BELOW- AND ABOVEGROUND PIGEONPEA PRODUCTIVITY IN ON-FARM SOLE AND INTERCROP SYSTEMS IN CENTRAL MALAWI By Chiwimbo P. Gwenambira A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Crop and Soil Sciences Master of Science 2015 ABSTRACT BELOW- AND ABOVEGROUND PIGEONPEA PRODUCTIVITY IN ON-FARM SOLE AND INTERCROP SYSTEMS IN CENTRAL MALAWI By Chiwimbo P. Gwenambira Smallholder farmers in Malawi face many challenges which include a degrading soil resource base. Pigeonpea is one legume that has shown promise in Malawi in terms of improving soil fertility but its below and aboveground productivity is not fully understood. On-farm trials were set-up in 2013/14 across three agro-ecologies in central Malawi. Pigeonpea was planted as a sole crop or in an additive intercrop system with soyabean, groundnut or maize (the farmer check system). The objectives of this study were to (1) assess the effect of cropping system and soil texture on pigeonpea root and shoot biomass and (2) to evaluate variability of pigeonpea growth within a smallholder farm context. Destructive harvest was conducted six months after planting to evaluate shoot parameters, and roots of the same plants were excavated fr om 0 60 cm. Cropping system and soil texture effected =0.05). Sole pigeonpea had the highest shoot biomass at 11.83 Mg ha -1 , root at 1.56 Mg ha -1 and pigeonpea/maize had the lowest shoot at 3.57 Mg ha-1 root at 0.53 Mg ha-1 . Root biomass was largely confined to the topsoil, with trends similar to that for aboveground biomass. The results confirm that intra-specific competition in a pigeonpea/maize intercrop is large, while pigeonpea productivity in pigeonpea/groundnut intercrop is comparable to sole cropped pigeonpea, with additional groundnut grain benefits. Promoting the later cropping system can enhance land productivity on smallholder farms in Malawi. iii To Melchisedec, the King Salem and all His loyal subjects, to my family and Dr. S.S Snapp iv ACKNOWLEDGMENTS My sincere thanks goes to my advisor, Dr. Sieglinde Snapp, for taking me on as her student. Her patience, wisdom, guidance, sense of humor and encouragement are what kept me going throughout the entire journey. I will always be grateful for her confidence in me and all the knowledge and skills I have gained under her tutelage. I really appreciate all the time my committee members, Dr. Richardson and Dr. Steinke invested in me. Their input and guidance contributed to the success of this work. Dr. Regis Chikowo offered immeasurable help during data collection and all analyses. I gained a lot from the knowledge he imparted on me. Members of the Snapp lab at Michigan State University are truly appreciated for their support, help and teamwork. They assisted me all the way, from my field work up to lab analyses. I would not have pulled through without them and special mention goes to these wonderful people Dan Kane, Erin Anders, Paul Rogé, Princess Adjei-Frimpong, Placid Mpeketula, Rich Price, Joel Clifton, Quinn Hanses, Paige Gurizzian, Mac Spitzley and Spencer Rosekrans This research would not have been possible without the Africa RISING mother and baby trial farmers. I am so thankful for their participation and for all that I learned from them. The Africa RISING team in Malawi, Lilongwe University of Agriculture and Natural Resources (LUANAR) faculty and staff and the agricultural extension personnel made my stay in Malawi a memorable one. Special thanks goes to Isaac Jambo, Emmanuel Jambo, Dr. Wezi Mhango, Edward Mzumara, v Elizabeth Bandason, Lackson Chirwa and Jeckner Phiri. Dr. Emmanuel Kaunda and family at LUANAR, saw to my every need during my data collection in Malawi and for that I will always be grateful. Carol Christofferson, Cal Bricker, Therese Iadipaolo, Darlene Johnson and Linda Colon from the Crop and Soil Sciences department at Michigan State University were a major source of support and help throughout my studies. Many thanks goes to Brad Peter for all the time he sacrificed in perfecting my maps as well as my climate data. His help is truly appreciated. I am very grateful to the MasterCard Foundation (MCF) for fully sponsoring me throughout my academic endeavors. I am thankful for that priceless scholarship and all the personal and professional development programs they offered me. Special mention goes to the MCF Scholars program team at MSU Dr. Kalumbu, Dr. Glew, Laura Wise, Jackie Thomas, Pam Farran and Dr. Onchiri. I also acknowledge the research funding I received from USAID. My family and friends have been a major source of strength and their faith in me helped me to press on. I am thankful to my parents, Mr. and Mrs. Gwenambira and my siblings Ruzivo and Joshua for their unconditional love, support and encouragement. vi TABLE OF CONTENTS LIST OF TABLES ................................................................................................ ..................... viii LIST OF FIGURES ..................................................................................................................... KEY TO ABBREVIATIONS .................................................................................................. CHAPTER ONE: LITERATURE REVIEW ............................................................................. 1 1.1.a Background .................................................................................................................... 1 1.1.b Malawi farming systems ................................................................ ................................ 1 1.1.c Legumes................................................................................................ .......................... 3 1.1.d Constraints to legume production .................................................................................. 3 1.1.e Long duration legumes ................................ ................................................................... 5 1.1.f Pigeonpea ....................................................................................................................... 6 1.1.g Underestimation of pigeonpea ................................................................ ....................... 8 1.1.h Pigeonpea-based cropping systems ............................................................................... 8 1.1.i Importance and uses of pigeonpea ............................................................................... 11 1.1.j Nutrient budgets ................................ ............................................................................ 11 1.1.k Research gaps .............................................................................................................. 13 CHAPTER TWO: BELOW AND ABOVEGROUND PIGEONPEA PRODUCTIVITY IN ON-FARM SOLE AND INTERCROP SYSTEMS IN CENTRAL MALAWI. ................... 15 2.1 INTRODUCTION............................................................................................................... 15 2.2 MATERIALS AND METHODS ................................................................ ........................ 19 2.2.1 Site description............................................................................................................. 19 2.2.2 Experimental Design ................................ .................................................................... 19 2.2.2.a Mother trials ............................................................................................................. 20 2.2.2.b Africa RISING Baby Trials ................................................................ ....................... 20 2.2.3. Agronomy .................................................................................................................... 21 2.2.4 Rainfall and temperature ................................................................ ............................. 22 2.2.5 Aboveground biomass assessment ............................................................................... 22 2.2.6 Belowground biomass assessment ............................................................................... 23 2.2.7 Soil Sampling and analyses ................................................................ ......................... 23 2.2.7.a Soil texture ................................................................................................................ 24 2.2.7.b Soil pH ...................................................................................................................... 24 2.2.7.c Inorganic N ............................................................................................................... 24 2.2.7.d Potentially mineralizable N (PMN) .......................................................................... 25 2.2.7.e Total soil N and Soil Organic Carbon (SOC) percent .............................................. 25 2.2.7 Biological N fixation .................................................................................................... 25 2.2.8 Statistical analysis ....................................................................................................... 26 2.3 RESULTS ................................................................................................ ........................... 27 2.3.1 Study locations and soil characteristics ...................................................................... 27 2.3.1.a Rainfall and temperature ................................................................ .......................... 27 2.3.2 Pigeonpea total shoot biomass .................................................................................... 28 vii 2.3.2.a Pigeonpea litter ................................ ................................ ................................ ......... 28 2.3.2.b Stems ................................ ................................ ................................ ......................... 29 2.3.2.c Twigs ................................ ................................ ................................ ......................... 30 2.3.2.d Leaves ................................ ................................ ................................ ....................... 30 2.3.2.e Pods ................................ ................................ ................................ ........................... 30 2.3.2.f Root biomass ................................ ................................ ................................ .............. 31 2.3.2.g Root shoot ratio ................................ ................................ ................................ ......... 32 2.3.2.h Soil pH ................................ ................................ ................................ ...................... 32 2.3.2.i Soil texture ................................ ................................ ................................ ................. 33 2.3.2.j Total soil N and SOC ................................ ................................ ................................ . 33 2.3.2.k NH 4 + - N, Soil NO 3 - - N and PMN ................................ ................................ ............. 34 2.4 DISCUSSION ................................ ................................ ................................ ..................... 35 2.4.1 Precipitation ................................ ................................ ................................ ................ 35 2.4.2 Pigeonpea shoot biomass ................................ ................................ ............................. 35 2.4.3 Root biomass ................................ ................................ ................................ ................ 38 2.4.4 Root shoot ratio ................................ ................................ ................................ ............ 40 2.4.5.a Soil pH ................................ ................................ ................................ ...................... 41 2.4.5.b Soil texture ................................ ................................ ................................ ................ 41 2.4.5.c Total N and SOC ................................ ................................ ................................ ....... 41 2.4.5.d Soil NO 3 - - N, NH 4 + - N and PMN ................................ ................................ ............. 42 2.4.6 Conclusions and future directions ................................ ................................ ............... 43 APPENDIX ................................ ................................ ................................ ................................ .. 44 BIBLIOGRAPHY ................................ ................................ ................................ ....................... 70 viii LIST OF TABLES Table 1 . Monthly precipitation in mm during the 2013/2014 growing season for the three sites. A historical 15 - yr monthly average precipitation is presented. Source: TRMM (2015). ................. 45 Table 2. for the three sites. A historical 15 - yr monthly average t emperature is presented. Source: NASA (2010). ................................ ................................ ................................ ................................ ........... 46 Table 3 . Soil properties measured in mother trials. Soil texture is indicated by silt, sand and clay percent. Total soil N and SOC are indicated by N % and C %, pH is the measure of the acidity or basicity of soil. Overall averages across mother trials are fol lowed by standard deviations in parentheses ................................ ................................ ................................ ................................ .... 47 Table 4 . Soil properties measured in baby trials. Soil texture is indicated by silt, sand and clay percent. Total soil N and SOC are indicated by N % and C %, pH is the measure of the acidity or basicity of soil. Overall averages across baby trials are followed by standard deviations in parentheses. ................................ ................................ ................................ ................................ ... 47 Table 5 . Cultivars of crops planted in the mother trials ................................ ............................... 48 Table 6 . Plant population densities in four pig eonpea - based cropping systems in mother trials 48 Table 7 . Planting dates across sites in mother trials ................................ ................................ ..... 48 Table 8 . Averages and ranges of shoot and root biomass across mother trials. Means are followed by standard deviations in par entheses. ................................ ................................ .......................... 49 Table 9 . Averages and ranges of shoot and root biomass in baby trials. Means are followed by standard deviations in parentheses. ................................ ................................ ............................... 49 Table 10 . Soil NO 3 - - N (mg/kg soil) and potential mineralizable N (PMN) (mg/kg soil/day) means by cropping system and soil depth in mother trials. Means are followed by standard deviations in parentheses. ................................ ................................ ................................ ................................ ... 50 Table 11 . Soil NO3 - - N (mg/kg soil) and potential mineralizable N (PMN) (mg/kg soil/day) means by soil texture and soil depth in baby trials. Means are followed by standard deviations in parenthe ses. ................................ ................................ ................................ ................................ ... 51 Table 12 . One - way ANOVA and mean response of pigeonpea litter, stems, twigs, leaves, pods, initial leaf fall, total shoot biomass, root biomass and root shoot ratio to cropping system. This data is from mother trials (n=39). Means are followed by standard errors ................................ .......... 52 ix Table 13 . One - way ANOVA and mean response of pigeonpea stems, twigs, leaves, pods, initial leaf fall, total shoot biomass, root biomass and root shoot ratio to soil texture. This data is from baby trials (n=40). Means are followed by standard errors. ................................ ......................... 54 Table 14 . Correlation matrix of explanatory variables in mother trials. ................................ ...... 55 Table 15 . Correlation matrix of explanatory variables in baby trials ................................ ........... 56 x LIST OF FIGURES Figure 1 . Map of Malawi showing the central region and the specific locations of mother and baby trials from the Linthipe site. ................................ ................................ ................................ .......... 57 Figure 2 . Map of Malawi showing the central region and specific locations of mother and baby trials for the Golomoti site. ................................ ................................ ................................ ........... 58 Figure 3 . Map of Malawi showing the central region and specific locations of mother and baby trials for the Kandeu site. ................................ ................................ ................................ .............. 59 Figure 4 . Map of Malawi showing temperature ranges across the country and the specific locations for which monthly temperature is presented for the three a gro - ecologies. ................................ ... 60 Figure 5 . Frequencies of soil types in baby trials for the 0 20 cm depth ................................ .. 61 Figure 6 . Shoot biomass in four cropping systems from mother trials. Total shoot biomass was separated into stems, twigs, leaves and pods initial leaf fall. Total litter that was collected over ten weeks from litter traps is also included in shoot biomass. ................................ ............................ 62 Figure 7 . Shoot biomass from light and heavy textured soils from the 0 20 cm depth in mother trials. Total shoot biomass was separated into stems, twigs, le aves and pods. Initial leaf fall is also included in shoot biomass. ................................ ................................ ................................ ............ 63 Figure 8. Shoot biomass averaged by sites with light and heavy textured soils, from baby trials. Total shoot biomass was separated into stems, twigs, leaves and pods. Initial leaf fall is also included in shoot biomass ................................ ................................ ................................ ............. 64 Figure 9 . Root biomass from the 0 20, 20 40 and 40 60 cm depths in four cropping systems from mother trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. ................................ ................................ ................................ .................. 65 Figure 10 . Root biomass from the 0 20, 20 40 and 40 60 cm depths, grouped by light and heavy textured soils from mother trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. ................................ ................................ .............. 66 Figure 11 . Root biomass from the 0 20, 20 40 and 40 60 cm depths in light and heavy textured soils from baby trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. ................................ ................................ ................................ ....... 67 Figure 12 . Soil NH 4 + - N by cropping systems from the 0 60 cm depths in mother trials. Error bars represent ± standard errors of treatment means. ................................ ................................ ... 68 xi Figure 13 . Soil NH 4 + - N by soil texture in baby trials from the 0 60 cm depths. Error bars represent ± standard errors of treatment means ................................ ................................ ............ 69 xii KEY TO ABBREVIATIONS BNF: Biological nitrogen fixation C: Carbon C/N Ratio: Carbon - to - nitrogen ratio EPA: Extension planning area GPS: Global positioning s ystem N: Nitrogen NH 4 + - N : A mmonium NO 3 - - N : N itrate PMN: Potentially mineralizable N SOC: Soil organic carbon SSA: Sub - Saharan Africa 1 CHAPTER ONE: LITERATURE REVIEW 1.1.a Background The maize ( Zea mays L.) based systems in Sub - Saharan Africa (SSA) are generally nutrient - depleted and farmers add nutrients that are inadequate for maintaining soil fertility (Vitousek, 2009). Legume rotations are used to maintain soil fertility in some cropping systems . However, synthetic nitrogen (N) fertilizers are an alternative to labor - and land - intensive legume rotations (Crews and Peoples, 2004) but, the economic challenges that small holder farmers face in the region lead to low inorganic fertilizer inputs. Sma ll holder farmers in Malawi are no exception to this phenomenon. A large gap in the literature is the extent to which leguminous crops grown on smallholder farms can he lp address the need for N inputs t hrough above a nd belowground biomass . To address this missing information , studying the productivity of l ong duration legume crops such a s pigeonpea ( Cajanus cajan L . ) under on - farm environmental conditions in Southern Africa is vital . 1. 1 . b Malawi farming systems Malawi has high population densities and very small land holdings (Orr and Ritchie 2004; contributes 35% of the gross domestic product (Ngwira et al ., 2012). However, th ere is a high degree of seasonality due to the unimodal rainfall pattern and frequent dry spells (Doward, 1999). Maize - based farming systems dominate the agricultural sector, particularly over the last century as maize has become the staple crop. The crop occupies about two - third s of the land area under Ngwira et 2 al ., 2012 ). Continuous maize cropping with limited fertilizers has led to poor soil health and low yields (Ngw ira et al ., 2012; Doward, 1999). Maize y ields can be as low as 0. 8 M g ha - 1 and 1.8 M g ha - 1 for local and hybrid varieties respectively (Orr and Ritchie, 2004). Other challenges to improving soil quality and crop productivity i nclude the hand - hoe agricultural system which involves form ing ridges every year, incorporati ng all crop residues and leaving soil bare after harvest (Ngwira et al ., 2012). Soil erosion is aggravated by this exposure of bare soil during the dry season and early rains, and a shortage of natural, organic sources of soil carbon (C) and N (Orr and Ritchie, 2004). Farm outputs are usually low for most smallholder farmers in Malawi, who have very limited access to both formal and informal credit sources (Doward, 1999). According to Ngwira et al . (2012), smallholders have limited access to adequate amounts of farm inputs such as fertilizer and improved seed due to low purchasing power and weak value chains. This holds true for many farmers even with the Farm Input Suppor t Programme (FISP) being implemented in Malawi. However, according to Orr and Ritchie (2004), Malawian farmers still prioritize soil fertility. Sustainable production practices that are resource - efficient are clearly needed in Malawi smallholder farming systems (Ngwira et al ., 2012). Sustainable intensification applies agro - ecological principles which can be implemented in SSA so as to produce more food from less land . This involves more efficient use of natural resources with minimal impact on the envir onment in order to meet the growing population demands (Ngwira et al ., 2012). The continuous monocropped maize systems in Malawi are not sustainable they degrade the natural base, and are labor and time intensive. Therefore, soil health cannot be improved with the use of inorganic fertilizes only. Legumes such as pigeonpea that biologically fix N from the atmosphere have to be included in these cropping systems. 3 1.1 . c Legumes Legumes have been promoted as a means to harness biological processes for N inp ut s that are less costly and not dependent on finite fossil - fuel reserves. Indeed, technologies that enhance reliance on biological N fixation (BNF) and nutrient cycling by legumes are a key sustainable option for maize production in SSA (Snapp et al ., 199 8). Grain legumes have been seen as less important and have received less emphasis than cereal grains, yet they are vital for food, feed and industrial purposes (Sinclair and Vadez, 2012). They have a high protein content compared to other crops and are a primary source of protein in human diets. Legumes are usually intercropped nd often not counted in government census data (Broughton, 2002). According to Mhango et al . (2012) smallholder farmers i n Malawi prefer high - yielding legumes with edible seed. Some legumes such as pigeonpea and cowpea ( Vigna unguiculata L . ) are valuable because both their leaves and grain are be a source of protein for farmers (Broughton, 2002). At the same time, legumes are a major source of N through BNF in most developing countries where synthetic fertilizers are expensive and inaccessible to many (Giller and Cadisch, 1995). Pypers et al . (2007) reported that if legumes are included in rotations they have positive soil - microbiological effects that promote maize growth and production. However, a number of obstacles prevent small holder farmers from producing more legumes even when they are willing to do so. 1. 1 . d Constraints to legume production Most legume seeds have a short storage life which is a challenge to most small holder farmers because they are resource constrained (Sinclair and Vadez, 2012). With the exception of soyabean ( Glycine max L . ), there is little demand for legumes in international trade (Broughton, 20 02). Legume seeds are more expensive than cereal grains and some legumes such as groundnuts 4 ( Arachis hypogaea L.) may have higher labor requirements (Snapp et al ., 2002). Most legumes are vulnerable to diseases and drought and their high protein and oil co ntent makes them susceptible to pests (Broughton, 2002; Sinclair and Vadez, 2012). For example, d espite considerable effort to develop pest resistant pigeonpea, new cultivars can be highly susceptible to pests, as was observed in on - farm trials conducted i n Malawi and in Kenya (Ritchie et al ., 2000). Goat and cattle destruction of pigeonpea is another constraint to adoption of pigeonpea in central Malawi (Snapp et al ., 2002). Pigeonpea is a valued fodder source in some locales, but grain production becomes problematic if livestock are not intensively supervised. Livestock control issues are frequently debated among communities that are experimenting with growing long - duration pigeonpea. Trade - offs can occur between use of pigeonpea as a fodder and a food sou rce. In on - farm trials , in areas new to pig eonpea production, the crop frequently survives for a mont h or two after the rains, but goats or cattle graze it just before harvest. Therefore, small holder farmers allocate little or no resources to legumes and invest more in cereal crops (Zingore et al , 2007). Field survey data from northern Malawi indicate that farmers allocate only about 10 to 15% of cropland to food legumes, and almost none to agroforestry legumes (Mhango et al , 2012). This is due to the ext reme land constraints, with typic ally one ha or less per farmer (FAO, 2012). There are also constraints in terms of legume species farmers can grow, both environmental resource constraints and market related barriers. Options for legumes suitable for the h ighly eroded sandy soils in SSA are limited (Chikowo et al ., 2004). For example, groundnut is productive in sandy soils but not common bean ( Phaseolus vulgaris L . ) . Access to the necessary infrastructure and markets limit small holder farmers, who specialize in producing traditional legume varieties (Broughton, 2002). However, there is growing evidence that long duration legumes such as pigeonpea have a lot to offer small 5 holder farmers in SSA, even with the social and ecological constraints involve d with legume production . 