An experiment is described in which the productivity of maize–bean and maize–potato mixtures in alternate rows was compared with pure stands of the component species in four cropping seasons. In seasons with low rainfall, yield from the mixtures fell short of that from pure stands, but in one exceptionally wet season a yield advantage was found for maize–bean mixtures. The poor performance of mixtures in low rainfall seasons appeared to result from large reductions in maize yield attributable to competition from the potato or bean. In the one high-rainfall season, no such reduction in maize yield occurred. An implication of these findings for farming systems with mixed cropping is discussed.