. 1998. Environm ental constraints to nodulation and nitrogen fixation of Phaseolus vulgaris L. in Tanzania: I. A survey of soil fertility, root nodulation and multi-loca tional responses to Rhizobium inoculation. Afr. J. Crop Sci. 6:159-169. Anderson, J.M., and J.S.E. Ingram. 1993. Tropical soil biology and fertility: A handbook of methods (2nd ed.). CAB International, Wallingford, UK. Anindo, D.O., and H.L. Potter. 1986. Milk production from napiergrass ( Pennisetum purpureum ) in a zero grazing system. East Afr. Agric. For. J. 52:102-111. Anindo, D.O., and H.L. Potter. 1994. Season al variation in produ ctivity and nutritive value of napiergrass at Muguga, Kenya. East. Afr. Agric. For. J. 59:177.

PAGE 210

195 AOAC (Association of Official Analytical Chemists). 1990. Official methods of analysis (15th Ed.). AOAC, Arlington, VA. Azhar, W.S., O. Van Cleemput, and W. Ve rstraste. 1986. Nitrification mediated nitrogen immobilization in soils. Plant Soil. 94:401-409. Balasubramanian V., and N.K.A. Blaise. 1993. Short season fallow management for sustainable production in Africa. p. 279-293. In J. Ragland, and R. Lal (eds.) Technologies for sustainable agriculture in the tropics. ASA Special publication No. 56, Madison, WI. Baligar, V.C., and O.L. Bennet. 1986. Outlook on fe rtilizer use efficiency in the tropics. Fert. Res. 10:83-86. Barrios, E., R.J. Buresh, and J.I. Spent. 1996. Organic matter in soil particle size and density fractions from maize and legume cropping systems. Soil Biol. Biochem. 28:185-193. Bationo, A., and A.U. Mokwunye. 1991. Role of manures and crop residues in alleviating soil fertility cons traints to crop production: With special reference to Sahelian and Sudanean zones of we st Africa. Fert. Res. 29:117-125. Becker, M., and D.E. Johnson. 1998. The role of legume fallow in intensified upland rice-based cropping systems in West Af rica. Nutr. Cycl. Agroecosys. 6:1-11. Becker, M., J.K. Ladha, and M. Ali. 1995. Green manure technology: Potential, usage, and limitations. A case study for lowland rice. Plant Soil. 174:181-194. Bergersen, F.J., G.L. Turner, R.R. Gault, M.P. Peoples, L. Morthorpe, and J. Brockwell. 1989. Contributions of nitrog en in soybean crop residues to subsequent crops and soils. Aust. J. Agric. Res. 40:763-780. Blair, G.T., and O.W. Boland. 1978. The rele ase of P from plant material added to the soil. Aust. J. Soil Res. 16:101-111. Boonman, J.G. 1993. East Af rica’s grasses and fodders: Their ecology and husbandry. Kluwer Academic Publishers, Dortrecht, Netherlands. Boonman, J.G. 1997. Farmers’ success with tr opical grasses: Crop/ pastures rotation in mixed farming in East Africa. Mini stry of Foreign Affairs, The Hague, Netherlands. Bowen, W.T., J.W. Jones, R.J. Carky, and J.O. Quntana. 1993. Evaluation of the nitrogen sub-model of CERES-maize following legume green-manure incorporation. Agron. J. 85:153-159.

PAGE 211

196 Brumby, P.J. 1986. ILCA and food production in Africa: A brief review. World Anim. Rev. 60:33-37. Buckles, D. 1995. Velvetbean: A new plant with a history. Econ. Bot. 49:13-25. Buresh, R.J., P.C. Smithson, and D.T. Hellu ms. 1997. Builiding so il phosphorus capital in Africa. p. 111-149. In R.J. Buresh, P.A. Sanchez, and F. Calhoun (eds.). Replenishing soil fertility in Africa. SSSA Special Publi cation No. 51. SSSA, Madison, WI. Burle, L.M., A.R. Suhet, M.S. Cravo, W. Bowen, D.R. Bouldin, and D.J. Lathwell. 1992. Legume green manures: dry season su rvival on the effe cts on succeeding maize crops. CRSP Bull. No. 92-04, Raleigh, N.C. Cahn, M.D., D.R. Bouldin, M.S. Cravo, and W.T. Bowen. 1993. Cation and nitrate leaching in an oxisol of the Brazilian Amazon. Agron. J. 85:334-340. Carsky, R.J., B. Oyewole, and G. Tian. 1999. Integrated soil management for the savanna zone of west Africa: Legume ro tation and fertilization. Nutr. Cycl. Agroecosys. 55:95-105. Chambers, R. 1997. Whose Reality Counts? Putting the first last. Intermediate Technology Publications, London, U.K. Chapman, A.L., and R.J.K. Myers. 1987. Nitr ogen contributed by grain legumes to rice grown in rotation on the cunnurra soils of the ord irrigation area, western Australia. Aust. J. Exp. Agric. 27:155-163. Chau, C.F., P.C.K. Cheung, and Y.S. Wong. 1998. Chemical composition of three underutilized legumes seeds grown in China. Food Chem. 61:505-509. Cheson, A. 1997. Plant degradation by ruminant s: parallels with li tter decomposition in soils. p. 47-66. In G. Cadisch, and K.E. Giller (eds.) Driven by nature: Plant litter decomposition. CAB Int., Wallingford, UK. Cheminingwa, G.N., and J.O. Nyabundi. 1994. Effect of proximity between intercropped maize and beans on growth and yield of maize under varying N levels. East Afr. For. J. 59:269-279. Cochran, W.G., and G.E. Cox. 1957. Experimental Designs (2nd edition). John Wiley and Sons, New York. Combellas, J., R.D. Baker, and J. H odgson. 1979. Concentrate supplementation, and herbage intake and milk production of heifers grazing Cenchrus ciliaris . Grass Forage Sci. 34:303-310.

PAGE 212

197 Constantinides, M., and J.H. Fowness. 1994. Nitrogen mineralization from leaves and litter of tropical plants: Re lationships to nitrogen, ligni n and soluble polyphenol concentrations. Soil Biol. Biochem. 26:49-55. Costa, F.J.S.A., D.R. Bouldin, and A.R. S uhet. 1990. Evaluation of N recovery from mucuna placed on the surface or incorporat ed in a Brazilian oxisol. Plant Soil 124:91-96. Coultas, C. L., T.J. Post, J.B. Jones, Jr., and Y.P. Hsieh. 1996. Use of velvet bean to improve soil fertility and weed control in corn production in northern Belize. Commun. Soil. Sci. Plant Anal. 27:2171-2196. D’Mello, J.P.F., and C. Devendra (eds.). 1995. Tropical legumes in animal nutrition. CAB Int., Wallingford, UK. Decker, M.A., A.J. Clark, F.R. Milford, a nd M.S. McIntosh. 1994. Legume cover crop contributions to no-tillage corn production. Agron. J. 86:126-135. Delve, R.J., G. Cadisch, J.C. Tanner, W. Thorpe, P.J. Thorne, and K.E. Giller. 2001. Implications of livestock feeding mana gement on soil fertility in the small farming systems of sub-Saharan Africa. Agric. Ecosyst. Environ. 84:227-243. Drechsel, P., K.G. Steiner, and F. Hagedorn. 1996. A review on the potential of improved fallows and green manure in Rwanda. Agrofor. Syst. 33:109-136. FAO (Food and Agriculture Organization) . 1995. FAO Yearbook 1994 production, Vol. 45. Food and Agriculture Organiza tion of the United Nations, Rome. FAO-UNESCO. 1994. Soil map of the world. Revised legend. Tech. Paper No. 20. FAO/Rome and ISRIC/Wage ningen, the Netherlands. Fernandez-Rivera, S., T.O. Williams, P. Hiernaux, and J.M. Powell. 1995. Fecal excretion by ruminants and manure availa bility for crop production in semi-arid West Africa. p. 149-169. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestock and sust ainable nutrient cycles in mixed-farming systems of sub-Saharan Africa. Volume II. Tech. Papers. Proc. Int. Conf. ILCA, 22-26 Nov. 1993, Addis Ababa, Ethiopia. Fischler, M. 1996. Research on green manures in Uganda: Results from experiments conducted in 1995. Unpublished report subm itted to the Rockefeller Foundation, Nairobi, Kenya. Fischler, M., and C.S. Wortmann. 1999. Gr een manures for maize-bean systems in eastern Uganda: Agronomic performance and farmers’ perceptions. Agrofor. Syst. 47:123-138.

