The yield increases often recorded in maize following grain legumes have been attributed to fixed-N and ‘other rotation’ effects, but these effects have rarely been separated. Field trials were conducted between 2003 and 2005 to measure these effects on maize following grain legumes in the northern Guinea savanna of Nigeria. Maize was grown on plots previously cultivated to two genotypes each of soybean (TGx 1448-2E and SAMSOY-2) and cowpea (IT 96D-724 and SAMPEA-7), maize, and natural fallow. The plots were split into four N fertilizer rates (0, 30, 60 and 90 kg N ha−1) in a split plot design. The total effect was calculated as the yield of maize following a legume minus the yield following maize, both without added N and the rotation effect was calculated as the difference between rotations at the highest N fertilizer rate. The legume genotypes fixed between 14 and 51 kg N ha−1 of their total N and had an estimated net N balance ranging from −29.8 to 9.5 kg N ha−1. Positive N balance was obtained only when the nitrogen harvest index was greater than the proportion of N derived from atmosphere. The results also indicated that the magnitude of the fixed-N and other rotation effects varied widely and were influenced by the contributions of the grain legumes to the soil N-balance. In general, fixed-N effects ranged from 124 to 279 kg ha−1 while rotation effects ranged between 193 and 513 kg ha−1. On average, maize following legumes had higher grain yield of 1.2 and 1.3-fold compared with maize after fallow or maize after maize, respectively.