1.1.e Long duration legumes Long duration legumes have the potential to improve both N and P sustainability in small holder farming systems in Africa (Snapp, 1998). Indeterminate, long duration legumes fix more N than early durati on legumes and provide another option for small holder farmers to use for improving soil fertility (Snapp and Silim, 2002). Perennial legumes enhance uptake of soil N through extended growth at a time when annual crops are being re - established (Crews and P eoples, 2004). They have the potential to reduce soil erosion on marginal lands (Giller and Cadisch, 1995). According to Snapp and Silim (2002) , long duration legumes may minimize erosion by the continuous soil cover they provide by their leaf biomass. Lon g duration legumes produce higher quality residues than short duration the field for BNF. They are tolerant to pests, drought stress and poor soil fertility but are usually associated with low to moderate yields, high labor demand and a longer waiting time for harvest (Snapp and Silim, 2002). There are tradeoffs, but perennial legumes have a unique role in terms of provision of ecosystem services to farmers c ompared to short duration annual s. This is especially true for agroforestry legume species. Agroforestry species show considerable potential to address environmental services; however, they do not all provide food, fodder and fuel, and this makes them unattractive to farmers. Smallholder farmer s that are often food insecure necessarily prioritize crops that produce food (Mhango et al . , 2013). There are viney and shrubby long duration legume crops that show potential to be multi - purpose, producing food as well as leaf and root residue inputs tha t can ameliorate soils . Pigeonpea is one such crop. 6 1.1.f Pigeonpea There is considerable evidence that peninsular India is the place where pigeonpea originated ( Nene et al . , 1990 ). This is highly likely due to the presence of several wild relatives, the l arge diversity of the crop genetic pool, ample linguistic evidence, some archaeological remains, and the wide usage in daily cuisine. The name "pigeonpea" probably originated in the Americas, where it reached sometime in the 15th Century, because the seeds were found to be favored by pigeons. Therefore, it may be satisfactorily concluded that pigeonpea originated in India and spread quite early. It is now widely grown in the Indian subcontinent. Other regions where pigeonpea is grown are Southeast Asia, Afr ica, and the Americas. There is a substantial area of pigeonpea in Kenya, Uganda, and Malawi in Africa, and in the Dominican Republic and Puerto Rico in Central America. In most other countries pigeonpea is grown in small areas and as a backyard crop. A secondary center of diversity of the species is found in eastern Africa and other authors contend that eastern Africa is the center of origin, since it occurs wild in Africa ( Nene et al . , 1990). Pigeonpea is a multi - purpose, tropical grain legume crop gr own under rain fed conditions in the semi - arid tropics ( Snapp et al ., 2003; Nam et al . , 1993). The legume is a semi - perennial whose height can reach up to 4 m. (Snapp et al ., 2003; Nene et al . , 1990). There is potential for pigeonpea to be adopted in many areas of the semi - arid tropics because the crop contains a wide range of maturity groups suitable for various agro - ecologies and cropping systems (Kumar Rao and Dart, 1987). At ICRISAT, Patancheru, India, ten maturity groups ranging from 60 days to more t han 160 days were identified, based on days to 50% flowering (Kumar Rao and Dart, 1987) . According to Snapp et al . (2003), the four main genotype categories are: extra short duration (<105 7 d), short duration (105 145 d), medium (146 199 d), and long - durati on (>200 d) cultivars. Pigeonpea is managed in agricultural systems as an annual or biannual because varieties can have either determinate or indeterminate growth habits but monoculture production of pigeonpea is rare. Pigeonpea grown on - farm, in poor soil s and without inputs, produces highly variable yields from 0.2 to 2.5 Mg ha - 1 grain and from 1.0 to 3.8 Mg ha - 1 of leaves and stems (Snapp et al ., 2003). Pigeonpea has extensive roots that reach to depths of 1 2m, with multiple branches, which increases its drought tolerance ( Nene et al . , 1990). The vigorous root system explores a large soil volume and recycles nutrients below the soil profile. It can form nodules up to a depth of 90 cm (Kumar Rao and Dart, 1987). Further, pigeonpea root exudates have th e ability to solubilize iron - bound phosphorus from some soil types (Snapp et al ., 2003). The pigeonpea root system also improves the soil structure by breaking plough pans ( Nene et al . , 1990), making the legume suitable for production in diverse soil types (Snapp et al ., 2003). That is why pigeonpea is also known as the "biological plough" ( Nene et al . , 1990). Pigeonpea is highly adaptable to a wide range of environments, it thrives in diverse cropping systems and is tolerant to abiotic and biotic stresses (Mhango et al ., 2012; Singh and Jauhar, 2005) . What sets it apart is its ability to produce fuel, fodder, and food on rocky, barren and infertile sites. It can be productive in a wide range of soil types from gravelly stones to heavy clay loams with a high moisture content, if there is no standing water on the soil surface ( Nene et al. , 1990). Farmers in India often allocate pigeonpea to fields with poor soils where other crops are not productive. Pigeonpea can tolerate salinity and alkalinity, but not extremely acidic soils (below a pH of 5.0). It tends to have fewer pest problems in drier areas, but droughts reduce the 8 grain yield. However, the crop has more drought tolerance than man y other grain legumes and is able to maintain vegetative growth during consecutive dry months because of its vigorous taproot system and osmotic adjustment (Snapp et al ., 2003). However, the food crop is underestimated in both research and agricultural sta tistics. 1.1 . g Underestimation of pigeonpea such as beans, peas, and chickpeas in terms of area and production (Snapp et al ., 2003; Nene et al . , 1990). It is used in more diverse ways than other legumes. However, research attention to the semi - perennial legume remains limited and its prominence is underestimated due to a number of reasons. The production systems pigeonpea is usually grown in (int ercrops, boundary markers, household vegetable and as a backyard crop) are often excluded in agricultural statistics. According to Nene et al . (1990), pigeonpea has been recorded as present in 40 African countries yet FAO reports production in five countri es only (Snapp et al ., 2003). Even though pigeonpea is understudied compared to cereals and other legumes, a handful of studies in both developed and characteristics make it a perfect fit in diverse intercropping situations in SSA . 1.1.h Pigeonpea - based cropping systems Pigeonpea is used within complex farming systems around the world, involving intercropping, relay cropping and double cropping (Snapp et al ., 2003). Long - duration pigeonpea cultivars are generally planted simultaneously as an intercrop with a cereal at the beginning of the 9 rainy season. In a unimodal system, cereals are harvested toward the end of the rainy season, and pigeonpea develops rapidly on residual moisture after harvest of the companion crop. The growth habit facilitates soil protection, as the canopy continues to expand for four months in the dry season after other crops are harvested . The living and senescent pigeonpea leaves during that extended period may be the only source of cover in semi - arid agro - ecosystems ( Odion et al ., 2007 ; Snapp et al ., 2003 ). The traditional pigeonpea mixed cropping systems used in the semi - arid to sub - humid tropics make efficient use of available natural resources. Th is addresses the needs of smallholder farmers to grow crops that provide stable returns while being ecologically sustainable ( Nene et al . , 1990). Landraces, traditional cultivars, and other long - duration, indeterminate types of pigeonpea have an ideal phe notype for intercrop production slow initial growth and a deep rooting habit (Snapp et al ., 2003; Snapp, 1998 ). This limits competition within an intercrop system. Branching of the pigeonpea shoot occurs late in the season, after harvesting of the cereal crop ( Nene et al . , 1990 ). In dryer areas, and especially in coarser - textured, infertile soils, farmers use wide spacing between plants to limit competition for soil moisture and nutrients. Studies by Snapp and Silim (2002) showed that smallholder farmers in SSA are interested in sustainable intensification systems involving pigeonpea. These systems have been reported to increase calorie production and yield of pigeonpea - based intercrops in on - farm trials in Malawi and Kenya. Pigeonpea significantly contrib utes to soil nutrie nt cycling particularly N and phosphorous cycling. Sinclair and Vadez (2012) reported that pigeonpea thrived for a month after sowing while four other, shallow rooted crop s pecies died from P deficiency. Earli er studies are also in agre ement with those findings as Ae and others (1990) found that pigeonpea enhanced P uptake by sorghum in an intercropping system. 10 A relatively new technology in Malawi is a legume/legume intercrop, which is known as - India but are relatively new in SSA. Farmers in Malawi are experimenting w ith pigeonpea/soyabean or pigeonpea/groundnut intercrops. Combining a short and long - season legume increases yields, woody stems for fuel wood and produces more high quality residues for soil fertility enhancement (Snapp et al ., 2003). Intercropping pigeon pea with cereals in semiarid regions with weathered soils has been proven to decrease soil erosion while increasing available N and P (Vance, 2001). Most small holder farmers prefer intercrops because they produce higher yields per unit area, minimize pest s, improve the N economy (in l egume associations), reduce risks and equally spread the farm resources ( Nene et al . , 1990). There is great potential for pigeonpea production to be economically beneficial to t continues to develop (Snapp et al ., 2003). Pigeonpea is endowed with several, unique characteristics, giving it an important place in the farming systems adopted by smallholder farmers in developing countries. The pigeonpea crop does not require special land preparation or harvesting. Deep ploughing to a depth of 15 cm is sufficient to obtain a good crop and farmers commonly hand harvest the grain ( Nene et al . , 1990). Pigeonpea is grown as an intercrop with maize in southern Malawi, where it accounts for 20% of household income among poor farmers (Orr et al ., 2000). It is also a constituent of boundary plan tings and vegetable gardens in n orthern Malawi . However, pigeonpea is rarely seen in c entral Malawi, even though the crop can be used in various ways in the small holder farming context. 11 1.1 . i Importance and uses of pigeonpea Pigeonpea is important for both human and soil nutrition. According to Snapp and others (2003), it has a high protein content (20 to 32%). More than 80 percent of the world's pigeo npea is produced and consumed in India where it is widely grown for dal (processed, dehulled, split seed) and immature, green pods are a vegetable source. Crushed grain is used for animal feed ( Nene et al . , 1990) and leaves for fodder and soil improvement (Snapp et al ., 2003). Pigeonpea leaves are an excellent source of fodder due to the high N content and low lignin levels. On smallholder farms in Malawi, goats and cattle preferentially feed on pigeonpea residues (Snapp et al ., 2003). Ten tonnes of dry pig eonpea sticks per hectare can be obtained. The stems are a source of fuel wood and are used to make huts, fences and baskets. Around small farms the shrubby legume is often used as a live fence. In India, pigeonpea plants are also used to culture the lac - p roducing insect. On mountain slopes, pigeonpea is cultivated t o reduce soil erosion. It is used as a green manure in some countries, as windbreak hedge, and shade for tree crops like vanilla ( Nene et al . , 1990). Farmers in eastern and southern Africa produ ce pigeonpea as a vegetable or export grain crop, which is intercropped with maize and other cereals or high value crops, such as tomatoes (Snapp et al ., 2003). 1.1.j Nutrient budgets To understand the impact of legumes on nutrient status of smallholder farms in Africa , nutrient budgets could provide important information. Nutrient budgets with cropping systems that involve pigeonpea in some studies . An experiment in India, where maize followed pigeonpea, residual N was estimated to be approximately 40 kg ha - 1 (Kumar Rao et al., 1981) . Kumar Rao and Dart (1987) conducted a comparative study on nodulation, N fixation and N uptake by 12 pigeonpea from different maturity groups. They reported that long duration varieties produced up to 11 Mg ha - 1 , while the ear ly duration varieties produced as little as 4 Mg ha - 1 .The study gave further insight on the importance of deep - rooted legumes in nutrient cycling. The researchers concluded that among all the pigeonpea maturity groups the total N uptake from the system va ried from 69 to 134 kg ha - 1 , which may otherwise have leached from the system. The long duration varieties fixed more N than the early duration varieties (Kuma Rao and Dart, 1987). In another study, Kuma Rao and others (1994), reported that pigeonpea fixed between 58 to 88 N kg ha - 1 when grown as a sole crop and from 30 to 50 N kg ha - 1 of residual N in crop rotations with maize, wheat and sorghum. Khan et al . (2002) found that pigeonpea had a higher belowground N value than fababean and mungbean, in yet ano ther study in India . Over the years, farmers in India have adopted pigeonpea for various reasons. Bantilan and Darthasarathy (1999) reported that the majority of farmers who quickly adopted pigeonpea in n orthern India did so mainly to improve soil fertility. The farmers reported a number of benefits that resulted from pigeonpea cultivation. The amount of inorganic fertilizer required for subsequent crops declined, soil structure improved, land preparation was easier, and other crops germinated well. According to Myaka et al . (2006), p igeonpea increased the recirculation of dry matter, N and P on on - farm trials at multiple sites in Malawi and Tanzania. Total soil C , total N and inorganic N were not a ffected by two seasons of pigeonpea /maize but were negatively affected by sole maiz e systems. They found the pigeonpea /maize intercrops to be sustainable, low risk and they mitigated topsoil nutrient leaching (Myaka et al, 2006). Adu - Gyamfi et al . (2007) conducte d 2 - yr , farmer - managed trials in Malawi and Tanzania, where they calculated N and P budgets of sole maize and maize - pigeonpea intercrops. The study consisted of two sites in each country and 20 farmers per site. They discovered that in Malawi, pigeonpea c ontributed about 38 117 kg ha - 1 of N through 13 BNF wh ile in Tanzania it was about 6 72 kg ha - 1 . The authors provided evidence that incorporating pigeon pea into the soil improves N budgets. In Malawi the N budgets of sole maize were negative and about fiv e times lower than those of the maize/pigeonpea intercrops which were positive (30.5 k g ha - 1 ). This was achieved by incorporating the whole aboveground pigeonpea biomass (excluding grain) into the soil. However, in Tanzania, N budgets in both intercrop and sole maize fields were negative but sole maize systems were six times more negative than the intercrops (Adu - Gyamfi et al., 2007). Based on these studies in Malawi, soil N and P budgets improved when maize was intercropped with long duration pigeonpea. S napp and Silim (2001), reported that pigeonpea residues provided 30 to 70 N kg ha - 1 and were particularly suited to the resource base of smallholders in a 3 - yr study at 40 farm sites in Malawi. After two years of intercrop ping or rotation s with pigeonpea, mai ze yields increased by 0.3 1.6 Mg ha - 1 compared with sole - cropped maize (Snapp and Silim, 2001). Other studies in Malawi and Benin have shown that after a pigeonpea fallow, maize grain yields reflect a n N fertilizer equivalency of 50 kg ha - 1 (McCall, 1989). Yield enhancement of cereals after a pigeonpea fallow has also been observed in Kenya and Cameroon (Degrande, 2001; Onim et al ., 1990). 