PAGE 213

198 Fischler, M., C.S. Wortmann, and B. Feil. 1999. Crotalaria ( C. ochroleuca G. Don) as green manure in maize-bean cropping systems in Uganda. Field Crops Res. 61:97-107. Frankenberger, W.T., and H.M. Abdelmagi d. 1985. Kinetic parameters of nitrogen mineralization rates of leguminous crops incorporated into soil. Plant Soil 87:257-271. Franzluebbers, K., L.R. Hassner, and S.R. Juo. 1998. Integrated nutrient management for sustained crop production in sub-Sahara n Africa (a review). Soil management CRSP, TropSoils/TAMU Tech. Bull. No. 98-03. Texas A and M University, TX., USA. Fribourg, H.A., J.R. Overton, W.W. McNei u, E.W. Culvahouse, M.J. Montgomery, M. Smith, R.J. Carlisle, and N.W. Robins on. 1984. Evaluation of the potential of hyacinth bean as an annual warm-season forage in the mid-south. Agron. J. 76:905-910. Fujii, Y., T. Shibuya, and T. Yasuda. 1992. Allellopathy of velvet bean: Its discrimination of L-DOPA as a candidate of allellopathic substances. Japan Agric. Res. Q. (JARQ) 25:238-247. Gachengo, C.N., C.A. Palm, B. Jama, and C. Othieno. 1999. Tithonia and senna green manures and inorganic fertilizers as p hosphorus sources for maize in western Kenya. Agrofor. Syst. 44:21-36. Gault, R.R., M.B. Peoples, G.L. Turner, D.M. Lilley, J. Brockwell, and F.J. Bergessen. 1995. Nitrogen fixation by irrigated lucer ne during the first three years after establishment. Aust. J. Agric. Res. 46:1401-1425. Giller, K.E., F. Amijee, S.J. Brodrick, and O.T. Odje. 1998. Environmental constraints to nodulation and nitrogen fixation of Phaseolus vulgaris L. in Tanzania. II. Response to N and P fertili zers and inoculation with Rhizobium . Afr. Crop Sci. J. 6:171-178. Giller, K.E., and G. Cadisch. 1995. Future bene fits from biological nitrogen fixation: An ecological approach to agricu lture. Plant Soil 174:255-277. Giller, K.E., G. Cadisch, C. Ehaliotis, E. Adams, W.D. Sakala, and P.L. Mafongoya. 1997. Building soil nitrogen cap ital in Africa. p. 151-192. In R.J. Buresh, P.A. Sanchez, and F. Calhoun (eds.). Replenishi ng soil fertility in Africa. SSSA Spec. Publ. No. 51. SSSA, Madison, WI.

PAGE 214

199 Giller, K.E., J.F. McDonald, and G. Cadisc h. 1994. Can biological nitrogen fixation sustain agriculture in the tropics? p. 173-191. In J.K. Syers, and D.L. Rimmer (eds.). Soil science and sustainable land management in the tropics. CAB Int., Wallingford, UK. Giller, K.E., J. Ormesher, and F.M. Awah. 1991. Nitrogen transfer from Phaseolus bean to intercropped maize measured using 15N-enrichment and 15N-isotope dilution methods. Soil Biol. Biochem. 23:339-346. Giller, K.E., and K.J. Wilson. 1991. Nitroge n fixation in tropical cropping systems. CAB Int., Wallingford, UK. Glasener, K.M., and C.A. Palm. 1995. A mmonia volatilization fr om tropical legume mulches and green manures and unlimed and limed soil. Plant Soil 177:33-44 Goering, H.K., and P.J. Van Soest. 1970. Fo rage fibre analysis. (Apparatus, reagents, procedures and some applications). USDA Agr. Handb. No. 379, U.S. Dept. of Agr., Washington, D.C., U.S.A. Goldson, J.R. 1977. Napiergrass ( Pennisetum purpureum K. Schum) in East Africa: A review. Pasture Res. Proj ect Tech. Rep. No. 24. Natl. Agric. Res. Stn., Kitale, Kenya. Goyal, S., M.M. Mishra, I.S. Hooda, and R. Singh. 1992. Organic matter-microbial biomass relationships in field experime nts under tropical conditions: Effects of inorganic fertilization an d organic amendments. Soil Biol. Biochem. 24:10811084. Graham, T.W.G., J.H. Wildin, S.J. Wood, a nd G.W. Blight. 1986. Animal Production in Central Queensland from the autumn-winter forage crops: lablab, molopo buffel grass and zulu sorghum. Trop. Grassl. 20:159-165. Hairiah, K., and K.M. Van Noordwijk. 1989. Root distribution of leguminous cover crops in the humid tropics and effect s on subsequent maize crop. p. 157-169. In J. Van der Heide (ed.). Nutrient mana gement for food crop production in tropical farming systems. Proc. Symp. in Mala ng, 19-24 Oct. 1987. Inst. Soil Fertil., Haren, Netherlands. Hairiah, K., M. Noondwijk, and S. Setijono. 1991 . Tolerance to aci d soil conditions of velvet beans ( Mucuna pruriens Var. utilis and M. deeringiana ): Root development. Plant Soil 134:95-105. Hairiah, K., W.H. Utomo, and J. Van de r Heide. 1992. Biomass production and performance of leguminous cover crops in on an ultisol in Lampung. Agrivita. 15:39-44.

PAGE 215

200 Handayanto, E., G. Cadisch, and K.E. Giller. 1994. Nitrogen release from pruning of hedgerow trees in relation to quality of the pruning and incubation method. Plant Soil 160:237-248. Handayanto, E., G. Cadisch, and K.E. Giller. 1995. Manipulation of quality and mineralization of tropical legume tree pr unings by varying nitrogen supply. Plant Soil 176:149-160. Handayanto, E., K.E. Giller, and G. Cadi sch. 1997. Nitrogen mineralization from mixtures of legume tree prunings of differe nt quality and recovery of nitrogen by maize. Soil Biol. Biochem. 29:1417-1426. Hanna, W.W., and V. Monson. 1988. Registra tion of dwarf napiergrass germplasm. Crop Sci. 28:870-871. Heal, O.W., J.M. Anderson, and M.J. Sw ift. 1997. Plant litter quality and decomposition: An historical ove rview and decomposition. p. 3-32. In G. Cadisch, and K.E. Giller (eds.). Driven by nature: Plant litter quality. CAB Int., Wallingford, UK. Heaney, D.G., and W.J. Pigden. 1963. Interr elationship and conversion factors between expressions of digestible energy of forages. J. Anim. Sci.. 22:956-960. Heisey, P.W., and W. Mwangi. 1996. Fert ilizer use and maize production in subSaharan Africa. International Maize and Wheat Improvement Center, Mexico, D.F. Henderson, G.R., and P.T. Prest on. 1977. Fodder farming in Kenya. East Afr. Lit. Bur., Nairobi. Hepperly, P., H. Aguilar-Erazo, R. Perez, M. Diaz, and C. Reyes. 1992. Pigeon pea and velvet bean allelopathy. p. 357-370. In S.J.H. Rizvi, and V. Rizvi (eds.). Allelopathy: basic and applied aspe cts. Chapman and Hall, London, UK. Hildebrand, P.E. 19984. Modified stability anal ysis of farmer-managed on-farm trials. Agron. J. 76:271-274. Hildebrand, P.E. 1996. Adaptability analys is: a method for the design, analysis, and interpretation of on-farm research –extension (1st edition). Iowa State University Press, Ames, IA. Hulugalle, N.R., R. Lal, and C.H.H. Ter Kuille. 1986. Amelioration of soil physical properties by mucuna after mechanized la nd cleaning of a tropical rain forest. Soil Sci. 141:219-224.

