For many years biologists have debated whether mast seeding (the synchronous intermittent production of large seed crops in perennial plants) results from weather conditions or is an evolved plant reproductive strategy. In this review, we analyze the evidence for the underlying causes of masting. In the absence of selection for higher or lower variability, plants will vary in tandem with the environment (resource matching). Two selective factors often favor the evolution of masting: increased pollination efficiency in wind-pollinated species, and satiation of seed predators. Other factors select against masting, including animal pollination and frugivore dispersal. A survey of 570 masting datasets shows that wind-pollinated species had higher seed production coefficients of variation (CVs) than biotically pollinated ones. Frugivore-dispersed species had low CVs whereas predator-dispersed plants had high CVs, consistent with gaining benefits from predator satiation rather than dispersal. The global pattern of masting shows highest seed crop variability at mid latitudes and in the Southern Hemisphere, which are similar to the patterns in variability of rainfall. We conclude that masting is often an adaptive reproductive trait overlaid on the direct influence of weather.