1.1.k Research gaps Legumes are widely grown and their importance is known but they are still understudied compared to cereals. A substantial number of studies on N contributions by legumes are available but most of them were carried out on research stations , where soil condi tions are often more favorable for growth than on fields . Long duration pigeonpea, a multi - purpose legume that is indeterminate in growth habit (8 or more months) is also understudied compared to rapid 14 cycling legume s (3 to 4 month s ), or agrofore stry species (multiple years) . Vitousek et al . (2009), carried out farm budgets in Kenyan and Chinese farming systems with zero contribution from legumes , an unrealistic approach showing that the role of legumes in nutrient cycling is often underestimated. The contribution of belowground legume biomass and its role in nutrient cycling is another key unknown factor. roots are very difficult to measure. The metho ds used to excavate roots are often time consuming, expensive and laborious. Taken together , there is urgent ne ed to conduct participatory on - farm research with smallholder farmers to investigate the overall productivity of aboveground and belowground biom ass of pigeonpea. 15 CHAPTER TWO: BELOW AND ABOVEGROUND PIGEONPEA PRODUCTIVITY IN ON - FARM SOLE AND INTERCROP SYSTEMS IN CENTRAL MALAWI. 2.1 INTRODUCTION Agro - ecological intensification improves the performance of agriculture through integration of ecological principles into farm and cropping s ystem management. Examples of agro - ecological principles include the biological conservation of resources such as s oil, nutrients and water. Increas ed use efficiency of limited nutrient and land resources is key to viable farming systems in densely populated African countries. Smallholder farmers face a rapidly changing physical environment in SSA. A degraded soil res ource base, climate change and a variable market environment are some of the challenges, and opportunities, faced by farmers (Mhango et al , 2012). The u se of legumes such as pigeonpea, which biologically fix N and provide high protein grain while recycl in g nutrients , is one approach that has shown promise in Malawi. Pigeonpea is used within mixed cropping systems around the world, such as intercropping, relay cropping and double cropping (Snapp et al ., 2003). L iving and senescent pigeonpea dry matter minim ize s soil erosion and may be the only source of cover in semi - arid agro - ecosystems (Odion et al ., 2007 ; Snapp et al., 2003; Snapp and Silim (2002) ; Giller and Cadisch, 1995 ). However, pigeonpea is not central Malawi include lack of intensive community management of animal grazing (Snapp et al ., 2002) , access to affordable seed and h igh susceptibility to pests for some cultivars (Ritchie et al ., 2000). Animal grazing is a major constraint in pigeonpea production as maize, the main crop is harvested months before pigeonpea matures. Community management of animal grazing is only effecte d till maize harvest. 16 Legumes are widely grown worldwide an d their importance on biological nitrogen fixation is known but are still understudied compared to cereals. Many legume studies are done on research stations . Knowledge of small holder farming systems is limited, and few studies take into account farmer decision making (Mhango et al ., 2012) . Long duration pigeonpea, a potential multi - purpose legume is understu d ied compared to other short duration legumes such as soyabean . E mpirical data on legume root and shoot biomass additions to the systems have also remained scarce (Myaka et al ., 2006). Additionally, r oots are understudied because the methods used to quantify them are time consuming and very laborious. A handful of studies on contribution of pigeonpea shoot biomass to N cycling have been done in Malawi, Tanzania and India. Adu - Gyamfi et al . (2007) quantified N exports from aboveground biomass of pigeo npea from farmer - managed pigeonpea/maize intercrops in Malawi. However, the authors did not have any sole pigeonpea or doubled - up legume cropping systems involving pigeonpea. The authors also did not quantify root biomass as well. Myaka et al . (2006) quant ified only the shoot biomass of pigeonpea in a pigeonpea/maize intercrop in on - farm trials in Malawi and Tanzania. Kumar Rao and Dart (1987) quantified both shoot and root biomass of pigeonpea but only from sole pigeonpea of different maturity groups and n o intercrops. The benefits of legumes on soil N and organic matter depend on the quality and quantity of both below and aboveground biomass that is returned to the soil. The quality of crop residues vary with legume species, soil nutrient levels, plant d ensity, planting type and field management practices (Reddy et al ., 2003). Quantifying the root and shoot biomass of pigeonpea in on - farm trials with various intercrops will help farmers in choosing the best cropping system especially if their goal is to i mprove soil fertility , particularly soil N . 17 This chapter seeks to fill these research gaps by quantifying aboveground and belowground biomass of pigeonpea in on - farm trials in central Malawi. Our findings are from research that was conducted on fields in a rain - fed system . The aim of this study is to improve the understanding of pigeonpea productivity in different cropping systems , to help farmers design more resource efficient cropping systems that suit fa r needs. Identifying combinations o f crops that can efficiently cycle or increase the availability of soil nutrients is relevant in the small holder farming context. On - farm, participatory research trials were set - up across three agro - ecologies in central Malawi, during the 2013/14 cropp ing season. Pigeonpea was planted as a sole crop or in an additive intercrop system with soyabean or groundnut . A pigeonpea /maize intercrop was also included. The objectives of this study were to: Assess the effect of the type of cropping system on pige onpea root and shoot biomass and other soil properties total soil N %, SOC %, inorganic N and PMN Determine the effect of soil texture on pigeonpea root and shoot biomass Evaluate variability (shoot and root biomass) of pigeonpea growth in different crop ping systems within a smallholder farm context We hypothesized that: Cropping system would have a significant effect on both root and shoot b iomass of pigeonpea and that biomass would be highest in the sole pigeonpea cropping system and lowest in the pig eonpea/maize intercrop 18 Heavy textured soils would have higher root and shoot biomass than light textured soils The pigeonpea/maize intercrop would have the highest variability compared to other cropping systems. 19 2.2 MATERIALS AND METHODS 2.2 . 1 Site description This study was conducted as part of the Africa Research in Sustainable Intensification for the Next Generation (Africa RISING), Malawi project. The participatory, research - for - - based agricul tural systems in SSA through sustainable technologies. There were four research sites from three different agro - ecological zones in central Malawi. The locations were Linthipe and Golomoti Extension Planning Areas (EPA) in the Dedza District and the Kandeu EPA in Ntcheu District. Dedza and Ntcheu districts are in the Lilongwe Agricultural Development Division (LADD). Linthipe is a high potential, sub - humid tropical site with well distributed rainfall in most years (Tables 1 and 2) . Kandeu is a medium potent ial, sub - humid tropical site. Golomoti is a low potential, semi - arid to sub - humid tropical site located at low altitude and with erratic rainfall. Soils at Linthipe are ferric luvisols, the Kandeu study site has a mix of chromic luvisols and orthic ferrals ols while Golomoti soils are a mix of eutric cambisols and eutric fluvisols (Lowole, 1984) . The GPS locations for each site are shown in Figures 1, 2 and 3. The 2013/14 growing season precipitation and temperatures, as well as historical 15 - yr monthly precipitation and temperature averages for three sites are presented in Tables 1 and 2 . The edaphic properties of the three sites are shown in Tables 3 and 4. 2.2 . 2 Experimental Design et al. , 2002) was used. It involved trials that test a full complement of technologies that are located on - farm a t central locations, with three replicates per trial, in a randomized complete block design the mother trials. The mother trials at each site were linked systematically with a cluster of 20 30 baby 20 trials. These are a type of on - farm trials where farmers choose a subset of technologies from the mother trial to test, where replication occurs across sites. In this study we focus on both mother and baby trials. 2.2 . 2 . a Mother trials Four cropping systems were tested. They consisted of pigeonpea grown as sol e crops , and three intercrop systems involving pigeonpea grown in an additive intercrop system with maize, groundnut and soy a bean. Intercropping two compatible grain legumes is known as the doubled - up legume technology. Many doubled - up legume cropping syst ems in Malawi involve pigeonpea, because the success of the approach hinges on the initially slow growth of pigeonpea, facilitating the growth of companion crops as if sole - cropped. The plot sizes were 5 m x 5 m and ridges were spaced at 0.75 m. 2.2 . 2 . b Af rica RISING Baby Trials The cropping systems from the Africa RISING baby trials and the number of farmers that adapted or adopted a particula r system differed. A total of 30 farmers adopted similar cropping systems from the mother trials while 10 cultivated pigeonpea around the borders of their fields a treatment referred to as pigeonpea/borders. Eleven farme rs had pigeonpea/groundnut, t en had pigeonpea/borders , eight had pigeonpea/maize, seven had sole pigeonpea and only four had pigeonpea/soyab ean. The management of baby trials varied with each farmer. Therefore, short interviews on nutrient management were conducted with each of the 40 farmers (data not shown) . Aboveground biomass, belowground biomass and soils samples were collected from all t he 40 baby trials. 21 2.2.3. Agronomy The variety of pigeonpea that was planted at the three sites is locally known as Mwaiwathu alimi ( pedigree ICEAP 00557 ) . It is a medium maturing (180 d) variety with a yield potential of about 2 .5 M g ha - 1 . According to Tropical Legumes II (2013), the variety is adapted to low to medium altitude areas. The varieties for the other crops pigeonpea was intercropped with are shown in Table 5 . Field experiments were set up in the 2013/14 growing season, with crops planted during December of 2013 (Table 5 ). Sole pigeonpea was planted at a spacing of 0.9 m x 0.75 m, with three plants per planting station to achieve a plant population density of 44 000 per ha. In all the - and groundnut or soy a bean were planted in the space between the pigeonpea plants, using an additive intercropping design . The plant population density of soya beans and groundnuts in the intercrops were 160 000 and 120 000 plants ha - 1 , respectively. The ra tio of pigeonpea to groundnut plants was about 1:3 and that of pigeonpea to soya beans was about 1:4 . In the pigeonpea/maize cropping system, both maize and pigeonpea were planted at a spacing of 0.9 m x 0.75 m in an additive design so pigeonpea population density was the same as sole pigeonpea at 44 000 plants ha - 1 . The combined pigeonpea and maize plant population density was 88 0 00 plants ha - 1 (Table 6 ) . All the intercrops were planted in the same row. Seeds were planted after the first effective rains, and all plots were planted on the same day at each site. Planting dates for all mot h er trials were recorded (Table 7 ). Weed and fertility man agement followed the Malawi agricultural recommendations (Malawi Guide to Agriculture, Government of Malawi, 2010). The plots were weeded by hand hoe three times at each site. The doubled - up legume intercrops (pigeonpea/groundnut and 22 pigeonpea/soybean) wer e fertilized just before planting with 23:21 N: P compound fertilizer at the rate of 11.5 kg ha - 1 N and 10.5 kg ha - 1 of P . The pigeonpea/maize intercrop was fertilized at the rate of 23 kg ha - 1 N and 21 kg ha - 1 P , with a side dress application of Urea at 100 kg ha - 1 which provided 46 kg ha - 1 N. 2.2.4 Rainfall and temperature The sources for the rainfall data for all sites were the Tropical Rainfall Measurement Mission Project (TRMM) (2015) and the Earth Observing System Data and Information System (EOSD IS) ( 2009 ) . Temperature data was sourced from NASA Land Processes Distributed Active Archive Center (LP DAAC ) ( 2010 ) . 2.2.5 Aboveground biomass assessment Wooden litter traps of 30 x 30 cm were placed in all treatments at the onset of pigeonpea flowering on 13 May 2014. The litter biomass was collected and weighed every fortnight, until senescence in mid - August. The leaves from the litter traps from each treatment were combined and ground using a Wiley laboratory mill (Thomas® Model 4 Wiley Mill, Swedesboro, NJ) . S ix months after planting, destructive sampling of three randomly selected plants per plot was conducted at peak flowering. The plants were cut at ground level, chopped, and fresh biomass was determined. The plant samples were oven - dried a t 75 °C to constant weight, and dry weighs recorded. The dry aboveground biomass was separated into stems, twigs, leaves, and pods. The 23 biomass was reported as Mg ha - 1 , for fresh and dry weights. The separate biomass components were ground to pass a 1 - mm s ieve with a Wiley laboratory mill and samples sent to the University of California Davis Stable Isotope Facility for natural abundance isotopic N analysis (data not shown) . 2.2.6 Belowground biomass assessment The method used for destructive root samplin g was similar to that of Taylor (1986). Pigeonpea plants were cut at ground leve l and an area measuring 45 cm x 37 . 5 c m was marked at the base of the plant to hand - dig a pit to a depth of 6 0 c m. Soil and roots were removed in three increments (0 2 0 cm, 2 0 4 0 cm and 4 0 6 0 c m). Large roots were removed from soils while dry sieving with a 2 - mm sieve. F ine roots were hand - picked using tweezers. The belowground biomass was separated into surface and deep roots. The roots were weighed fresh, oven - dried at 75 o C to constant weight and reweighed. They were ground with a 1 - mm sieve Wiley laboratory mill and analyzed for total N concentration (data not shown) . 2.2 . 7 Soil Sampling and analyses Soil samples were collected from each plot at all sites in increments of 0 2 0 cm, 2 0 4 0 cm and 4 0 6 0 c m . D uring root excavations, all the soil from a volume of 0.135 m 3 per layer was spread on plastic sheets. After all the roots were removed, the soil w as mixed thoroughly and composite samples of about 2 kg were collected. The samples were air - dried for 48 h and sieved through a 2 - mm sieve at the Lilongwe University of Agriculture and Natural Resources 24 (LUANAR) soil science laboratory. After sieving, roc ks and large pieces of organic matter were discarded. 