PAGE 216

201 Hulugalle, N.R., D.L. Larsen, and S. Hengge ler. 1996. Effect of broad beds and dolichos residue management on properties of an irrigated vertis ol. Soil Technol. 8:275-286. Humphreys, L.R. 1994. Tropical forages: Their role in sustainable agriculture. Longman, Harlow, UK. Ibewiro, B., N. Sanginga, N., B. Vanlauwe, and R. Merckx. 2000a. Nitrogen contributions from decomposing cover crop residues to maize in a tropical derived savanna. Nutr. Cycl. Agroecosyst. 57:131-140. Ibewiro, B., N. Sanginga, B. Vanlauwe, and R. Merckx. 2000b. Evaluation of symbiotic dinitrogen inputs of herbaceous legumes in to tropical cover-cr op systems. Biol. Fertil. Soils 32:234-242. Ibewiro, B., B. Vanlauwe, N. Sanginga, and R. Merckx. 1998. Nitrogen contribution of roots to succeeding maize in herbace ous legume cover cropping systems in a tropical derived savanna. p. 123-128. In G. Renard, A. Neef, K. Becker, and M. von Oppen (eds.). Soil fertility management in West African land use systems. Proc. Workshop, Niamey, Niger, 4-8 Mar. 1997. Ibrahim, A.E.S. 1994. Influence of nitrogen application and stag e of harvesting on the dry matter production of Sorghum sudanese Lablab purpureus mixtures in central Sudan. Exp. Agric. 30:447-452. Ile, E., M.K. Hamadina, Z. Zafa, and J. Henrot. 1996. Note on effects of a Mucuna pruriens var utilis crop on the growth of maize ( Zea mays ) on an acid ultisol in southeastern Nigeria. Field Crops Res. 48:135-140. Infante, M.E., A.M. Perez, M.R. Simao, F. Manda, E.F. Baquete, and A.M. Fernandes. 1990. Outbreak of acute toxic psychosis attributed to Mucuna pruriens . The Lancet 336:1129. Jaetzold, R., and H. Schmidt. 1983. Farm management handbook. Part B: Rift valley and central provinces. Mi nistry of Agriculture, Republic of Kenya, Nairobi, Kenya. Jama, B., R.A. Swinkels, and R.J. Buresh. 1997. Agronomic and economic evaluation of organic and inorganic sour ces of phosphorus in western Kenya. Agron. J. 89:597604. Jones, G.A., T.A. McAllister, A.D. Muir, and K.J. Cheng. 1994. Effects of sainfoin ( Onobsychis viciifolia Scop) condensed tannins on gr owth and proteolysis by four strains of ruminal bacteria. Ap pl. Environ. Microbiol. 60:1374.

PAGE 217

202 Juo, A.S.R., A. Dabiri, and K. Franzluebbers. 1995a. Acidification of a kaolinitic Alfisol under continuous croppi ng and nitrogen fertilization in west Africa. Plant Soil 171:245-253. Juo, A.S.R., K. Franzluebbers, A. Dabiri, and B. Ikhile. 1995 b. Changes in soil properties during long-term fallow and co ntinuous cultivation after fore st clearing in Nigeria. Agric. Ecosyst. Environ. 56:9-18. Juo, A.S.R., K. Franzluebbers, A. Dabiri, and B. Ikhile. 1996 . Soil properties and crop performance on a kaolinitic Alfisol afte r 15 years of fall ow and continuous cultivation. Plan t Soil 180:209-217. Kachaka, S., B. Vanlauwe, and R. Me rckx. 1993. Decomposition and nitrogen mineralization of prunings of different quality. p. 199-208. In K. Mulongoy, and R. Merckx (Eds.). Soil organic matter dyna mics and sustainability of tropical agriculture. John Wiley a nd Sons, Chichester, U.K. Kang, B.T. 1993. Changes in soil chemical properties and crop performance with continuous cropping on an Entisol in the humid tropics. p. 297-305. In K. Mulongoy, and R. Mercks (eds.). Soil or ganic matter dynamics and sustainability of tropical agriculture , John Wiley and Sons, Chichester, U.K. Kapkiyai, J.J., K.N. Karanja, J.N. Qureshi, P.C. Smithson, and P.L. Woomer. 1999. Soil organic matter and nutrient dynamics in a Kenyan nitosol under long-term fertilizer and organic input manage ment. Soil Biol. Biochem. 31:1773-1782. Kapkiyai, J.J., K.N. Karanja, P.L. Woomer, and J.N. Qureshi. 1998. Soil organic matter fractions in a long-term experiment and the potential for their use as a diagnostic assay in highland farming systems of central Kenya. Afr. J. Crop Sci. 6:19-28. KARI (Kenya Agricultural Rese arch Institute). 1991. Ke nya’s Agricultural Research Priories to the year 2000. KARI, Nairobi, Kenya. KARI (Kenya Agricultural Rese arch Institute). 2000. Eff ect of organic and inorganic fertilizers and their combinations on mai ze yields on smallholder farms. p. 15-17. 1997 KARI Ann. Rep., KARI, Nairobi, Kenya. Kariuki, J.N., C.K. Gachuiri, G.K. Gitau, S. Tamminga, J. Van Bruchem, J.M.K. Muia, and K.R.G Irungu. 1998. Effect of feeding napiergrass, lucern, and sweet potato vines as sole diets to da iry heifers on nutri ent intake, weight gain and rumen degradation. Livest. Prod. Sci. 55:13-20. Kariuki, J.N., G.K. Gitau, C.K. Gachuiri, S. Tamminga, J.M.K. Muia. 1999. Effect of supplementing napiergrass with desmodi um and lucerne on DM, CP and NDF intake and weight gains in dairy heifers. Liv. Prod. Sci. 60, 81-88.

PAGE 218

203 Karlen, D.L., R.L. Flannery, and E.J. Sadler . 1988. Aerial accumulation and partitioning of nutrients by corn. Agron. J. 80:232-242. Kenya, Republic of. 2000. Economic Survey. Ce ntral Bureau of Statistics, Ministry of Planning and National Devel opment, Nairobi, Kenya. Khalili, H., P.O. Osuji, N.N. Umunna, and S. Crosse. 1994. The effects of forage type (maize-lablab or oat-vetch) and level of supplementation (wheat middlings) on food intake, diet apparent digestibilit y, purine excretion and milk production of crossbred ( Bos taurus x Bos indicus ) cows. Anim. Prod. 58:321-328. Kiflewahid, B., and B. Mosimanyana. 1992. Dolichos lablab ( Lablab purpureus) in byproduct based diets for lactating cows in Botswana. p. 155-71. In ARNAB (African Research Network for Agri cultural by-products). Overcoming constraints to the efficient utilization of agricultural by-products as animal feed. Proc. 4th annu. Workshop, Inst. Anim. Res., Mankon Stn., Bameda, Cameroon, 20-27 Dec. 1987. Kolawole, G.O., and B.T. Kang. 1997. Effect s of seed size and phosphorus fertilization on growth of selected legumes. Co mmun. Soil. Sci. Plant Anal. 28:1223-1235. Kumar Rao, J.V.D.K., and P.J. Dart. 1987. Nodulation, nitrogen fixation and nitrogen uptake in pigeon pea [ Cajanus cajan (L.) Millsp] of diff erent maturity groups. Plant Soil. 99:255-266. Kunkle, W.E., R.S. Sand, and D.O. Ral. 1994. Effects of body condition on productivity in beef cattle. Florida cooperative ex tension service, Inst. Food Agric. Sci., University of Florida. Lal, R. 1989. Agroforestry systems and soil surface management of a tropical Alfisol. III. Changes in soil chemical properties. Agrofor. Syst. 8:113-132. Lal, R. 1995. Erosion-crop productivity relations hips for soils of Africa. Soil Sci. Soc. Am. J. 59:661-667. Larson, B.A., and G.B. Frisvold. 1996. Fertili zers to support agricu ltural development in sub-Saharan Africa: What is n eeded and why. Food policy 21:509-525. Ledgard, S.J., and K.E. Giller. 1995. Atmospheric N2-fixation as an alternative nitrogen source. p.. 443-486. In P. Bacon (ed.). Nitrogen fertilization and environment. Marcel Dekker, NY. Lekasi, J.K., J.C. Tanner, S.K. Kimani, P.J. C. Harris. 1998. Manure management in the Kenyan highlands: Practices and potent ial. DFID/NRSP Project R6731, HDRA Publications, Kenilworth, UK.

