Evaluating the use of two contrasting legume species as relay intercrop in upland rice cropping systems

Abstract

In the savanna zone of West Africa with a prolonged dry season, the successful growth of legumes as short-season fallow necessitates relay establishment of cover crops into the preceding food crop. Inappropriate choice of legume species and establishment dates may result in severe interspecific competition for light, water and nutrients and a subsequent yield loss in the main crop. In the current study, the highly competitive Cajanus cajan L. and the weakly competitive Stylosanthes hamata Taub. were relay-seeded into two upland rice cultivars with different competitive ability (WAB56-50, WAB450-24-3-2-P18-HB syn. V4 ) at different sowing dates, ranging from 0 to 84 days after rice sowing (DARS). Using a logistic function was adequate to describe rice grain yield, final rice biomass and legume biomass at 1 week after rice harvest as a function of legume sowing date. Rice biomass and grain yield were significantly ( P<0.001) reduced when legumes were established between 0 and 28 DARS. Beyond 56 DARS no competition effects on rice were observed. At these late sowing dates only relatively low amounts of legume biomass were obtained, due to a combined effect of a more severe competition by the rice crop and a shorter growing period. The derived yield loss–legume biomass relationship showed that C. cajan was the most appropriate legume species of the two under study, to suit a relay intercropping system with upland rice, as it produced reasonable amounts of biomass at low levels of rice grain yield loss. This conclusion was confirmed by results on the after-effect of the legumes, where only the improved fallow by C. cajan gave a significant increase in the yield of the subsequent rice crop. Optimum sowing time of C. cajan was between 30 and 35 DARS. The effect of selecting a specific legume was found to surpass that of choosing either one of the two rice cultivars, for which contrasting results were obtained in 2 years of experimentation.