2.2.7.a Soil texture Soil texture was determined using the hydrometer method (Kellogg Biological Station LTER, 2008) at LUANAR. Soil subsample s were shipped to Michigan State University (MSU) for further physical and chemical analysis. 2.2.7.b Soil pH The 1:5 soil: water suspension method (Department of Su stainable Natural Resources, 2013) was used to determine soil pH, and a Metler Toledo SevenEasy S20 pH meter was used. 2.2.7.c Inorganic N Sub - samples of 10 g were weighed into a 100 mL p lastic centrifuge cups and 40 mL of 2M KCl was added. The cups we re shaken for one hour on a reciprocal shaker at approx. 180 strokes per minute. After shaking, cups were allowed to settle for 15 mins. The supernatant was filtered through a Whatman No. 1 filter paper (GE Healthcare Bio - Sciences, Pittsburg, PA) and poure d into small plastic vials. The samples were frozen until they were analyzed. KCl extractant was analyzed for inorganic N ( NO 3 - - N and NH 4 + - N ) concentrations using the colormetric method described in Doane and Horwath (2003) , and a Thermo Multiskan TM 96 - well plate reader (Kane et al . , 2015). 25 2.2.7.d Potentially mineralizable N (PMN) Sub - samples of 10 g were weighed into a 100 m L plastic centrifuge cups and 10mL of distil led water was added. Anaerobic conditions were created by adding N 2 gas and the samples were incubated at 30 °C for seven days. After incubation, NH 4 + - N was extracted from samples by adding 30 mL of 2.66 M KCl to each sample, effectively bringing the mola rity of the sample solution to 2 M KCl. Extracts were then analyzed for NH 4 + - N only since the anaerobic condition created during the incubation inhibits nitrification. The initial NH 4 + - N concentration of each sample was then subtracted from the concentratio n of the corresponding incubated sample to determine the amount mineralized during incubation. The samples were frozen until they were analyzed as described above. 2.2.7.e Total soil N and S oil O rganic C arbon (SOC) percent Soil subsamples were ground to pa ss a 1 - mm sieve in a shatter mill. The dried, ground soils were weighed on a microbalance into tin capsules for analysis and total C/N content was determined by combustion. A Carlo Erba NA1500 SeriesII Combustion Analyzer (Kellogg Biological Station LTER, 2003) was used. 2.2.7 Biological N fixation The amount of N biologically fixed by pigeonpea was determined using the natural abundance method. The samples were sub sampled and weighed into capsules before the stable isotope analysis was conducted at the UC Davis stable isotope facility. Maize and a local non - N - (data not shown) . 26 2.2.8 Statistical analysis Data were analyzed using the MIXED and CORR procedure in SAS 9.4 ( SAS Institute, 2002 ) statistical package. A one - way analysis of variance was carried out to assess effect of cropping system on pigeonpea shoot and root biomass, where sole - cropped and intercropped pigeonpea w as compared. Soil texture effects on pigeonpea root and shoot biomass were a lso tested using the same procedures. In addition to aboveground and belowground biomass, the amount of senescence biomass , soil N and C cont ent from different cropping systems and soil textures were analyzed by ANOVA. The Least Significant Difference (LSD) at 5% and 10 % level of significance was used to test mean differences. Graphs were created using R version 3.2.1 (Team R, 2015) and SigmaPlot version 13.0 (SigmaPlot, 2015). Maps for the global positioning system coordinates (GPS) and for temperature gradients across Malawi were created using ArcMap (2010). 27 2.3 RESULTS 2.3.1 Study locations and soil characteristics The Linthipe mother trial was located at 1 4 S latitude and 3 4 ° longitude . The location of baby trials from Linthipe ranged from 1 4° 11 to 1 4 S latitude and from 3 4 ° 3 4 ° longitude . The GPS coordinates for the Linthipe mother and baby trials are shown on a map in Figure 1. Golomoti mother trials were both at 1 4° 2 6 S latitude and 3 4 ° 3 4 ° longitude . Baby trials from Golomoti were between 1 4° 2 6 S latitude and 3 4 ° 3 4 ° longitude . The locations of the Golomoti mother and baby trials are presented in Figure 2. The Kandeu mother trial was located at 1 4° 37 S latitude and 3 4 ° longitude . The Kandeu baby trials in this study were located between 1 4° 35 and 14° 37 S latitude and 3 4 ° 3 4 ° longitude (Figure 3). All the mother trials are located in the Dedza and Ntcheu districts of central Malawi. The soil characteristics for all mother trials are shown in Table 3 and for baby tria ls they are presented in Table 4 . 2.3 . 1.a Rainfall and temperature Av erage monthly precipitation and temperature for a 15 - yr period for all study locations are shown in Tables 1 and 2 respectively . The rainfall was unevenly distributed with the total rainy days for Linthipe being 59 and 49 for both Golomoti and Kandeu. Tota l rainfall for the 2013/2014 gro wing season for Linthipe was 979 mm (higher than the 15 - yr average), 848 mm for Golomoti (lower than the 15 yr average) and 909 mm for Kandeu, which was well in its hist orical precipitation range. Averages and ranges for mon thly precipitation and temperature for the 2013/2014 growing season are shown in Tables 1 and 2 . 28 2.3.2 Pigeonpea total shoot biomass The mother trial data was used to investigate the effect of cropping system and of soil properties on pigeonpea biomass ( Table s 8 , 10 and 12 ). Response of pigeonpea biomass could only be assessed for soil texture in ba by trials due to no replication within a site, and not all cropping systems being represented at each baby trial site. Therefore, the statistical analysis of d ata from baby trials focused on effects of soil texture ( an important determinant of plant growth ) on pigeonpea biomass. Cropping system had the following effect on shoot biomass: the largest amount was accumulated in sole pigeonpea ( 11.83 Mg ha - 1 ± 1.2 ) , and doubled up pigeonpea / groundnut intercrop ( 6.99 Mg ha - 1 ± 1.24 ) had comparable amounts of b iomass to sole and to pigeonpea/ soyabean intercrop ( 5.08 Mg ha - 1 ± 0.77 ) . Pigeonpea / soyabean biomass w as comparable to pigeonpea /maize ( 3.57 Mg ha - 1 ± 0.33 ) , whic h saw the smallest accumulation overall (Table 12 ). Soil texture was also evaluated for effects on shoot biomass using data from the four mother trials, where response in light - textured soils (mother t rial sites Kandeu and Golomoti - E) was compared to heavy - textured soils (mother trial sites Lin thipe and Golomoti - B) . Total shoot biomass was ( 4.75 Mg ha - 1 ± 0.61 ) in l ight textured soils and ( 8.17 Mg ha - 1 ± 1.48 ) in heavy texture d soils ( p = 0.05 ) . 2.3.2.a Pigeonpea litter Overall, response in the mother trials shows that initial litter biomass averaged 0.28 Mg ha - 1 and accumulated litter biomass was 0.7 Mg ha - 1 (Table 8 ). Cropping system had a significant effect on initial litter and on litter biomass from litter tr aps , and followed the same pattern (highest 29 in sole pigeonpea, intermediate in doubled - up legumes and lowest in the pigeonpea/maize cropping system) observed for total shoot biomass ( P = 0.0005 , Table 12). Initial litter ranged from 0.002 Mg ha - 1 in a pig eonpea/soyabean intercrop to 0.866 Mg ha - 1 in a sole pigeonpea cropping system. The lowest pigeonpea litter biomass of 0.01 Mg ha - 1 was from the pigeonpea/soyabean intercrop and the highest litter biomass of 2.16 Mg ha - 1 was from the sole pi geonpea cropping system (Table 8 ). The initial leaf fall average for baby trials was 0.34 Mg ha - 1 and values ranged from 0.001 to 1.55 Mg ha - 1 (Table 9 ). Soil texture did not have any effect on initial leaf fall in baby trials ( P = 0.3716 , Table 13 ). 2.3.2.b Ste ms Overall, the average stem biomass in mother trials was 3.27 Mg ha - 1 and it ranged from 0.45 Mg ha - 1 in a pigeonpea/soyabean intercrop to 7.4 Mg ha - 1 in a sole pi geonpea cropping system (Table 8 ). Pigeonpea stem biomass from mother trials was affected by the type of cropping system , in a similar manner to the re s ponse of total shoot ( P =0.0059, Table 12). The overall average stem biomass for baby trials was 3.86 Mg ha - 1 and the lowest stem biomass was at 0.53 Mg ha - 1 , while the highest was 12.32 Mg ha - 1 ( Tab le 9 ). For baby trials, soil texture did not have any significant effect on stem biomass ( P = 0.1119, Table 13). 30 2.3.2.c Twigs The average twig biomass from mother trials was 1.07 Mg ha - 1 and the lowest was 0.05 Mg ha - 1 in a pigeonpea/soyabean intercrop while the highest was 4.37 Mg ha - 1 in a sole pi geonpea cropping system (Table 8 ). There was a significant effect of cropping system on twig biomass in mother trials ( P = 0.0025, Table 12). Soil texture had a significant effect on twig biomass in baby trials ( P =0.0799, Table 13 ). The overall mean for twig biomass from baby trials was 1.03 Mg/ ha - 1 , the minimum was 0.12 Mg ha - 1 and the maximum was 3.62 Mg ha - 1 (Table 9 ). 2.3.2.d Leaves There was a significant effect of cropping system on leaf biomass in mother trials ( P = 0.0387, Table 12 ). Average leaf biomass from mother trials was 0.73 Mg ha - 1 , minimum was 0.02 Mg ha - 1 in a pigeonpea/maize intercrop and the maximum of 2.76 Mg ha - 1 was from a sole pigeonpea cr opping system (Table 8 ). Soil texture did not have a significant effect on leaf biomass in baby trials ( P =0.1546, Table 13 ). The overall average leaf biomass from baby trials was 0.97 Mg ha - 1 and the biomass ranged from 0.002 to 4.87 Mg ha - 1 (Table 9 ). 2.3.2.e Pods The average pod biomass from mother trials was 1.33 Mg ha - 1 , and cropping system had a marked effect, where pod biomass was highly suppressed in a pigeonpea/maize intercrop (0.02 31 Mg ha - 1 ), compared to 3.98 Mg ha - 1 in a sole pigeonpea cropping system (Table 8 , Figure 6 ). This supp ression effect in the pigeonpea/ maize systems was greater than that obser ved for overall shoot biomass (Figure 6 ). In contrast to the dramatic effect of cropping system, soil texture did not have a n effect on pod biomas s in any trial ( P =0.9655, Table 13 ). The overall pod biomass from baby trials was 0.57 Mg ha - 1 , the lowest biomass was 0.02 Mg ha - 1 and the highest was 2 . 02 Mg ha - 1 (Table 9 ). 2.3.2.f Root biomass Root biomass was largely present in the 0 20 cm depth , as observed for both mother and baby trials across all soil textures present . In mother trials, cropping system had a significant effect on root biomass , which was primarily due to response at the 0 20 cm depth ( P = 0.0 052, Table 12 ). For mother trials, av erage root biomass from the 0 20 cm depth was 882.17 kg ha - 1 , with ranges from 56.00 kg ha - 1 (pigeonpea/soyabean) to 2877.78 kg ha - 1 (sole pigeonpea). Root biomass from the 20 40 cm depth ranged from 0.44 kg ha - 1 to 241.93 kg ha - 1 (both pigeonpea/soyabea n), and the average was 45.28 kg ha - 1 (Table 8 ). The 40 60 cm depth had very little b iomass with ranges from 0.15 kg ha - 1 in a pigeonpea/soyabean intercrop to 43.26 kg ha - 1 in a sole pigeonpea cropping system. Figures 9 and 1 0 show effect of cropping system and soil texture on root biomass by depth for mother trials. In baby trials, soil texture had a significant effect on root biomass from the 0 20 cm depth ( P = 0.0429, Table 13 ) but not on the root biomass from the 20 40 cm depth ( P =0.9398, Table 13 ) or the 40 60 cm depth ( P = 0.7696, T able 13 ). Soil texture had a significant effect on total root 32 biomass of pigeonpea from baby trials ( P =0.0407, Table 13 ). The overall average for root biomass from the 0 20 cm depth was 927.5 2 kg ha - 1 , and the biomass ranged from 31.54 to 3331.11 kg ha - 1 . Root biomass from the 20 40 cm depth ranged from 1.62 to 302.70 kg ha - 1 and the average was 80.99 kg ha - 1 . The 40 60 cm depth had very little biomass compared to the other two depths, with ranges from 0.86 to 147.41 kg ha - 1 and the average was 24.77 kg ha - 1 (Table 9 ). Overall, total root biomass from baby trials had ranges from 0.03 to 3.78 Mg ha - 1 and an average of 1.03 Mg ha - 1 . Figures 11 shows effect of soil texture on root biomass by dep th for baby trials. 2.3.2.g Root shoot ratio There was no significant effect of cropping system on root shoot ratios from the mother trials ( P =0.1728, Table 12 ). Overall, the average root shoot ratio from mother trials was 0.14 and the ranges were from 0.05 (pigeon/maize intercrop) to 0.26 (pigeonpea/groundnut and pigeonpea/maize intercrop). Root shoot ratios for each cropping system from mother trials are shown in Table 12 . Soil texture did not have any significant effect on root shoot ratios of pigeonpea from baby trials ( P =0.9663, Table 13 ). The average root shoot ratio from baby trials was 0.1 7 with ranges from 0.12 to 0.35 (Table 9) . 2.3.2.h Soil pH Soil p H was slightly acidic for both mother and baby trials. The overall soil pH average for mother trials was 5.20 and it had ra nges from 4.51 to 6.35 (Tables 3 ). 33 In baby trials, soil texture did not have any significant effect on pH ( P = 0.6907, Table 13 ). Th e overall soil pH average for baby trials was 5.37 and the pH had ranges from 4.36 to 7.63 (Table 4) . 2.3.2.i Soil texture Clay content increased down the soil profile for both mother and baby trials. A threshold of clay content which was equal to, or gre ater than 20% was set as the heavy textured soils. Soils with a clay content lower than 20% were classified as light textured soils. From the 0 20 cm depth in mother trials, 38% were light textured soils while 62 % were heavy textured. The 20 40 cm depth had 18 % of light textured soils and 82% heavy textured soils. Only 3% were light textured soils in mother trials from the 40 60 cm depth, while 97% were heavy textured soils. Averages of percent silt, percent clay and percent sand by depth for each moth e r trial are presented in Table 3 . Soil texture f rom baby trials varied (Figure 5 ). From the 0 20 cm depth, 52.5 % of baby trials were light textured while 47.5 % were heavy textured. For the 20 40 cm depth 35 % of the baby trials were light textured w hile 65 % were heavy textured. For the 40 60 cm depth, only 27.5 % of baby trials were light textured while were 72.5 % heavy textured. Averages of percent clay and percent sand by depth for baby trials from each location are presente d in Table 4 . 2.3.2 .j Total soil N and SOC As expected, for recently established trials (November, 2012), the c ropping system did not have a n effect on either soil N percent ( P =0.8235 ) or SOC ( P =0.9112). Total soil N ranged from 34 0.01 % in a pigeonpea/groundnut intercrop to 0. 