lntercropping trials incorporating upland rice (Oryza sativa L.) and common bean (Phaseolus vulgaris L.), and maize (Zea mays L.) and common bean were carried out during two seasons under rainfed conditions in southern Sudan. The performance of improved rice varieties (ITA 118 and ITA 122), intercropped with various densities of bean, was measured against that of local varieties selected from on-farm variety trials. During the first season the two improved rice varieties greatly out-yielded a local variety which was severely affected by rice blast (Pyricularia oryzae Cavara). Neither rice yield nor bean yield was affected by bean density, but bean yield was significantly greater when intercropped with ITA 118 than with ITA 122, or a farmer's variety. Farmers' varieties under their own management were generally more productive than the improved varieties. In on-station trials during the second season, three farmers' varieties were compared with the two improved varieties intercropped at two densities of bean. Farmers' varieties out-yielded the improved ones at both bean densities. Bean yields were greater when grown with the improved rice varieties, apparently due to the shorter stature of the latter, and greater at the higher density. Maize-bean intercropping was studied simultaneously. Line-sown maize, intercropped with bean, yielded better than when sown in hills at the same density, but was more susceptible to lodging. Bean yields were unaffected by sowing arrangement, but yielded more when sown at higher densities. Sowing maize at two densities did not affect yield of the bean intercrop, but the maize yielded more at the higher density. Bean generally grew and produced better when intercropped with rice than with maize.