PAGE 219

204 Lensi, R., A.M. Domenach, and L. Abbadie. 1992. Field study of nitrification and denitrification in wet savanna of west Africa. Plant Soil 147:107-113. Lorenzetti, F., S. MacIsaac, J.T Arnason, D.V.C. Awang, and D. Buckles. 1998. The phytochemistry, toxicology and food potenti al of velvetbean (Mucuna adans spp., Fabaceae). p. 67-84. In D. Buckles, A. Eteka, O. Osiname, M. Galiba, and G. Galiano (eds.). Cover crops in West Africa contributing to sustainable agriculture. IDRC/IITA/Sasaka wa Global 2000, Ottawa, Canada. Maasdorp, B.V., and M. Titterton. 1997. Nutri tional improvement of maize silage for dairying: Mixed-crop silage s from sole and intercropped legumes and a long season maize. 1. Biomass yield and nutri tive value. Anim. Feed Sci. Technol. 69:241-261. Mafongoya, P.L., P. Barak, J.D. Reed. 2000. Carbon, nitrogen, and phosphorus mineralization of tree leaves and ma nure. Biol. Fert il. Solis 30:298-305. Mafongoya, P.L, K.E. Giller, and C.A. Palm. 1997b. Decomposition and nutrient release patterns of prunings and litter of agroforestry trees. Agrofor. Syst. 38:7797. Mafongoya, P.L., K.E. Giller, and C.A. Palm. 1998. Decomposition and nitrogen release patterns of tree prunings and li tter. Agrofor. Syst. 38:77-97. Mafongoya, P.L., and P.K.R. Nair. 1997. Mu ltipurpose tree prunings as a source of nitrogen to maize under semiarid conditions in Zimbabwe. 1. Nitrogen recovery in relation to pruning quality and method of application. Agrofor. Syst. 35:31-46. Mafongoya, P.L, P.K.R. Nair, and B.H. Dzow ela. 1997a. Multipurpose tree prunings as a source of nitrogen to maize under semi arid conditions in Zimbabwe. 2. Nitrogen-recovery rates and crop growth as influenced by mixtures and prunings. Agrofor. Syst. 35:47-56. Makembe, N.E.T., and R. Ndlovu. 1996. Dolichos lablab ( Lablab purpureus cv. Rongai) as supplementary feed to mai ze stover for indigenous female goats in Zimbabwe. Small Rumin. Res. 21:31-36. Mandimba, G.R. 1995. Contribution of nodulated legumes on the growth of Zea mays L. under various cropping systems. Symbiosis 19:213-222. Manyong, V.M., A.V. Houndekon, A. Govan, M.V. Versteeg, and F. Van der Pol. 1996. Determinants of adoption for a resource management technology: the case of mucuna in Benin Republic. p. 86-93. In S. Zhang, and Y. Wang (eds.). Proc. Adv. Agric. Boil. Environ. Engin., Agric. Univ. Press, Beijing, China.

PAGE 220

205 Mausolff, C., and S. Farber. 1995. An ec onomic analysis of ecological agricultural technologies among peasant farmers in Honduras. Ecol. Econ. (Amsterdam) 12:237-248. Mbogoh, S.G. 1992. An analysis of some socio-economic aspects of dairy production and marketing in Kenya. A paper pr esented at a KARI/ISNAR workshop on program planning and dairy research pr iority setting held on 4-9, May, 1992, Nairobi, Kenya. McIntire, J., and J.M. Powell. 1995. Af rican semi-arid agriculture cannot produce sustainable growth without ex ternal inputs. p. 539-554. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestock and sustainable nutrient cycles in mixed-fa rming systems of sub-saharan Africa. Volume II. Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993, Addis Ababa, Ethiopia. McLeod, M.N., P.M. Kennedy, and D.M.J. Mins on. 1990. Resistance of leaf and stem fractions of tropical forages to chewing a nd passage in cattle. Br. J. Nutr. 63:105119. McSorley, R., D.W. Dickson, and J.A. Brito. 1994. Host status of selected tropical rotation crops to four populations of r oot-knot nematodes. Nematropica 24:45-53. Meentemeyer, V. 1998. Macroclimate and ligni n control of litter d ecomposition rates. Ecology 59:465-472. Melillo, J.M., J.D. Aber, A.E. Linkins, A. Ri cca, B. Fry, and K.J. Nadelhoffer. 1989. Carbon and nitrogen dynamics along a decay continuum: Plant litter to soil organic matter. p. 53-62. In Clarkholm, M. and Bergstrom, L. (eds.) Ecology of arable land. Kluwer Academic Pub lishers, Dordrecht, the Netherlands. Milera, M., V. Remy, H. Santana, J. Martinez, and N. Cabrena. 1989. The effect of the inclusion of Lablab Purpureus forage as 30 % of dairy cow rations. (In Portuguese, with English abstract .) Pastos Y Forrages 12:65-70. Mitchell, C.C., R.L. Westerman, J.R. Brow n, and T.R. Peck. 1991. Overview of longterm agronomic research. Agron. J. 83:24-29. Mertens, D.R., 1994. Regulation of forage intake. p. 450-493. In G.C. Fahey, Jr., M. Collins, D.R. Mertens, and L.E. Moser (eds.). Forage quality, evaluation, and utilization. ASA, CSSA, SSSA , Madison, Wisconsin, USA. Minson, D.J., Milford, R., 1967. The voluntar y intake and digestibility of diets containing different prop ortions of legume and mature pangola grass ( Digitaria decumbens ). Aust. J. Exp. Agric. Anim. Husb. 7:546-551.

PAGE 221

206 Mohamed-Saleem, M.A. 1984. Integration of forage legumes into the cropping systems of Nigerian sub-humid zone. p. 304-325. In R. Von Kaufmann, S. Charter, and R. Blench (eds.). Livestock system res earch in Nigerian sub-humid zone. Proc. ILCA/NAPRI Symposium, 29 Oct. Nov. 1984, ILCA, Addis Abbaba. Mohamed-Saleem, M.A. 1998. Nutrient balan ce patterns in Africa livestock systems. Agric. Ecosyst. Environ. 71:241-254. Mokwunye, A.U., and L.L. Hammond. Ferti lizer use in the tr opics. p. 121-134. In R. Lal, and P.A. Sanchez (eds.). Myths and science of soils of the tropics. SSSA Spec. Publ. No. 29, SSSA, ASA, Madisson, WI. Mugendi, D.N., and P.K.R. Nair. 1997. Pr edicting decomposition patterns of tree biomass in tropical highland microregions of Kenya. Agrofor. Syst. 35:187-201. Mugwira, L.M., and H.K. Murwira. 1997. Us e of cattle manure to improve soil fertility in Zimbambwe: Past and current research and future research needs. CIMMYT, Harare, Zimbambwe. Muhr, L., M. Peters, S.A. Tarawali, and R. Schultze-Kraft. 1999a. Forage legumes for improved fallows in agropastoral syst ems of subhumid West Africa. I. Establishment, herbage yield and nutritive value of legumes as dry season forage. Trop. Grassl. 33:222-233. Muhr, L., S.A. Tarawali, M. Peters, and R . Schultze-Kraft. 1999b. Forage legumes for improved fallows in agropastoral systems of subhumid West Africa. II. Green manure production and decomposition after incorporation into the soil. Trop. Grassl. 33:234-244. Muhr, L., S.A. Tarawali, M. Peters, and R. Schultze-Kraft. 1999c. Forage legumes for improved fallows in agropastoral systems of sub-humid west Africa. III. Nutrient import and their rotational effects on subsequent maize. Trop. Grassl. 33:245256. Muia, J.M.K., S. Tamminga, P.N. Mbugua. 2000a. Effect of supplementing napiergrass ( Pennisetum purpureum ) with sunflower meal or pou ltry litter-based concentrates on feed intake, live-weight changes and economics of milk production in Friesian cows. Livest. Prod. Sci. 67:89-99. Muia, J.M.K., S. Tamminga, P.N. Mbugua, and J.N. Kariuki. 1999. Optimal stage of maturity for feeding napiergrass ( Pennisetum purpureum ) to dairy cows in Kenya. Trop. Grassl. 33:182-190. Muia, J.M.K., S. Tamminga, P.N. Mbugua, and J.N. Kariuki. 2000b. The nutritive value of napiergrass ( Pennisetum purpureum ) and its potential for milk production with or without supplementation: a review. Trop. Sci. 40:109-131.