16 % in a pigeonpea/maize intercrop, and the average was 0.07 % . Overall, SOC average for mother trials was 1.01 % with a range of 0.05 to 2.93 % (Table 3 ). Soil texture was associated with soil organic matter in baby trials . C lay % was positively correlated with soil N ( P =0.0287) and SOC ( P =0.001 ), while sand % was negatively correlated to both (Table 15) . The average percent N for baby trials was 0.05 with ranges from 0.004 to 0.17. SOC in baby trials had ranges from 0.007 to 2.52 and an average o f 0.8 (Table 4 ). Correlations between variables from both mother and ba by trials are shown in Tables 14 and 15 respectively. 2.3.2.k NH 4 + - N , Soil NO 3 - - N and PM N Ammonium response is shown in F igure 1 2 , where the average level was quite low at 2.05 mg/kg soil with rang es from 0.41 to 9.38 mg/kg soil in mother trials. The soil NO 3 - - N levels follow ed ammonium levels closely, with the lowest value being unde te ctable and the highest at 6.69 mg/kg soil . The overall NO 3 - - N average was 0.9 mg/kg soil. Potentia l mineralized soil N from moth er trials had an average of 0.45 mg /kg /day soil, with ranges from - 0.19 mg/kg soil /day indicating immobilization, to a high value of 1.96 mg/ kg soil/ day . Figure 13 shows the soil NH 4 + - N for light and heavy textured soils fro m baby trials by soil depth. The soil NH 4 + - N had ranges from 0. 18 4. 14 mg/kg soil, with an average of 1.73 mg/kg soil. Baby trials had a soil nitrate average of 0. 77 mg/kg soil, with ranges from 0.00 8.55 mg/kg soil . M ineralize d N from baby trials had an average of 0.47 mg/kg soil with ranges from - 0.0 9 1. 71 mg/kg soil /day . The results also indicate N immobilization in baby trials. 35 2.4 DISCUSSION 2.4.1 Precipitation Precipitation was variable and unevenly distributed (Table 1 , Figure 4 ). This is expected based on the highly variable weather patterns experienced in this region, which have been ex acerbated in recent years by global climate change as indicated by global climate models which all predict gains in rainfall variability in the coming decades (Burke et al., 2009. Climate variability has a negative effect on crop productivity. Root and shoot biomass variability from our mother and baby trials may have been as a result of rainfall patterns , and heterogeneity in s oil as well as farm management decisions . Other studies have highlighted the variability of the maize - based farming systems in SSA , and vulnerability to rainfall patterns ( Funk et al., 2008) . 2.4.2 Pigeonpea shoot biomass Total shoot biomass of pigeonpea from both baby and mother trials was quite high , in the range of 3.57 to 11.83 Mg ha - 1 ( Tables 12 and 13 ), and highest for sole pigeonpea as hypothesized . Pigeonpea in the intercrop systems produced aboveground biomass similar to previous reports from a m aize - pigeonpea intercrop field study conducted in Tanzania and Malawi (Myaka et al ., 2006) , and higher than the biomass accumulation o bserved in on - farm trials in n orthern Malawi (Mh ango, 2012). The high biomass observed may be due in part to our collection of litter over t e n weeks, as pigeonpea growth pattern includes leaf senescence during the growing season which can complicate measurement of primary productivity and lead to under - estimation. According to our knowledge, t his is the first study that monitored both on - farm root and shoot biomass, and included collection of leaf litter over the growing season. Kumar Rao and Dart ( 1987 ) measured root, shoot 36 biomass and litter of pigeonpea but this was conducted on a research station. Overall the hig hest levels of biomass we observ ed were in the sole pigeonpea (11.83 Mg ha - 1 ) and pigeonp ea - groundnut intercrop (6.99 Mg ha - 1 ) . The total shoot biomass from sole pigeonpea is within the range reported for a long duration cultivar of pigeonpea (220 d ays ) co nducted on a research station in India (Kumar Rao and Dart, 1987). Research stations might have high soil fertility status fields , thus it is interesting that we observed high biomass accumulation potential in this 180 d ays pigeonpea g enotype grown on Malawi smallholder farms. Others have reported highly variabl e pigeonpea biomass from 1.5 Mg ha - 1 to greater than 7 Mg ha - 1 ( Giller et al ., 1997 ). In our study we quantified shoot and root biomass from high plant population density field stands, showing the high growth and yield potential that is possible on - farm but may not be achieved by farmers constrained in access to seed. The negative effect of low plant population density on pigeonpea shoot biomass and N fixation estimates have been observed previously in on - farm studies in Malawi (Mhango, 201 2 ). Overall, findings are consistent with the ability of pigeonpea to perform well on smallholder f ields in c entral Malawi and suggest the need for in - depth monitoring to fully document the growth of indeterminate plant life forms. This is notably true for sole pigeonpea stands which were highly productive. However, pigeonpea is widely grown in mixed in tercrop systems due to the complementarity of its slow initial growth pattern with the rapid early growth of short - season annual crops such as maize and grain legumes. Crop species effect on pigeonpea and inter vs intra competition is an important aspect t o understand in pigeonpea production. Although pigeonpea/maize is the most widely used pigeonpea production approach in Malawi, we noted very low pod biomass in this system relative to all others and relative to stem or leaf biomass (Figure 6). Competition from maize appears to differentially impact pod production which we 37 expect to have negative implications for pigeon pea grain production (this was not measured here due to livestock pressure). Overall, p igeonpea biomass (particularly root biomass) in the pigeonpea/groundnut system was quite similar t o sole cropped pigeonpea. In addition to the forage and p igeon pea grain associated with sole p igeon pea, a pigeonpea/groundnut intercrop produces additional groundnut grain that can be sold or consumed ( Snapp et al ., 2010). Our findings are thus consistent with the pigeonpea/groundnut intercrop as option for small holder farmers whose goal is to improve soil fertility (Gilbert, 2004) . From a study that compared pigeonpea with three other shrubby legu mes ([ Sesbania sesban L) ], [ Crotalaria grahamiana ], and [ Tephrosia vogelii ], Gathumbi et al . (2002) also concluded that g roundnut yields were substantially better when intercropped with species with an open canopy structure, such as pigeonpea . In addition to mean response we examined variability of shoot growth. The pigeonpea/soyabean intercrop was the most variable cropping system ( Table 8 ). In terms of total shoot biomass, sole pigeonpea had a 30% coefficient of variation while the pigeonpea/soyabe an int ercrop had 52%. On - farm experiments have been shown to exhibit high variability, usually in the range of 15 to 40% (Snapp et al ., 2010), similar to our field results. We observed very high pigeonpea growth variability in the pigeonpea/soyabean system, whic h is a novel finding and requires further research to determine if this is related to variability in soybean growth under different environmental conditions or some other factor. These results confirm previous research on pigeonpea as a crop that can th rive in unfavorable environmental conditions (Singh and Jauhar, 2005) , such as poor fertile soils typica l of smallholder farms (soil NH 4 + - N data, Figures 1 2 and 1 3 ) . At the same time, aboveground 38 biomass of pigeonpea on h eavy textured soils from baby trials produced more shoot biomass than on light textured soils. Higher aboveground biomass from heavy textured soils may have been due to the high soil nutrient status associated with heavier soils, which has previously been shown to enhance growth of legumes such as pig eonpea on smallholder farms in n orthern Malawi (Mhango et al ., 2012). No difference in shoot biomass was observed relative to soil texture from pigeonpea cropping systems evaluated in mother trials. These trial s were also conducted on - farm, however in researcher - designed systems that included fertilizer application which may have mitigated any d ifference due to soil texture on plant growth. 2.4.3 Root biomass Root biomass findings are consistent with studies in India by Kumar Rao and Dart (1987). Overall mean total root biomass was highest in sole pigeonpea and lowest in the pigeonpea/maize intercrop. Root biomass from both mother and baby trials was largely confined to the 0 20 cm soil layer. The pigeonpea/ maize and pigeonpea/soyabean intercrops performed the same way as short duration cultivars from the Kumar Rao and Dart (1987) study. The same authors had the same range of root biomass for long duration cultivars as we had in sole pigeonpea and the pigeonp ea/groundnut intercrop. In all the cropping systems from mother trials, overall root growth followed s hoot growth in our study (Figures 6 and 9 ). Previous research has also found high correlations between aboveground biomass and belowground biomass, with r oot to shoot ratios that vary within a narrow range ( Ravindranath and Ostwald , 2007 ) . S oil texture had a significant influence on total root biomass. Heavy textured soils were associated with higher amounts of root biomass than light textured soils in the topsoil of all trials. However, at lower depths in light textured soils there was evidence from the mother trials of the 39 opposite response: higher amounts of roots were observed compared to heavy textured soils . Th e topsoil response overall was consistent with shoot and root growth being in synchrony. However, there was modest evidence from the mother trial sites but not the baby trials that a deep rooting pattern was preferentially observed for low fertility, light soils. This is consistent with the f oraging theory which suggests that allocation of a large proportion of the root biomass and root surface area occurs relative to shoots, specifically in low - nutrient soils. In the literature support for this hypothesis i s mixed, with either a positive or n o correlation found between foraging precision and plant growth rate ( de Kroon and Mommer, 2005) . Overall, our total shoot and root biomass findings provide evidence that pigeonpea may be a highly effective agroforestry crop compared to many other widely promoted agroforestry shrub species. For instance, we found shoot biomass from pigeonpea to be in the range of 3 to 12 Mg ha - 1 , which is substantially more than the biomass reported by Schroth and Zech (1995) who conducted an above and belowground biomas s study on gliricidia ( Gliricidia sepium ) in West Africa. A research station study conducted in Zimbabwe found that pigeonpea consistently produced more aboveground biomass than either Acacia angustissima or Sesbania sesban agroforestry species in an improved fallow evaluation (Mafangoya and Dzowela , 1999) . This is also consistent with Gathumbi et al . (2002) who concluded pigeonpea performed better in intercrop systems than three other agroforestry species. Our findings of h igh biomass on on - farm trials are also consistent with another study in Malawi (Chirwa et al ., 2003). Chirwa and others (2003) found pigeonpea b iomass production to be higher on - farm than on - station and also higher than gliricidia biomass. 40 2.4.4 Root sh oot ratio According to Cairns et al . (1997) root shoot ratios are the relative biomass allocation between roots and shoots and the ratios are derived from dividing the total root biomass by the shoot biomass (Mokany et al ., 2006). Root shoot ratios are imp ortant as they are the evidence of plant partitioning of photosynthates between roots and shoots. There is a relationship between root shoot ratio and the factors that affect partitioning of aboveground and belowground biomass. Some of the factors include soil texture and plant species (Kuyah et al ., 2012) . We hypothesized higher root shoot ratios from the pigeonpea/maize intercrop (high competition and plant stress) in mother trials and from light textured soils in baby trials. However, contrary to our hy pothesis, there were no significant effects of cropping systems in mother trials or soil texture in baby trials on root shoot ratios of pigeonpea. Our findings did not show a trend of decreasing root shoot ratios with increasing shoot biomass in mother tri als. The root shoot ratio of sole pigeonpea was equal to that of the pigeonpea/maize intercrop in mother trials (Table 12 ). Root shoot ratios from baby trials in light and heavy textured soils were also equal. These results conflict with previous meta - anal ysis reviews by Mokany et al . ( 2006 ), who concluded that f ertile and productive soils enhance aboveground biomass while sacrificing belowground biomass and that l imit ing water and nutrients in light textured soils result in larger root shoot ratios than those in heavy textured soils. These results could also suggest that there might have been underestimations of root biomass such as f rom C losses ( root exudation and respiration or root sloughing ). Soil physical factors can modify the quantity, densi ty, branching patterns and depth of roots and this indirectly affects sampling methods (Vogt et al ., 1998). This may explain the lack of significant difference of root biomass from the different soil textures and cropping systems. However, our findings are consistent with Cairns et al . (1997). In a comparative review with more than 160 studies, the authors concluded 41 that root shoot ratios did not vary significantly with latitude (tropical, temperate or boreal), soil texture (fine, medium or coarse) or tree type (angiosperms or gymnosperms). 2.4.5.a Soil pH In agreement with previous authors ( Kamanga et al ., 2014; Mhango et al ., 2012; Adu - Gyamfi et al ., 2007; Snapp, 1998), soil pH from all the mother and baby trials was moderately acidic . The overall pH val ues for mother and baby trials were 5.20 and 5.37 respectively (Table s 3 and 4 ). According to Snapp (1998), moderate soil acidity is not a major edaphic problem for small holder farmers in Malawi. Poor soil fertility is one of the major challenges the farm ers in Malawi face instead. 2.4.5.b Soil texture Soil texture varies vastly throughout the landscape and changes in soil quality are heavily influenced by soil texture. Like other studies from Malawi (Mhango et al ., 2012; Snapp, 1998), our findings show that s oils on smallholder farms include a wide range of textures. Our silt, clay, and sand content from both mother and baby trials were consistent with findings from Kamanga et al . (2014). As with the former authors, clay content from both mother and ba by trials increased down the soil profile as sand content decreased. 