A trial focusing on the management of Mucuna pruriens (mucuna) and maize stover residues was conducted over two seasons on a low fertility soil in the North Rift region of Kenya. The effects of factorial combinations of legume and/or stover residues and their method of placement on a subsequent maize crop were evaluated. When the maize was harvested, the stovers were either removed or applied as mulch between the rows of mucuna. Mucuna aboveground biomass was removed to simulate grazing, or applied as surface mulch or incorporated into the soil, 7 days before planting a subsequent maize crop. Mucuna accumulated 5.35 and 1.03 t of above ground dry matter in the first and second year, respectively. The bulk of the DM consisted of low quality woody stem and litter. The aboveground legume residues contained an estimated 148 and 32.5 kg N ha-l in the first and second year, respectively. The removal of the aboveground mucuna biomass significantly affected maize yields in the first year. Maize in the simulated grazing plots produced an average of 1.6 t ha-l less grain than the treatments in which the biomass was retained(4.1 compared to 5.7 t ha-\\ Legume biomas retained treatments produced 30% more grain than the no legume unfertilised check. In the second year, however, when mucuna performed poorly, the previous maize crop's stovers depressed ~rain yields by 16% compared to when the stover was removed at harvest. Incorporated residues produced 13% more grain than residues applied as surface mulch. The between year differences in mucuna DM and biomass total N content seemed to be the main factor affecting the subsequent maize yields. Possible ways in which current cropping system characteristics and the legume's dry season niche are likely to affect the green manure system are highlighted.