PAGE 222

207 Muia, J.M.K., S. Taminga, P.N. Mbugua, and J.N. Kariuki. 2001. Effect of supplementing napiergrass ( Pennisetum purpureum ) with poultry litter and sunflower meal based concentrates on feed intake and rumen fermentation in Friesian steers. Anim. Feed Sci. Technol. 92:113-126. Muinga, R.W., W. Thorpe, and J.H. Topps. 1992. Voluntary food intake, live-weight change and lactation performance of crossbred dairy cows given ad libitum Pennisetum purpureum (napiergrass var. Bana) supplemented with leucaena forage in the lowland semi-humid tropics. Anim. Prod. 55:331-337. Muinga, R.W., W. Thorpe, and J.H. Topps . 1993. Lactational performance of Jersey cows given napiergrass fodder ( Pennisetum purpureum ) with and without protein concentrates in the semi-humid tropi cs. Trop. Anim. Health Prod. 25:118-128. Muinga, R.W., J.H. Topps, J.A. Rooke, and W. Thorpe. 1995. The effect of supplementation with Leucaena leucocephala and maize bran on voluntary food intake, digestibility, live weight and milk yield of Bos indicus and Bos taurus dairy cows and rumen fermentation in steers offered Pennisetum purpureum ad libtum in the semi – humid tropics. Anim. Sci. 60:13. Mukisira, E.A., L.E. Phillip, and B.N. Mita ru. 1995. The effect of feeding diets containing intact or part ially detoxified lupin on voluntary intake and milk production by Friesian dairy cows. Anim. Sci. 60:69-175. Mupangwa, J.F., N.T. Ngongoni, J.H. Topps, and H. Hamudikuwanda. 2000. Effects of supplementing a basal diet of Chloris gayana hay with one of three protein-rich legume hays of Cassia rotundifolia , Lablab purpureus , and Macroptilium atropurpureum forage on some nutritional parameters in goats. Trop. Anim. Health Prod. 32:245-256. Mupangwa, J.F., N.T. Ngongoni, J.H. Topps, and P. Ndlovu. 1997. Chemical composition and dry matter degradabili ty profiles of forage legumes Cassia rotundifolia cv. Wynn, Lablab purpureus cv. Highworth, and Macropitilium atropurpureus cv. Siratro at 8 weeks of growth (pre-anthesis). Anim. Feed Sci. Technol. 69:167-178. Muriuki, E.M. 1998. A private sector perspectiv e on fertilizer delivery constraints Soil fertility working group workshop in Kenya , Continental Hotel, Mombasa, Kenya, 15-19 Nov. 1998. CMRT – Egerton Univ ersity, KARI and CIMMYT, Njoro, Kenya. Murwira, H.K. 1995. Ammonia losses from Zimbambwean cattle manure before and after incorporation to the soil. Trop. Agric. 72:269-276.

PAGE 223

208 Murwira, H. K., and H. Kirchmann. 1993. Nitrogen dynamics and maize growth in a Zimbabwean sandy soil under manure ferti lization. Commun. Soil. Sci. Plant Anal. 24:2343-2359. Muyekho, F.N., J. Mbugua, L. Karani, D. Wekesa, J. Apongo, and J. Musuya. 1998. The farming systems of Tumaini farm , Kaplamai Division: Summary from community-based activities. p. 1-10. In D.J. Rees, C. Nkonge, J.L. Wandera, V.I Manson, and F. Muyekho (eds.). Particip atory rural appraisal of the farming systems of the north of Rift Valle y Province, Kenya, 1995-1997, KARI, Kitale, Kenya. Muza, L. 1998. Selecting green-manure legumes for relay and intercropping systems with maize on sandy soils in Zimbabwe. p. 251-257. In D. Buckles, A. Eteka, O. Osiname, M. Galiba, and G. Galiano (eds.). Cover crops in west Africa: Contributing to sustainable agricult ure. IDRC-IITA-Sasakawa Global 2000, Ottawa, Canada. Myers, R.J.K, C.A. Palm, E. Cuevas, I.U.N. Gunakillrke, and M. Brossard. 1994. The synchronization of nutrient mineralizati on and plant nutrient demand. p. 81-116. In P.L. Woomer, and M.G. Swift (eds.). Biological management of soil fertility, Wiley-Sayce co-publication, New York, USA. NARC (National Agricultural Research Cent re). 1995. Participatory rural appraisal reports of Anin sub-location in Keiyo Di strict, Chobosta village in Uasin Gishu District, Matunda village in Trans-Nzoia District and Cheptuya village in West Pokot District. KARI-Rockefeller Foundati on on-farm soil management project, NARC, Kitale, Kenya. Nandwa, S,M., and M.A. Bekunda. 1998. Resear ch on nutrient flows and balances in east and southern Africa: state-of-the-a rt. Agric. Ecosyst. Environ. 71:5-18. Nnadi, L.A., and I. Haque. 1988. Root nitrogen transformation and mineral composition in selected forage legumes. J. Agric. Sci. (Camb.) 111:513-518. Norby, R.J., and M.F. Cotrufo. 1998. A que stion of litter quality. Nature 396:17-18. Norton, B.W., and J.H. Ahn. 1997. A comparison of fresh and dried Calliandra calothyrsus supplements for sheep given a basal di et of barley straw. J. Agric. Sci. (Cambridge) 129:485-494. Norton, B.W., and D.P. Poppi. 1995. Compositi on and nutritional attributes of pasture legumes. p. 23-47. In J.F.P. D’Mello, and C. Deve ndra (eds.). Tropical legumes in animal nutrition. CAB Int., Wallingford, U.K. NRC (National Research Council). 1989. Nutr ient requirements of dairy cattle. 6th edition. National Academy Press, Washington, D.C.

PAGE 224

209 Nsahlai, I. V., N.N. Umunna, and M.L.K. Bonsi. 1998. The utilization of teff ( Eragrotis tef ) straw by sheep fed supplementary fora ge legumes with or without either crushed maize grain or wheat bran. Small Rumin. Res. 29:303-315. Nyambati, E.M. 1997. Dairy cattle research achievements, feeding practices, constraints and strategies for future research. p. 188-201. In D. Rees, C. Nkonge, and J.L. Wandera (eds.). A review of agricultural practices and constrai nts in the north of Rift Valley Province, Kitale , Kenya. Kenya Agric. Res. Inst., Kitale, Kenya. Nyambati, E.M., L.E. Sollenberger, C.K. Hiebsch, D.M. Sylvia, and S.C. Rono. 2001. Nitrogen contribution from relay-cropped mu cuna and lablab to succeeding maize in the sub-humid highlands of Ke nya. Agron. Abstr. CD-ROM, a06sollenberger142348-0. Nziguheba, G., C.A. Palm, R.J. Buresh, and P.C. Smithson. 1998. Soil phosphorus fractions and adsorption as affected by organic and inorganic sources. Plant Soil 198:159-168. Obeid, J.A., J.A. Gomide, and M.E. Cruiz. 1992. Maize ( Zea Mays ) and legume silage mixtures for feeding housed steers. (I n Portuguese, with English abstract.) Revista da sociedade Brasilei ra de Zootecnia. 21:33-38. Oglesby, K.A., and J.H. Fowness. 1992. Effects of chemical composition on nitrogen mineralization from green manures of seven tropical leguminous tr ees. Plant Soil 143:127-132. Oikeh, S.O., V.O. Chude, R.J. Carsky, C.K. Weber, and W.J. Horst. 1998. Legume rotation in the moist tropical savanna: Managing soil nitrogen dynamics and cereal yields in farmers’ fi elds. Expl. Agric. 34:73-83. Osuji, P.O., Nsahlai, I.V. and Khalili, H. 1993. Feed Evaluation. ILCA manual 5. ILCA (International Livestock Centre for Africa), Addis Ababa, Ethiopia. Osuji, P.O., and A.A. Odenyo. 1997. The role of forages as supplements to low quality roughages – ILRI experience. Anim. Feed Sci. Technol. 69:27-38. Otsyula, R.M., S.I. Ajanga, R.A. Buruchar a, C.S. Wortman. 1998. Development of an integrated bean root rot control strate gy for western Kenya. Afr. Crop Sci. J. 6:61-67. Otsyula, R.M., J.H. Ndiritu, and R.A. Buruch ara. 1999. Interaction between bean stem maggot, bean root ro t and soil fertility. In G. Farrell, and G.N. Kibata (eds.). Proceedings of the 2nd biennial crop protection c onference, 16-17 Sept. 1998, Nairobi, Kenya. KARI/DFID NARP II.