2.4.5.c Total N and SO C In agreement with Myaka et al ., (2006), we also found more N and C in the upper soil layer and the two decreased with increasing soil depth (Table s 3 and 4 ). SOC ranges from both mother 42 and baby trials were consistent with those of Snapp (1998). SOC and N were both significantly affected by soil texture in baby trials. Both parameters were higher in heavy textured soils than in light textured soil as hypothesiz ed . Our soil N percentages were in the same range as in previous studies from Malawi (Kamanga et al ., 2014; Myaka et al ., 2006 , Sakala et al ., 2000 ). As expected, cropping system had no effect on SOC or N in this study . Previous on - farm studies in Malawi also did not observe significant differences in total N and C content fr om plots planted with maize and with pigeonpea /maize intercrops after only two cropping seasons (Myaka et al ., 2006; Adu - Gyamfi et al . , 2007). This suggests that it takes time for soil properties to chang e in highly eroded SSA soils. 2.4.5.d Soil NO 3 - - N , NH 4 + - N and PM N There were no significant effects of cropping system on soil NO 3 - - N , NH 4 + - N or PM N. However, NH 4 + - N and PM N were higher in the topsoil and in heavier soils (F igure 12 ) . Harawa et al . (2006) also observed similar trends for inorganic N in southern Malawi under agroforestry cropping systems. Our results from mother trials are in agreement with Myaka et al . (2006), who did not find any effect of cropping systems (sole ma ize and pigeonpea/maize intercrops) on soil inorganic N in Malawi on - farm studies. However, long - term studies have shown that agroforestry intercrop systems with pigeonpea produce higher soil inorganic N levels than sole maize (Beedy et al ., 2014) . 43 2.4.6 Conclusions and future directions The type of cropping system had a significant effect on the total shoot and root biomass. Pigeonpea productivity in terms of root and shoot biomass was highest in the sole pigeonpea cropping system, it was intermedia te in the doubled - up legume intercrops and lowest in the pigeonpea/maize intercrop. Soil texture also had a significant effect on total shoot and root biomass in baby trials. T he pigeonpea/soyabean intercrop was the most variable cropping system. There is need to explore more on the causes of variability of the pigeonpea/soyabean intercrop. Further research is needed to evaluate combined shoot and root biomass of all crops involved in the pigeonpea - based intercrops. O f the four cropping systems, there was n o significant effect on soil N and this suggests that small holder farmers can adopt or adapt any of the cropping systems Overall, p igeonpea productivity in the pigeonpea/groundnut system was comparable to s ole cropped pigeonpea , and a bonus crop of groundnut grain was produced as well . 44 APPENDIX 45 Table 1 . Monthly precipitation in mm during the 2013/2014 growing season for the three sites. A historical 15 - yr monthly average precipitation is presented. Source: TRMM (2015). Location Oct Nov Dec Jan Feb Mar Apr May Total 2013 2013 2013 2014 2014 2014 2014 2014 (mm) 15 - yr Average (2000 2014) Linthipe 4.97 55.15 146.43 289.08 307.87 76.79 96.49 1.72 978.50 Golomoti 8.66 45.62 181.19 224.61 250.09 77.20 57.40 3.66 848.43 Kandeu 13.04 38.45 188.91 261.79 290.85 68.75 40.55 6.39 908.72 933.26 902.59 911.08 46 Table 2 . for the three sites. A historical 15 - yr monthly average temperature is presented. Source: NASA (2010). Year Month Lo cation Mean Min Max C C C 2013 Oct Linthipe 30.05 (1.57) 24.83 31.79 Golomoti 35.32 (0.40) 34.13 36.22 Kandeu 33.68 (1.01) 31.76 35.10 2013 Nov Linthipe 29.82 (1.70) 24.70 31.76 Golomoti 36.85 (0.71 ) 35.11 38.40 Kandeu 35.51 (1.16) 32.76 37.70 2013 Dec Linthipe 26.06 (1.20) 22.35 27.42 Golomoti 31.42 (0.37) 30.52 32.57 Kandeu 30.74 (1.20) 28.29 32.86 2014 Jan Linthipe 20.81(0.92) 17.76 22.35 Golomoti 19.64 (0.65) 18.35 21.83 Kandeu 19.53 (0.67) 18.30 21.17 2014 Feb Linthipe 20.54 (0.59) 18.80 21.43 Golomoti 22.97 (0.69) 21.95 24.89 Kandeu 23.00 (0.86) 20.80 24.78 2014 Mar Linthipe 21.30 (0.77) 18.73 22.27 Golomoti 23.56 (0.34) 22.89 24.35 Kandeu 23.70 (0.42) 22.76 24.53 2014 Apr Linthipe 20.76 (0.77) 18.27 21.92 Golomoti 23.00 (0.54) 21.43 24.71 Kandeu 23.44 (0.49) 22.57 24.49 2014 May Linthipe 21.30 (0.81) 18.08 22.58 Golomoti 23.93 (0.51) 23.18 25.49 Kandeu 24.08 (0.50) 22.97 25.10 2000 - 2014 Oct - May Linthipe 22.97 (0.89) 20.31 23.77 Golomoti 26.34 (0.11) 26.11 26.70 Kandeu 26.00 (0.51) 24.76 26.57 47 Table 3 . Soil properties measured in mother trials. Soil texture is indicated by silt, sand and clay percent. Total soil N and SOC are indicated by N % and C %, pH is the measure of the acidity or basicity of soil. Overall averag es across mother trials are fol lowed by standard deviations in parentheses . Depth (cm) Silt % Clay % Sand % pH N % SOC % C/N Ratio Linthipe mother trial 0 20 13 33 55 5.3 0.14 2.49 18 20 40 14 47 39 5.3 0.13 2.30 18 40 60 13 49 38 5.4 0.10 1.65 17 Golomoti mother trial 1 (E) 0 20 9 22 69 5.6 0.06 0.84 14 20 40 11 22 67 5.3 0.01 0.13 13 40 60 10 28 62 5.2 0.01 0.09 9 Golomoti mother trial 2 (B) 0 20 12 31 58 5.4 0.08 1.03 13 20 40 12 41 47 5.4 0.01 0.1 10 40 60 12 42 46 5.4 0.01 0.09 9 Kandeu mother trial 0 20 6 10 83 5.0 0.03 0.33 11 20 40 8 23 69 4.7 0.02 0.24 12 40 60 8 36 55 4.9 0.02 0.16 8 Average 10 (3.96) 23.69 (10.01) 66.36 (13.91) 5.2 (0.38) 0.08 (0.05) 1.23 (0.94) 14 (2.49) Average 10 (3.96) 23.69 (10.01) 66.36 (13.91) 5.2 (0.38) 0.08 (0.05) 1.23 (0.94) 14 (2.49) Table 4 . Soil properties measured in baby trials. Soil texture is indicated by silt, sand and clay percent. Total soil N and SOC are indicated by N % and C %, pH is the measure of the acidity or basicity of soil. Overall averages across baby trials are followe d by standard deviations in parentheses . Depth (cm) Silt % Clay % Sand % pH N % SO C % C/N Ratio 0 20 8 20 72 5.4 0.06 0.99 17 20 40 9 23 68 5.3 0.05 0.82 16 40 60 10 27 63 5.4 0.03 0.59 20 Average 8 (3.11) 23 (10.81) 67.58 (0.05) 5.4 (0.66) 0.05 (0.03) 0.80 (0.50) 17 (2.26) 48 Table 5 . Cultivars of crops planted in the mother trials . District EPA Pigeonpea Groundnut Soyabean Maize Dedza Linthipe Mwaiwathu alimi Nsinjiro Nasoko PAN 53 Dedza Golomoti Mwaiwathu alimi JL24 Nasoko DKC 8033 Ntcheu Kande u Mwaiwathu alimi JL24 Nasoko DKC 8033 Table 6 . Plant population densities in four pigeonpea - based cropping systems in mother trials . Crop Cropping system Plant population density Total Pigeonpea Sole pigeonpea 44 000 44 000 Groundnut Pigeonpea/Gnut 120 000 164 000 Soyabean Pigeonpea/Soya 160 000 204 000 Maize Pige onpea/Maize 15 200 59 200 Table 7 . Planting dates across sites in mother trials . District EPA Planting date Dedza Linthipe 6 December 2013 Dedza Golom oti - B 19 December 2013 Dedza Golom oti - E 20 December 2013 Ntcheu Kandeu 17 December 2013 49 Table 8 . Averages and ranges of shoot and root biomass across mother trials. Means are followed by standard deviations in parentheses. Variable Mean Min Max Litter (Mg ha - 1 ) 0.70 (0.64) 0.01 2.16 Stems (Mg ha - 1 ) 3.27 (1.55) 0.45 7.40 Twigs (Mg ha - 1 ) 1.07 (0.91) 0.05 4.37 Leaves (Mg ha - 1 ) 0.73 (0.60) 0.02 2.76 Pods (Mg ha - 1 ) 1.33 (1.17) 0.02 3.98 Initial leaf fall (Mg ha - 1 ) 0.28 (0.26) 0.002 0.86 Total shoot biomass (Mg ha - 1 ) 6.73 (4.16) 0.48 19.44 Root biomass kg ha - 1 (0 20 cm) 882.17 (604.76) 5 6.00 2877.78 Root biomass kg ha - 1 (20 40 cm) 45.28 (57.89) 0.44 241 .93 Root biomass kg ha - 1 (40 60 cm) 11.41(12.57) 0.15 43.26 Total root biomass (Mg ha - 1 ) 0.94 (0.65) 0.06 3.11 Root shoot ratio 0.14 (0.05) 0.05 0.26 Table 9 . Averages and ranges of shoot and root biomass in baby trials. Means are followed by standard deviations in parentheses . Variable Mean Min Max Stems (Mg ha - 1 ) 3.86 (2.64) 0.53 12.32 Twigs (Mg ha - 1 ) 1.03 (0.79) 0.12 3.62 Leaves (Mg ha - 1 ) 0.97 (0.93) 0.002 4.87 Pods (Mg ha - 1 ) 0.57(0.61) 0.02 2.02 Initial leaf fall (Mg ha - 1 ) 0.34 (0.39) 0.001 1.55 Total shoot biomass (Mg ha - 1 ) 6.37 (5.04) 0.10 21. 55 Root biomass kg ha - 1 (0 20 cm) 927.52 (770.83) 31.54 3331.11 Root biomass kg ha - 1 (20 40 cm) 80.99 (76.09) 1.62 302.70 Root biomass kg ha - 1 (40 60 cm) 24.77 (30.57) 0.86 147.41 Total root biomass (Mg ha - 1 ) 1.03 (0.82) 0.03 3.78 Root shoot ratio 0.17 (0.04) 0.12 0.35 50 Table 10 . Soil NO 3 - - N (mg/kg soil) and potential mineralizable N (PMN) (mg/kg soil/day) m eans by cropping system and soil depth in mother trials. Means are followed by standard deviations in parentheses. Cropping system Variable Depth Mean Min Max Sole Pigeonpea Soil NO 3 - - N 0 20 cm 0.94 (0.49) 0.17 1.57 20 40 cm 0.42 (0.53) 0.00 1.31 40 60 cm 0.13 (0.18) 0.00 0.47 Pigeonpea/Gnut 0 20 cm 2.07 (2.74) 0.03 6.69 20 40 cm 0.76 (1.23) 0.00 3.71 40 60 cm 0.44 (0.69) 0.00 2 .13 Pigeonpea/Soya 0 20 cm 3.00 (2.18) 0.23 5.51 20 40 cm 0.86 (1.32) 0.00 4.45 40 60 cm 0.25 (0.38) 0.00 1.05 Pigeonpea/Maize 0 20 cm 1.32 (1.00) 0.17 2.79 20 40 cm 0.11 (0.15) 0.00 0.25 40 60 cm 0.002 (0.01) 0.00 0.02 Sole Pigeonpea PM N 0 20 cm 0.52 (0.34) 0.13 1.08 20 40 cm 0.34 (0.24) 0.05 1.60 40 60 cm 0.19 (0.14) 0.04 0.41 Pigeonpea/Gnut 0 20 cm 0.75 (0.44) 0.09 1.46 20 40 cm 0.54 (0.45) 0.0 5 1.60 40 60 cm 0.31 (0.33) 0.02 1.13 Pigeonpea/Soya 0 20 cm 0.88 (0.49) 0.35 1.96 20 40 cm 0.46 (0.42) 0.18 1.52 40 60 cm 0.09 (0.15) - 0.19 0.32 Pigeonpea/Maize 0 20 cm 0.73 (0.33) 0.29 1.21 20 40 cm 0.36 (0.16) 0.00 1.21 40 60 cm 0.24 (0.20) - 0.03 0.59 51 Table 11 . Soil NO3 - - N (mg/kg soil) and potential mineralizable N (PMN) (mg/kg soil/day) means by soil texture and soil depth in baby trials. Means are followed by standard deviations in parentheses. Soil texture Variable Depth Mean Min Max Light Soil NO 3 - - N 0 20 cm 1.22 (1.52) 0.00 5.96 20 40 cm 0.48 (1.12) 0.00 4.30 40 60 cm 0.21 (0.32) 0.00 1.29 Heavy 0 20 cm 2.20 (1.97) 0.27 8.55 20 40 cm 0.41 (0.86) 0.00 3.98 40 60 cm 0.18 (0.34) 0.00 1.29 Light PM N 0 20 cm 0.62 (0.34) 0.27 1.71 20 40 cm 0.32 (0.29) 0.01 1.09 40 60 cm 0.24 (0.41) - 0 .0 9 1.31 Heavy 0 20 cm 0.83 (0.46) 0.27 1.71 20 40 cm 0.52 (0.34) 0.08 1.51 40 60 cm 0.26 (0.21) 0.00 0.80 52 Table 12 . One - way ANOVA and mean response of pigeonpea litter, stems, twigs, leaves, pods, initial leaf fall, total shoot biomass, root biomass and root shoot ratio to cropping system. This data is from mother trials (n=39). Means are followed by standard errors . Variable Cropping System Mean Pr > F Litter (Mg ha - 1 ) Sole Pigeonpea 1.37 ± 0.19 0.0005* ** Pigeonpea/Gnut 0.64 ± 0.19 Pigeonpea/Soya 0.60 ± 0.17 Pigeonpea/Maize 0.22 ± 0.07 Stems (Mg ha - 1 ) Sole Pigeonpea 4.84 ± 0.57 0.0059* ** Pigeonpea/Gnut 3 .46 ± 0.40 Pigeonpea/Soya 2.56 ± 0.38 Pigeonpea/Maize 2.44 ± 0.18 Twigs (Mg ha - 1 ) Sole Pigeonpea 2.01 ± 0.35 0.0025* ** Pigeonpea/Gnut 1.31 ± 0.29 Pigeonpea/Soy a 0.62 ± 0.14 Pigeonpea/Maize 0.50 ± 0.06 Leaves (Mg ha - 1 ) Sole Pigeonpea 1.07 ± 0.23 0.0387* * Pigeonpea/Gnut 0.80 ± 0.24 Pigeonpea/Soya 0.69 ± 0.14 Pigeonpea/Maize 0.33 ± 0.11 Pods (Mg ha - 1 ) Sole Pigeonpea 2.66 ± 0.29 0.0036* ** Pigeonpea/Gnut 1.39 ± 0.69 Pigeonpea/Soya 0.88 ± 0.25 Pigeonpea/Maize 0.09 ± 0.03 Initial leaf fall (Mg ha - 1 ) Sole Pigeonpea 0.55 ± 0.08 0.0005* ** Pigeonpea/Gnut 0.26 ± 0.07 Pigeonpea/Soya 0.24 ± 0.07 Pigeonpea/Maize 0.09 ± 0.03 Total shoot biomass (Mg ha - 1 ) Sole Pigeonpea 11.83 ± 1.20 <0.001* ** Pigeonpea/Gnut 6.99 ± 1.24 Pigeonpea/Soya 5.08 ± 0.77 Pi geonpea/Maize 3.57 ± 0.33 Root biomass kg ha - 1 (0 20 cm) Sole Pigeonpea 1436.12 ± 225.44 0.0052* ** Pigeonp ea/Gnut 1087.88 ± 201.96 Pigeonpea/Soya 589.70 ± 104.53 Pigeonpea/Maize 512.48 ± 79.98 53 Table 12 . Variable Cropping System Mean Pr > F Root biomass kg ha - 1 (20 40 cm) Sole Pigeonpea 96.99 ± 22.24 0.0139* * Pigeonpea/Gnut 36.08 ± 10.13 Pigeonpea/Soya 36.12 ± 19.18 Pigeonpea/Maize 15.00 ± 4.94 Root biomass kg ha - 1 (40 60 cm) Sole Pigeonpea 25.43 ± 5.07 0.0269* Pigeonpea/Gnut 8.10 ± 2.03 Pigeonpea/Soya 7.72 ± 2.87 Pigeonpea/Maize 5.63 ± 2.45 Total root biomass (Mg ha - 1 ) Sole Pigeonpea 1.56 ± 0.24 0.0043* ** Pigeonpe a/Gnut 1.13 ± 0.21 Pigeonpea/Soya 0.63 ± 0.12 Pigeonpea/Maize 0.53 ± 0.08 Root shoot ratio Sole Pigeonpea 0.13 ± 0.02 0.1728 Pigeonpea/Gnut 0.17 ± 0.02 Pigeonpea/Soya 0.13 ± 0.01 Pigeonpea/Maize 0.15 ± 0.02 * Significant at 1 **Significant at ***Significant at 0.0 1 54 Table 13 . One - way ANOVA and mean response of pigeonpea stems, twigs, leaves, pods, initial leaf fall, total shoot biomass, root biomass and root shoot ratio to soil texture. This data is from baby trials (n=40). Means are followed by standard errors. Variable Cropping System Mean Pr > F Stems (Mg ha - 1 ) Light 3.22 ± 0.57 0.1119 Heavy 4.57 ± 0.60 Twigs (Mg ha - 1 ) Light 0.82 ± 0.17 0.0799 * Heavy 1.26 ± 0.18 Leaves (Mg ha - 1 ) Light 0.76 ± 0.20 0.1546 Heavy 1.19 ± 0.21 Po ds (Mg ha - 1 ) Light 0.56 ± 0.16 0.9655 Heavy 0.57 ± 0.16 Initial leaf fall (Mg ha - 1 ) Light 0.40 ± 0.09 0.3716 Heavy 0.28 ± 0.09 Total shoot biomass (Mg ha - 1 ) Light 4.75 ± 0.61 0.0420* * Heavy 8.17 ± 1.48 Root biomass kg ha - 1 (0 20 cm) Light 677.19 ± 76.32 0.0429* * Heavy 1 204.19 ± 232.90 Root biomass kg ha - 1 (20 40 cm) Light 79.71 ± 20.86 0.9398 Heavy 81.66 ± 15.31 Root biomass kg ha - 1 (40 60 cm) Light 26.64 ± 9.45 0.7696 Heavy 23.37 ± 5.82 Total root biomass (Mg ha - 1 ) Light 0.77 ± 0.09 0.0407* * Heavy 1.33 ± 0.24 Root shoot ratio Light 0.18 ± 0.02 0.9663 Heavy 0.18 ± 0.02 * Significant at 1 **Significant at 55 Table 14 . Correlation matrix of explanatory variables in mother trials. Pearson Correlation Coefficients, N = 39 Prob > |r| under H0: Rho=0 Total shoot biomass Total root biomass pH Clay % Sand % N % SO C % Total shoot biomass 1 0.88724 0.25039 - 0.0426 - 0.0195 0.25993 0.31107 <.0001 0.1242 0.7967 0.9061 0.11 0.0539 Total root biomass 0.88724 1 0.18352 0.21851 - 0.2648 0.49623 0.53236 <.0001 0.2634 0.1814 0.1033 0.0013 0.0005 pH 0.25039 0.18352 1 0.37946 - 0.3863 0.28903 0.22382 0.1242 0.2634 0.0172 0.0151 0.0743 0.1708 Clay % - 0.0426 0.21851 0.37946 1 - 0.9789 0.79979 0.74488 0.7967 0.1814 0.0172 <.0001 <.0001 <.0001 Sand % - 0.0195 - 0.2648 - 0.3863 - 0.9789 1 - 0.8161 - 0.7706 0.9061 0.1033 0.0151 <.0001 <.0001 <.0001 N % 0.25993 0.49623 0.28903 0.79979 - 0.8161 1 0.98834 0.11 0.0013 0.0743 <.0001 <.0001 <.0001 SO C % 0.31107 0.53236 0.22382 0.74488 - 0.7706 0.98834 1 0.0539 0.0005 0.1708 <.0001 <.0001 <.0001 56 Table 15 . Correlation matrix of explanatory variables in baby trials . Pearson Correlation Coefficients, N = 40 Prob > |r| under H0: Rho=0 Total shoot biomass Total root biomass pH Clay % Sand % N % SO C % Total shoot biomass 1 0.88724 0.25039 - 0.0426 - 0.0195 0.25993 0.31107 <.0001 0.1242 0.7967 0.9061 0.11 0.0539 Total root biomass 0.88724 1 0.18352 0.21851 - 0.2648 0.49623 0.53236 <.0001 0.2634 0.1814 0.1033 0.0013 0.0005 pH 0.25039 0.18352 1 0.37946 - 0.3863 0.28903 0.22382 0.1242 0.2634 0.0172 0.0151 0.0743 0.1708 Clay % - 0.0426 0.21851 0.37946 1 - 0.9789 0.79979 0.74488 0.7967 0.1814 0.0172 <.0001 <.0001 <.0001 Sand % - 0.0195 - 0.2648 - 0.3863 - 0.9789 1 - 0.8161 - 0.7706 0.9061 0.1033 0.0151 <.0001 <.0001 <.0001 N % 0.25993 0.49623 0.28903 0.79979 - 0.8161 1 0.98834 0.11 0.0013 0.0743 <.0001 <.0001 <.0001 SO C % 0.31107 0.53236 0.22382 0.74488 - 0.7706 0.98834 1 0.0539 0.0005 0.1708 <.0001 <.0001 <.0001 57 Figure 1 . Map of Malawi showing the central region and the specific locations of mother and baby trials from the Linthipe site. 58 Figure 2 . Map of Malawi showing the central region and specific locations of mother and baby trials for the Golomoti site. 59 Figure 3 . Map of Malawi showing the central region and specific locations of mother and baby trials for the Kandeu site. 60 Figure 4 . Map of Malawi showing temperature ranges across the country and the specific locations for which monthly temperature is presented for the three agro - ecologies. 