We report a study in northern Thailand to examine the effects of fertilizer N, applied both to paddy rice and to a subsequent soybean crop on symbiotic and yield characteristics of soybean and on the differences between inputs of fixed N 2 and the removal of N as harvested product. Treatments were a factorial arrangement of 0, 100 and 300 kg N ha -1 applied to the rice (designated R0, R100 and R300, respectively), and 0, 25 and 50 kg N ha -1, applied as 'starter' fertilizer to the soybean ($0, $25 and $50, respectively).Nitrogen applied to the rice increased rice yields by up to 74% but proportions recovered by the rice were low (45% [R100] and 14% [R300]). The rice N treatments had only marginal effects on soybean nodulation (up to 17% reduction in early growth) and above-ground dry matter (up to 9% increase). Effects on soybean seed yield and total N 2 fixed were insignificant. Starter N, applied to the soybean at sowing, also marginally reduced nodulation and enhanced above-ground dry matter. Total N 2 fixed was unaffected but seed yield was increased by up to 6%. For all treatments, total above-ground N ranged from 145 to 179kgha -~ with 72 to 85% (122 and 140kgha -~) derived from N 2 fixation. When harvested product consisted of seed only, differences between inputs of fixed N 2 and removals of seed N were close to zero (-10 to +9 kg N ha -l) with little effect of fertilizer N. The N balances were reduced by an average of 18 kg N ha -1 when straw was included as harvested product. We concluded that N applied to the rice and to the following soybean was inefficiently used by those crops and had only marginal effects of symbiotic activity of the soybean. Furthermore, the benefit of the N 2 fixing soybean in this system was to slow the decline of, rather than enhance, the N fertility of the soil