PAGE 225

210 Pachauri, V.C., and R.S. Upandhyaya. 1982. A note on feeding value of velvet bean ( Stizolobium deeringianum ) hay for babari goats. Forage Res. 8:71-74. Palm, C.A. 1995. Contribution of agroforest ry trees to nutrient requirements of intercropped plants. Agrofor. Syst. 30:105-124. Palm, C.A., C.N. Gachengo, R.J. Delve, G. Cadisch, and K.E. Giller. 2001. Organic inputs for soil fertility management in tr opical agroecosystems: application of an organic resource database. Agri c. Ecosyst. Environ. 83:27-42. Palm, C.A., R.J.K. Myers, and S.M. Nandw a. 1997. Combined use of organic and inorganic nutrient sources for soil fertil ity maintenance and replenishment. p. 193-215. In R.J. Buresh, P.A. Sanchez, and F. Calhoun (eds.). Replenishing soil fertility in Africa. SSSA Spec. Publ. No. 51. SSSA, Madison, WI. Palm, C.A., and A.P. Rowla nd. 1997. A minimum dataset fo r characterization of plant quality for decomposition. p. 379-392. In G. Cadisch, and K.E. Giller (eds.). Driven by Nature: Plant Litter Qual ity and Decomposition. CAB Int, Wallingford, UK. Palm, C.A., and P.A. Sanchez. 1991. Nitroge n release from the leaves of some tropical legumes as affected by their lignin and polyphenolic contents. Soil Biol. Biochem. 23:83-88. Parkinson, J.A., and S.E. Allen. 1975. A wet oxidation procedure suitable for the determination of nitrogen and mineral nutri ents in biological materials. Commun. Soil Sci. Plant Anal. 6:1-11. Paustian, K., W.J. Parton, and J. Persson. 1992. Modeling soil organic matter in organic amended and nitrogen-fertilized long-term plots. Soil Sci. Soc. Am. J. 56: 476488. Pearson, D., 1976. The Chemical Analysis of Foods. 7th edition. Chemical publishing Co. Inc, New York. Peoples, M.P., D.F. Herridge, and J.K. Ladha . 1995. Biological ni trogen fixation: An efficient source for nitrogen for agri cultural production. Plant Soil 174:3-28. Poppi, D.P., and B.W. Norton. 1995. In take of tropical legumes. p. 23-47. In J.F.P. D’Mello, and C. Devendra (eds.). Tropi cal legumes in animal nutrition. CAB Int., Wallingford, U.K. Powell, J.M., S. Fernandez-Rivera, and S. Ho fs. 1994. Effects of sheep diet on nutrient cycling in mixed farming systems of semi -arid west Africa. Agric. Ecosyst. Environ. 48:263-271.

PAGE 226

211 Powell, J.M., and M.A. Mohamed-Saleem. 1987. Nitrogen and phosphorus transfers in a crop-livestock system in West Africa. Agric. Syst. 25:261-277. Powell, J.M., and C. Valentine. 1998. Eff ects of livestock on soil fertility in West Africa. p. 319-338. In G. Renard, A. Neef, K. Beck er, and M. von Oppen (eds.). Soil fertility management in West African land use systems. Proc. Workshop, 4-8 Mar. 1997, Niamey, Niger. Powell, J.M., and T.O. Williams. 1993. Li vestock, nutrient cycling and sustainable agriculture in west African Sahel. Gate keeper series No. 37. Int. Inst. Environ. Dev., London. Powell, J.M., and T.O Williams. 1995. An overview of mixed farming systems in subSaharan Africa. p. 21-36. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestock and sust ainable nutrient cycles in mixed-farming systems of sub-Saharan Africa. Volume II: Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993, Addis Ababa, Ethiopia. Probert, M.E., J.R. Okalebo, and R.K. Jones. 1995. The use of cattle manure on smallholder farms in semi-arid easte rn Kenya. Exp. Agric. 31: 371-381. Rajaram, N., and K. Jamardhanan. 1991. The biochemical composition and nutritional potential of the tribal pulse, Mucuna gigantean (Wild) DC. Plant Food Hum. Nutrn. 4:45-51. Ramos, M.G., M.A.A. Villatoro, S. Urquiag a, B.J.R. Alves, R.M. Bondey. 2001. Quantification of BNF to tropical green ma nure crops and the residual benefit to a subsequent maize crop using 15N-isotope technique. J. Biotech. 91:105-115. Ravindran, V. 1988. Studies of Mucuna Pruriens (L) DC as forage alternative in tropical countries: Evaluation of productivity and forage quality at four different stages. Beitrage Zur Tropschen Land wirtschaft Und Vetenarmedizin 26:381-386. Ravindran, V., and G. Ravindran. 1988. Nutri tional and antnutritiona l characteristics of mucuna ( Mucuna utilis ) bean seeds. J. Sci. Food Agric. 46:71-79. Reed, J.D. 1995. Nutritional toxicology of tannins and related polyphenols in forage legumes. J. Anim. Sci. 73:1516-1528. Reed, J.D., H. Soller, and A. Woodward. 1990. Fodder tree and straw diets for sheep: Intake, growth, digestibility and the e ffects of phenolics on nitrogen utilization. Anim. Feed Sci. Technol. 30:39-50. Rees, D.J., C. Nkonge, and J.L. Wandera. 1997. A review of agricultural practices and constraints in the north of Rift Valley Pr ovince, Kitale, Kenya. Kenya Agric. Res. Inst., Kitale, Kenya.

PAGE 227

212 Russell, C. A., and I.R.P. Fillery. 1994. Es timates of lupin below-ground biomass N, dry matter and N turnover to wheat. Aust. J. Agric. Res. 45: Ruto, J.K. 1992. Coordination and management of the maize research program. Proc. Workshop, 19-23 Nov. 1990, Golf Hotel, Kakamega, Kenya. Salawu, M.B., T. Acamovic, C.S. Stewart, and R.L. Roothaert. 1999. Composition and degradability of different fractions of Calliandra leaves, pod, and seeds. Anim. Feed Sci. Technol. 77:181-199. Sanchez, P.A., A.M. Izac, I. Valencia, and C. Pieri. 1996. Soil fertility replenishment in Africa: A concept note. p. 200. In S.A. Beth, S.A (ed.). Achieving greater impact from research investments in Africa. Sasakawa Africa Association, Mexico City, Mexico. Sanchez, P.A., K.D. Shepherd, M.J. Soule, F.M. Place, A.U. Mokwunye, R.J. Bureshi, F.R. Kwesiga, A.M.N., Izac, C.G. Ndiritu, and P.L. Woomer. 1997. Soil fertility replenishment in Africa. p. 1-46. In R.J. Buresh, P.A. Sanchez, and F. Calhoun (eds.). Replenishing soil fertility in Africa. SSSA Spec. Publ. No. 51. SSSA, Madison, WI. Sanginga, N, B. Ibewiro, P. Houngnandan, B. Vanlauwe, J.A. Okogum, I.D. Akobundu, and M. Versteeg. 1996a. Evaluation of symbiotic propert ies and nitrogen contribution of mucuna to maize grown in the derived savanna of west Africa. Plant Soil. 179:119-129. Sanginga, N., J.A. Okogun, I.O. Akobundu, and B.T. Kang. 1996b. Phosphorus requirement and nodulation of herbaceous and shrub legumes in low P soils of a Guinean Savanna in Nigeria. Appl. Soil Ecol. 3:247-255. SAS (Statistical Analysis Systems Institute). 2001. SAS system for windows. Release 8.12, SAS Inst., Cary, NC, USA. Scalbert, A. 1991. Antimicrobial prope rties of tannins. Phytochem. 30:3875-3883. Schreuder, R., P.J.M. Snijders, A.P. Wouters, A. Steg, and J.N. Kariuki. 1993. Variation in DM digestibility, CP, yield a nd ash content of napiergrass ( Pennisetum purpureum ) and their production from chemical and environmental factors. Res. Rept., Natl. Anim. Husb. Res. Stn., KARI, Naivasha, Kenya. Siddhuraju, P., and K. Becker. 2001. Rapi d reversed-phase high performance liquid chromatographic method for the qua ntification of L-Dopa (L-3,4dihydroxyphenylalanine), non-methylated a nd methylated tetrahydroisoquinoline compounds from Mucuna beans. Food Chem. 72:389-394.