61 Figure 5 . Frequencies of soil types in baby trials for the 0 20 cm depth . 62 Figure 6 . Shoot biomass in four cropping systems from mother trials. Total shoot biomass was separated into stems, twigs, leaves and pods initial leaf fall. Total litter that was collected over ten weeks from litter traps is also included in shoot biomass . 63 Figure 7 . Shoot biomass from light and heavy textured soils from the 0 20 cm depth in mother tr ials. Total shoot biomass was separated into stems, twigs, leaves and pods. Initial leaf fall is also included in shoot biomass. 64 Figure 8 . Shoot biomass averaged by sites with light and heavy textured soils, from baby trials. Total shoot biomass was separated into stems, twigs, leaves and pods. Initial leaf fall is also included in shoot biomass . 65 Figure 9 . Root biomass from the 0 20, 20 40 and 40 60 cm depths in four cropping systems from mother trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. 66 Figure 10 . Root biomass from the 0 20, 20 40 and 40 60 cm depths, grouped by light and heavy textured soils from mother trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. 67 Figure 11 . Roo t biomass from the 0 20, 20 40 and 40 60 cm depths in light and heavy textured soils from baby trials. Error bars represent ± standard errors of treatment means. Note: the scale across soil depths is different. 68 Figure 12 . Soil NH 4 + - N by cropping systems from the 0 60 cm depths in mother trials. Error bars represent ± standard errors of treatment means. 69 Figure 13 . Soil NH 4 + - N by soil texture in baby trials from the 0 60 cm depths. Error bars represent ± standard errors of treatment means . 70 BIBLIOGRAPHY 71 BIBLIOGRAPHY Adu - Gyamfi, J. J., Myaka, F. A., Sakala, W. D., Odgaard, R., Vesterager, J. M., & Høgh - Jensen, H. (2007). Biological N fixation and N and phosphorus budgets in farmer - managed intercrops of maize pigeonpea in semi - arid southern and eastern Africa. Plant and soil , 295 (1 - 2), 127 - 136. ArcMap. (2014) . ArcMap ESRI (Environmental Systems Resource Institute), Version 10.0 10.2 ., ESRI. Redlands, California. Bantilan, M. C. S., & Parthasarathy, D. (1999). Efficiency and sustainability gains from adoption of short - duration pigeonpea in nonlegume - based cropping systems (No. 5). Patancheru: ICRISAT Beedy, T. L., Nyamadzawo, G., L uedeling, E., Kim, D. G., Place, F., & Hadgu, K. M. (2014). 11 Agroforestry for Small Landholders of Eastern and Southern Africa. Soil Management of Smallholder Agriculture, 237. Broughton, W. J., Hernandez, G., Blair, M., Beebe, S., Gepts, P., & Vanderleyden, J. (2003). Beans (Phaseolus spp.) model food legumes. Plant and soil , 252 (1), 55 - 128. Chikowo, R., Mapfumo, P., Nyamugafata, P., & Giller, K. E. (2004). Maize productivity and mineral N dynamics following different soil fertility management practices on a depleted sandy soil in Zimbabwe. Agriculture, ecosystems & environment , 102 (2), 119 - 131. Burke, M. B., Lobell, D. B., & Guarino, L. (2009). Shifts in African crop climates by 2050, and the implications for crop improvement and genetic resou rces conservation. Global Environmental Change, 19(3), 317 - 325. Crews, T. E., & Peoples, M. B. (2004). Legume versus fertilizer sources of N : ecological tradeoffs and human needs. Agriculture, Ecosystems & Environment , 102 (3), 279 - 297. Cairns, M. A., Brown, S., Helmer, E. H., & Baumgardner, G. A. (1997). Root biomass allocation in the world's upland forests. Oecologia , 111 (1), 1 - 11. Chirwa, P. W., Black, C. R., Ong, C. K., & Maghembe, J. A. (2003). Tree and crop productivity in gliricidi a/maize/pigeonpea cropping systems in southern Malawi. Agroforestry Systems, 59(3), 265 - 277. Degrande, A. (2001). Farmer assessment and economic evaluation of shrub fallows in the humid lowlands of Cameroon. Agroforestry systems , 53 (1), 11 - 19. de Kroon, H ., & Mommer, L. (2006). Root foraging theory put to the test. Trends in Ecology & Evolution, 21(3), 113 - 116. 72 Department of Sustainable Natural Resources (2013): http://www.environment.nsw.gov.au/resources/soils/testmethods/phw.pdf Doane, T. A., & Horwáth, W. R. (2003). Spectrophotometric determination of nitrate with a single reagent. Analytical Letters, 36(12), 2713 - 2722. Dorward, A. (1999). Farm size and produ ctivity in Malawian smallholder agriculture. The Journal of Development Studies , 35 (5), 141 - 161. Earth Observing System Data and Information System (EOSDIS). (2009). Earth Observing System ClearingHouse (ECHO) / Reverb Version 10.X [online application]. Greenbelt, MD: EOSDIS, Goddard Space Flight Center (GSFC) National Aeronautics and Space Administration (NASA). URL: https://wist.echo.nasa.gov/api/ Funk, C., Dettinger, M. D., Michaelsen, J. C., Verdin, J. P., Brown, M. E., Barlow, M., & Hoell, A. (2008). Warming of the Indian Ocean threatens eastern and southern African food security but could be mitigated by agricultural development. Proceedings of the Na tional Academy of Sciences, 105(32), 11081 - 11086. Gathumbi, S. M., Ndufa, J. K., Giller, K. E., & Cadisch, G. (2002). Do species mixtures increase above - and belowground resource capture in woody and herbaceous tropical legumes?. Agronomy Journal, 94(3), 5 18 - 526. Gilbert, R. A. (2004). Best - bet legumes for smallholder maize - based cropping systems of Malawi. In Green Manure/Cover Crop Systems of Smallholder Farmers (pp. 153 - 174). Springer Netherlands. Giller, K. E., & Cadisch, G. (1995). Future benefits from biological N fixation: an ecological approach to agriculture. In Management of Biological N Fixation for the Development of More Productive and Sustainable Agricultural Systems (pp. 255 - 277). Springer Netherlands. Giller, K. E., Cadisch, G., Ehaliotis, C., Adams, E., Sakala, W. D., & Mafongoya, P. L. (1997). Building soil nitrogen capital in Africa. Replenishing soil fertility in Africa, (replenishingsoi), 151 - 192. ICRISAT (2013): http://www.icrisat.org/tropicallegumesII/pdfs/November - 2013.pdf Kamanga, B. C. G., Waddington, S. R., Whitbread, A. M., Almekinders, C. J. M., & Giller, K. E. (2014). Improving the efficiency of use of small amounts of N and phosphorus fertiliser on smallholder maize in central Malawi. Experimental Agriculture, 50(02), 229 - 249. Kane, D. A., Snapp, S. S., & Davis, A. S. (2015). Ridge Tillage Conce ntrates Potentially Mineralizable Soil N , Facilitating Maize N Uptake. Soil Science Society of America Journal, 79(1), 81 - 88. 73 Khan, D. F., Peoples, M. B., Chalk, P. M., & Herridge, D. F. (2002). Quantifying below - ground N of legumes. 2. A comparison of 15 N and non - isotopic methods. Plant and Soil , 239 (2), 277 - 289. Kumar Rao, J. K., & Dart, P. J. (1987). Nodulation, N fixation and N uptake in pigeonpea (Cajanus cajan (L.) Millsp) of different maturity groups. Plant and Soil , 99 (2 - 3), 255 - 266. Kuyah , S., Dietz, J., Muthuri, C., Jamnadass, R., Mwangi, P., Coe, R., & Neufeldt, H. (2012). Allometric equations for estimating biomass in agricultural landscapes: II. Belowground biomass. Agriculture, Ecosystems & Environment , 158 , 225 - 234. MacColl, D. (1989 ). Studies on maize (Zea mays) at Bunda, Malawi. II. Yield in short rotations with legumes. Experimental Agriculture , 25 (03), 367 - 374. Mafongoya, P. L., & Dzowela, B. H. (1999). Biomass production of tree fallows and their residual effect on maize in Zimb abwe. Agroforestry systems, 47(1 - 3), 139 - 151. Mhango, W. G., Snapp, S. S., & Phiri, G. Y. (201 2 ). Opportunities and constraints to legume diversification for sustainable maize production on smallholder farms in Malawi. Renewable Agriculture and Food Syste ms , 28 (03), 234 - 244. Mokany, K., Raison, R., & Prokushkin, A. S. (2006). Critical analysis of root: shoot ratios in terrestrial biomes. Global Change Biology , 12 (1), 84 - 96. Myaka, F. M., Sakala, W. D., Adu - Gyamfi, J. J., Kamalongo, D., Ngwira, A., Odgaard , R., ... & Høgh - Jensen, H. (2006). Yields and accumulations of N and P in farmer - managed intercrops of maize pigeonpea in semi - arid Africa. Plant and soil , 285 (1 - 2), 207 - 220. Nam, N. H., Chauhan, Y. S., & Johansen, C. (1993). Comparison of extra - short - du ration pigeonpea with short - season legumes under rainfed conditions on Alfisols. Experimental Agriculture , 29 (03), 307 - 316. Nene, Y. L., Hall, S. D., & Sheila, V. K. (1990). The pigeonpea . CAB International. NASA Land Processes Distributed Active Archiv e Center (LP DAAC), MODIS Land Surface Temperature & Emissivity (MOD11A2) . (2010). USGS/Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota Ngwira, A. R., Aune, J. B., & Mkwinda, S. (2012). On - farm evaluation of yield and eco nomic benefit of short term maize legume intercropping systems under conservation agriculture in Malawi. Field crops research , 132 , 149 - 157. Odion, E. C., Asiribo, O. E., Ogunlela, V. B., Singh, B. B., & Tarawali, S. A. (2007). Strategies to improve and s ustain food production capacity in the savanna: The role of leguminous fodder crops in maintaining soil fertility and health. Journal of Food Agriculture and Environment , 5 (2), 338. 74 Onim, J. F. M., Mathuva, M., Otieno, K., & Fitzhugh, H. A. (1990). Soil fertility changes and response of maize and beans to green manures of leucaena, sesbania and pigeonpea. Agroforestry Systems , 12 (2), 197 - 215. Orr, A., Mwale, B., & Saiti, D. (2000). Markets and livelihoods: Lessons from farmi ng systems research in Blantyre/Shire Highlands. p 279 291. In Integrated crop management research in Malawi: Developing technologies with farmers. Proc. of the final workshop of the farming systems integrated pest management project. Natural Resources Ins t., Chatham, UK . Orr, A., & Ritchie, J. M. (2004). Learning from failure: smallholder farming systems and IPM in Malawi. Agricultural systems , 79 (1), 31 - 54. Pypers, P., Huybrighs, M., Diels, J., Abaidoo, R., Smolders, E., & Merckx, R. (2007). Does the enhanced P acquisition by maize following legumes in a rotation result from improved soil P availability? Soil Biology and Biochemistry , 39 (10), 2555 - 2566. Ravindranath, N. H., & Ostwald, M. (2007). Methods for Below - Ground Biomass. C Inventory Methods Ha ndbook for Greenhouse Gas Inventory, C Mitigation and Roundwood Production Projects, 149 - 156. Rochester, I. J., Peoples, M. B., Hulugalle, N. R., Gault, R., & Constable, G. A. (2001). Using legumes to enhance N fertility and improve soil condition in cotton cropping systems. Field Crops Research , 70 (1), 27 - 41. Reddy, B. V. S., Reddy, P. S., Bidinger, F., & Blümmel, M. (2003). Crop management factors influencing yield and quality of crop residues. Field Crops Re search, 84(1), 57 - 77. SAS Institute. 2002. SAS release 9.3. The SAS Inst., Cary, NC. Sakala, W. D., Cadisch, G., & Giller, K. E. (2000). Interactions between residues of maize and pigeonpea and mineral N fertilizers during decomposition and N mineralizat ion. Soil biology and biochemistry, 32(5), 679 - 688. Schroth, G., & Zech, W. (1995). Above - and below - ground biomass dynamics in a sole cropping and an alley cropping system with Gliricidia sepium in the semi - deciduous rainforest zone of West Africa. Agrofo restry Systems, 31(2), 181 - 198. SigmaPlot. (20 15 ). SigmaPlot for Windows. Ver. 13.0. Sinclair, T. R., & Vadez, V. (2012). The future of grain legumes in cropping systems. Crop and Pasture Science , 63 (6), 501 - 512. Snapp, S. (1999). Mother and baby trials: a novel trial design being tried out in Malawi. TARGET. The Newsletter of the Soil Fertility Research Network for Maize - Based Cropping Systems in Malawi and Zimbabwe . 75 Snapp, S. S. (1998). Soil nutrient status of smallholder farms in Malawi 1. Commun ications in Soil Science & Plant Analysis , 29 (17 - 18), 2571 - 2588. Snapp, S. S., Blackie, M. J., Gilbert, R. A., Bezner - Kerr, R., & Kanyama - Phiri, G. Y. (2010). Biodiversity can support a greener revolution in Africa. Proceedings of the National Academy of Sciences, 107(48), 20840 - 20845. Snapp, S. S., Jones, R. B., Minja, E. M., Rusike, J., & Silim, S. N. (2003). Pigeonpea for africa: a versatile vegetable and more. HortScience , 38 (6), 1073 - 1079. Snapp, S. S., Mafongoya, P. L., & Waddington, S. (1998). Organic matter technologies for integrated nutrient management in smallholder cropping systems of southern Africa. Agriculture, ecosystems & environment , 71 (1), 185 - 200. Snapp, S. S., Rohrbach, D. D., Simtowe, F., & Freeman, H. A. (2002). Sustainable soil management options for Malawi: can smallholder farmers grow more legumes? Agriculture, ecosystems & environment , 91 (1), 159 - 174. Snapp, S., Kanyama - Phiri, G., Kamanga, B., Gilbert, R., & Wellard, K. (2002). Farmer and researcher partnerships in Malawi: d eveloping soil fertility technologies for the near - term and far - term. Experimental agriculture , 38 (04), 411 - 431. Snapp, S. S., & Silim, S. N. (2002). Farmer preferences and legume intensification for low nutrient environments. In Food Security in Nutrient - Genetic Capabilities (pp. 289 - 300). Springer Netherlands. Team, R. C. (2015). R: A language and environment for statistical computing. Vienna, Austria; 2014. URL http://www. R - project. org. Tropical Rainfall Measurement Mission Project (TRMM). (2015) Goddard Space Flight Center Distributed Active Archive Center (GSFC DAAC). Vance, C. P. (2001). Symbiotic N fixation and phosphorus acquisition. Plant nutrition in a world of declining renewable resources. Plant physiology , 127 (2), 390 - 397. Vitousek, P. M., Naylor, R., Crews, T., David, M. B., Drinkwater, L. E., Holland, E., ... & Zhang, F. S. (2009). Nutrient imbalances in agricultural development. Science, 324 (5934), 1519. Vogt, K. A., Vogt, D. J., & Bloomfield , J. (1998). Analysis of some direct and indirect methods for estimating root biomass and production of forests at an ecosystem level. In Root Demographics and Their Efficiencies in Sustainable Agriculture, Grasslands and Forest Ecosystems (pp. 687 - 720). S pringer Netherlands. Zingore, S., Murwira, H. K., Delve, R. J., & Giller, K. E. (2007). Influence of nutrient management strategies on variability of soil fertility, crop yields and nutrient balances on smallholder farms in Zimbabwe. Agriculture, ecosystem s & environment , 119 (1), 112 - 126.