Niche-based assessment of contributions of legumes to the nitrogen economy of Western Kenya smallholder farms

Nitrogen (N) deficiency is a major constraint to the productivity of the African smallholder farming systems. Grain, green manure and forage legumes have the potential to improve the soil N fertility of smallholder farming systems through biological N-2-fixation. The N-2-fixation of bean (Phaseolus vulgaris), soyabean (Glycine max), groundnut (Arachis hypogaea), Lima bean (Phaseolus lunatus), lablab (Lablab purpureus), velvet bean (Mucuna pruriens), crotalaria (Crotalaria ochroleuca), jackbean (Canavalia ensiformis), desmodium (Desmodium uncinatum), stylo (Stylosanthes guianensis) and siratro (Macroptilium atropurpureum) was assessed using the N-15 natural abundance method. The experiments were conducted at three sites in western Kenya, selected on an agro-ecological zone (AEZ) gradient defined by rainfall. On a relative scale, Museno represents high potential AEZ 1, Majengo medium potential AEZ 2 and Ndori low potential AEZ 3. Rainfall in the year of experimentation was highest in AEZ 2, followed by AEZ 1 and AEZ 3. Experimental fields were classified into high, medium and low fertility classes, to assess the influence of soil fertility on N-2-fixation performance. The legumes were planted with triple super phosphate (TSP) at 30 kg P ha(-1) stop, with an extra soyabean plot planted without TSP (soyabean-P), to assess response to P, and no artificial inoculation was done. Legume grain yield, shoot N accumulation, %N derived from N-2-fixation, N-2-fixation and net N inputs differed significantly (P < 0.01) with rainfall and soil fertility. Mean grain yield ranged from 0.86 Mg ha(-1), in AEZ 2, to 0.30 Mg ha(-1), in AEZ 3, and from 0.78 Mg ha(-1), in the high fertility field, to 0.48 Mg ha(-1) stop, in the low fertility field. Shoot N accumulation ranged from a maximum of 486 kg N ha(-1) in AEZ 2, to a minimum of 10 kg N ha(-1) in AEZ 3. Based on shoot biomass estimates, the species fixed 25-90% of their N requirements in AEZ 2, 23-90% in AEZ 1, and 7-77% in AEZ 3. Mean N-2-fixation by green manure legumes ranged from 319 kg ha(-1)(velvet bean) in AEZ 2 to 29 kg ha(-1) (jackbean) in AEZ 3. For the forage legumes, mean N-2-fixation ranged from 97 kg N ha(-1) for desmodium in AEZ 2 to 39 kg N ha(-1) for siratro in AEZ 3, while for the grain legumes, the range was from 172 kg N ha(-1) for lablab in AEZ 1 to 3 kg N ha(-1) for soyabean-P in AEZ 3. Lablab and groundnut showed consistently greater N-2-fixation and net N inputs across agro-ecological and soil fertility gradients. The use of maize as reference crop resulted in lower N-2-fixation values than when broad-leaved weed plants were used. The results demonstrate differential contributions of the green manure, forage and grain legume species to soil fertility improvement in different biophysical niches in smallholder farming systems and suggest that appropriate selection is needed to match species with the niches and farmers' needs.