PAGE 228

213 Singh, D., and L.L. Relwani. 1978. Mixed cropping of maize ( Z. mays L.) with cow pea ( Vigna sinensis ) and velvet bean ( S. deeringianum ) on the yield and chemical composition of fodder. Indian J. Dairy Sci. 31: 28-33. Singh, Y., B. Singh, and C.S. Khind. 1992. Nu trient transformations in soil amended with green manure. Adv. Soil Sci. 20:237-309. Skerman, P.J., Riveros, F., 1990. Tropical Grasses. FAO, Rome. Smaling, E.M.A. 1993. Soil nutrient depleti on in subSaharan Africa. p. 5367. In H. Van Reuter, and W.H. Prins (eds.). The ra te of plant nutrient for sustainable crop production in sub-Saharan Africa. Dutch Association of fert ilizer producers, Leilschendam, Netherlands. Smaling, E.M.A., and A.R. Braun. 1996. Soil fe rtility research in sub-Saharan Africa: New dimensions, new challenges. Co mmun. Soil Sci. Plant Anal. 27:365-386. Smaling, E.M.A., L.O. Fresco, and A. de Jager. 1996. Classifying, monitoring and improving soil nutrient stocks and flows in African agriculture. Ambio 25:492496. Smaling, E.M.A., S.M. Nandwa, B.H. Janssen. 1997. Soil fertility in Africa is at stake. p. 47-61. In R.J. Buresh, P.A. Sanchez, and F. Calhoun (eds.). Replenishing soil fertility in Africa. SSSA Spec. Publ. No. 51. SSSA, Madison, WI. Smith, J.L., R.I. Papendick, D.F. Bezdicek, and J.M. Lynch. 1993. Soil organic matter dynamics and crop residue management. p. 65-95. In M.F. Blaine Jr., (ed.). Soil microbial ecology: Applications in ag ricultural and environmental sciences. Marcel Decker, NY. Smyth, T.J., M.S. Cravo, and R.J. Melgar. 1991. Nitrogen supplied to corn by legumes in a central Amazon Oxisol. Tr op. Agric. (Trinidad) 68:366-372. Snapp, S.S., P.L. Mafongoya, and S. Wa ddington. 1998. Organic matter technologies for integrated nutrient management in sm allholder cropping systems of southern Africa. Agric. Ecosyst. Environ. 71:185-200. Snijders, P.J.M., A. Nahuis, F. Wekesa F., a nd A.P. Wouters. 1992. Yield and quality of napiergrass and green leaf desmodium at two cutting regimes. Res. Rep., Naivasha, Kenya. Soil Survey Staff. 1994. Key to soil taxonomy, 6th edition. USDA/S CS. Washington, USA. Sollenberger, L.E., and C.S. Jones. 1989. Beef production from N fertilized Mott elephantgrass. Trop. Grassl. 23:129-134.

PAGE 229

214 Somda, Z.C., J.M. Powell, S. Fernandez-Ri vera, and J.D. Reed. 1995. Feed factors affecting nutrient excretion by ruminants and fate of nutrients when applied to the soil. p. 227-246. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestoc k and sustainable nutrient cycles in mixed-farming systems of sub-Saharan Africa. Volume II. Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993. Addis Ababa, Ethiopia. Sotomayor, R.A., B.A. Quil, and W.W. Ha nna. 1997. Agronomic comparison of tall napiergrass in Puerto Rico. J. Agric. (Puerto Rico) 81:9-18. Spain, A.V., and R.P. Le Feure. 1987. Breakdow n of four litte r of contrasting quality in a tropical Australian rainfore st. J. Appl. Ecol. 24:278-288. Stamford, N.P., M.H. Albuquerque, and D.R. Santos. 1994. Nitrogen use by sorghum following legumes incorporated after cutti ng on different dates. (In Portuguese with English abstract). Revista Br asileira de Ciencia do solo 18:221-227. Stoorvogel, J.J., E.M.A. Smaling, and B.H. Janssen. 1993. Calc ulating soil nutrient balances in Africa at different scales. I. Supra-national scal e. Fert. Res. 35:227235. Strickland, M.J. 1975. Experiments with dairy cattle. In: P.D. Wood (ed.). Considerations for design and interpretati on of cattle experiments. Proceedings of a symposium on cattle experimentation, 8th Dec. 1975. Br. Soc. Anim. Prod. Swift, M.J., L. Bohren, and S.E. Carter 1994a. Biological management of tropical soils: Integrating process research and farm practice. p. 209-227. In P.L. Woomer, and M.J. Swift (eds.). The biological manageme nt of tropical soil fertility. J. Wiley and Sons, Chichester, UK. Swift, M.J., P.G.H. Frost, B.M. Campbell, J.C. Hatton, and K. Wilson. 1989. Nutrient cycling in farming systems derived from savanna. p. 63-76. In M. Clarhalm, and D. Berstrom, (eds.). Ecology of arid lands : Perspectives and ch allenges. Kluwer Academic Publishers, Dordrecht, Germany. Swift, M.J., P.D. Seward, P.G.H. Frost, J.N. Qureshi, and F.N. Muchena. 1994b. Longterm experiments in Africa: Developing a database for sustainable land use under global change. In R.A. Leigh, and A.E. Johnstone (eds.). Long-term experiments in agricultural and ecological scie nces. CAB Int. Wallingford, UK. Takasaki, S., and S. Kawakishi. 1997. Formation of protein-bound 2,4dihydroxyphenylalanine and 5-S-Cystei nyl-3,4-dihydroxyphenylalanine as new cross-linkers in gluten. J. Agric. Food Chem. 45:3472-3475.

PAGE 230

215 Tanner, J.C., J.D. Reed, and E. Owen. 1990. The nutritive valu e of fruits (Pods with seeds) from four Acacia spp . compared with extracted noug ( Guizotia abyssinica ) meal as supplements to maize stover for Ethiopian highland sheep. Anim. Prod. 51:127-133. Tarawali, G., and M.A. Mohamed Saleem. 1995. The role of forage fa llows in supplying improved feed and recycling nitrogen in sub-hum id Nigeria. p. 263-276. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestock and sustainable nutrient cycles in mixed-fa rming systems of sub-saharan Africa. Volume II. Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993, Addis Ababa, Ethiopia. Tarawali, S.A., and M. Peters. 1996. The potential contributions of selected forage legume pastures to cereal production in cr op-livestock systems. J. Agric. Sci. (Cambridge), 127:175-182. Ternouth, J.H. 1989. Endogenous losses of phosphorus by sheep. J. Agric. Sci. 113:291-297. Thomas, R.J., and C.E. Lascano. 1995. The benefits of forage legumes for livestock production and nutrient cycling in pastur e and agropastoral systems of Latin America. p. 277-291. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard (eds.). Livestoc k and sustainable nutrient cycles in mixed-farming systems of sub-Saharan Africa. Volume II: Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993. Addis Ababa, Ethiopia. Thomas, D., and J.E. Sumberg. 1995. A revi ew of the evaluation and use of tropical forage legumes in sub-Saharan Africa. Agric. Ecosyst. Environ. 54:151-163. Tian, G., L. Braussaard, and B.T. Kang. 1993. Biological effects of plant residues with contrasting chemical compositions unde r humid tropical conditions: effects on soil fauna. Soil Biol. Biochem. 25:731-737. Tian, G., and B.T. Kang. 1998. Effects of soil fertility and fertilizer application on biomass and chemical composition of legume cover crops. Nutr. Cycl. Agroecosyst. 51:231-238. Tian, G., B.T. Kang, and L. Braussaard. 1992. Effects of chemical composition or N, Ca and Mg release during incubation of l eaves from selected and fallow plant species. Biogeochem. 16:103-119. Tian, G., B.T. Kang, and L. Braussaard. 1997. The effect of mulch quality on earthworm activity and nutrient supply in the humid tropics. Soil Biol. Biochem. 29:369-373.

