Nine trials were conducted, in like number of localities in Central America in 1990 and 1991, to evaluate the effect of legume intercropping on corn yield. The first cycle was evaluated in 1989 and 1990, when trhree legumes were planted with doses of 20 and 40 kg of P /ha each, The legumes stubble was left on their respective plots at the end of the first cycle and in following year, corn was planted on the same plots as the preeeding eycle, keeping them well identified. No fertilizer was applied to any treatment in this residual cycle. The results show that the legumes increased the corn yield in the residual cycle, although only Canavalia ensiformis showed to have a net positive effeet after the two cycles. This net effect is of 203 ± 314 kg/ha as an average of the nine localities.