PAGE 231

216 Tian, G., G.O. Kolawole, B.T. Kang, a nd G. Kirchhof. 2000. Nitrogen fertilizer replacement indexes of legume cover crops in the derived savanna of west Africa. Plant Soil 224:287-296. Tian, G., G.O. Kolawole, F.K. Salako, and B.T. Kang. 1999. An improved cover cropfallow system for sustainable management of low activity clay soils of the tropics. Soil Sci. 164:671-682 Tilley, J.M.A., and R.A. Terry. 1963. A two stage technique for the in-vitro digestion of forage crops. J. Br. Grassl. Soc. 18:104-111. Tracy, S.M., and H.S. Coe. 1918. Velvet beans. Farmers Bull. 962, USDA, Washington, DC, USA. Tothill, J.C., B.H. Dzowela, and A.K. Diallo. 1989. Present and future role of grasslands in inter-tropical countries with special reference to ecological and sociological constraints. p. 1719-1724. Proc. XVI In t. Grassl. Congr., 4-11 Oct. 1989, Nice, France. Umoh, J.E. 1975. A short term comparison of the value of maize/mucuna and elephant grass ( Pennisetum purpureum ) silage for milk production. Inst. Agric. Res., Nigeria. Samaru Agric. Newsl. 17:57-61. Unkovich, M.T., J.S. Pate, and J. Hamblin. 1994. The nitrogen economy of broadacre lupin in southwest Australia. Aust. J. Agric. Res. 45: 149-164. Utomo,W.H, S.M. Sitompul, and M.V. Noor dwijk. 1992. Effects of leguminous cover crops on subsequent maize and soy bean crops on an ultisol in Lampang. Agrivita 15:44-53. Van den Bosch, H., V.N. Ogaro, S. Maobe, and J. Vlaming. 1998. Monitoring nutrient flows and economic performance in Afri can farming systems (NUTMON). III. Monitoring nutrient flows and balances in three districts in Kenya. Agric. Ecosyst. Environ. 71:63-80. Van Noordwijk, K.M., and P. Purnomisidi. 1992. Litter fall in cassava based cropping systems in Lampung. Agrivita 15:29-33. Vanlauwe, B., J. Diels, N. Sanginga, and R. Merckx. 1997. Residue quality and decomposition: An unsteady relationship? p. 157-166. In G. Cadisch, and K.E. Giller (eds.). Driven by Nature: Plant Litter Quality and Decomposition. CAB Int., Wallingford, UK. Vanlauwe, B., O.C. Nwoke, N. Sanginga, a nd R. Merckx. 1996. Impact of residue quality on the C and N mine ralization of leaf and root residues of three agroforestry species. Plant Soil 183:221-231.

PAGE 232

217 Van Soest, P.J., J.B. Robertson, and B.A. Lewis. 1991. Methods for dietary fiber, neutral detergent fiber, and non-starch polysaccharides in relation to animal nutrition. J. Dairy Sci. 74, 3583-3597. Versteeg, M. N., F. Amadji, A. Eteka, A. Gogan, and V. Koudokpon. 1998a. Farmers’ adoptability of Mucuna fallowing and agroforestry te chnologies in the coastal savanna of Benin. Agric. Syst. 56(3):269-287. Versteeg, M. N., F. Amadji, A. Eteka, V. Houndekon, and V.M. Manyong. 1998b. Collaboration to increase the use of mucuna in production systems in Benin. p. 33-43. In D. Buckles, A. Eteka, O. Osiname, M. Galiba, and G. Galiano (eds.). Cover crops in West Africa contribu ting to sustainable agriculture. IDRC/IITA/Sasakawa Global 2000, Ottawa, Canada. Vine, H. 1953. Experiments on the maintena nce of soil fertility at Ibadan, Nigeria, 1922-1951. Empire J. Exp. Agric. 21:65-85. Vlek, P.L.G. 1993. Strategies for sustaining agriculture in sub-Saharan Africa: The fertilizer technology issue. p. 265-276. Technologies for sustainable agriculture in the tropics. ASA Spec. Publ. No. 56, Madison, WI. Wade, M.K., and P.A. Sanchez. 1983. Mu lching and green manure applications for continuous crop production in the Amazon basin. Agron. J. 75:39-45. Wandera, J.L. 1996. General farming envir onment, constraints and opportunities to livestock feeds development in the high m ilk production regions of Central, Rift valley and Nyanza Provinces of Kenya. (A report of information generated using PRA techniques). Nat. Agric. Res. Cent., Kitale, Kenya. Ware–Austin, W.D. 1963. Napiergrass for m ilk production in Trans-Nzoia. East Afr. Agric. For. J. 26:223. Warren, G. 1992. Fertilizer phosphorus sorptio n and residual value in tropical African soils. NRI Bull. 37. Nat Re source Inst., Chatham, UK. Waterman, P.G., and S. Mole. 1994. Analysis of phenolic plant metabolites. Blackwell Scientific Publications, Oxford, UK. Weber, G. 1996. Legume based technologies for African savannas: Challenges for research and development. Biol. Agric. Hort. 13:309-333.

PAGE 233

218 Williams, T.O., J.M. Powell, and S. Fern andez-Rivera. 1995. Manure utilization, drought cycles and herd dynamics in th e Sahel: Implications for cropland productivity. p. 393-409. In J.M. Powell, S Fernandez-Rivera, T.O. Williams, and C. Renard. (eds.). Livestock and sust ainable nutrient cycles in mixed-farming systems of sub-saharan Africa. Volume II: Tech. Papers. Proc. Int. Conf., ILCA, 22-26 Nov. 1993, Addis Ababa, Ethiopia. Woods, I.M. 1983. Lablab bean ( Lablab purpureus ) for grain and forage production in the ord irrigation area. Aust. J. Exp. Agric. Anim. Husb. 23:162-171. Woodward, A., and J.D. Reed. 1997. Nitrogen metabolism of sheep and goats consuming Acacia brevispica and Sesbania sesban . J. Anim. Sci. 75:1130-1139. Wortmann, C.S., H.E. Gridley, S.M. Musaana. 1996. Seed yield a nd stability of bean multilines. Field Crops Res. 46:153-159. Wortmann, C.S., B.D. McIntyre, and C. K. Kaizzi. 2000. Annual soil improving legumes: agronomic effectiveness, nutr ient uptake, nitroge n fixation and water use. Field Crops Res. 68:75-83. Wouters, A.P. 1987. Dry matter yield and quality of napiergrass on farm level 19831986. Research report, Ministry of Li vestock Development, Nairobi, Kenya.

PAGE 234

219 BIOGRAPHICAL SKETCH Elkana Momanyi Nyambati was born on 20 D ec. 1961, in Kisii District in Kenya. Nyambati attended Ngenyi Primary School between 1968 and 1974. His seccondary and high school education were obtained between 1975 and 1980 at Nyansabakwa in Kisii and Strathmore School in Nairobi, respectiv ely. Nyambati attended Egerton University in 1981and graduated with a Distinction Dipl oma in Range Management in 1984 and was immediately employed by Kenya’s Ministry of Agriculture and Livestock Development as a Technical Officer. He was posted to Kajiado District where he worked as an Agricultural Extension Officer . In 1985, he was transferred to the Scientific Research Division of the same ministry (later the Kenya Agricultural Research Institute, KARI) and posted to the National Agricultural Rese arch Centre (NARC) at Kitale. In 1986 Nyambati joined the University of Nairobi on a study leave where he graduated with a BSc degree (honors) in 1989. He was immedi ately awarded a University of Nairobi scholarship to pursue an MSc degree whic h he obtained in 1993. Between 1994 and 1997, Nyambati worked as a Research Scientis t in KARI at NARC-Kitale. Nyambati is currently on a study leave at th e University of Florida wher e he is pursuing a doctoral degree in agronomy on a Rockefeller Foundati on Scholarship. Nyambati is married to Florence and they have three ch ildren, Kelvin, Newton, and Nancy.