The theoretical ideas underpinning the evolution of seed dormancy and seed mass are briefly reviewed. Comparative tests of these theoretical ideas are then presented; all tests use modern comparative methods to allow accurate, meaningful comparisons across species. The comparative analyses of seed dormancy demonstrate that species that average the environment across time, by having long-lived adults or space as a result of clonal growth, typically have less dormant seeds. Amongst species with seeds unspecialized for spatial dispersal, seed mass is not related to dormancy. However, in species with wind- or animal-dispersed seeds those with heavy seeds typically have less dormancy. This is consistent with heavy seeds having better establishment success and/or suffering higher levels of herbivory. The relationships between seed mass and plant height, lateral spread and adult longevity are explored using the Sheffield dataset. The complex pattern of relationships between these variables is interpreted in terms of dispersal limitation, escape from competition with clonal ramets and differences in establishment conditions. Some of the problems that are frequently encountered in the interpretation of comparative data are briefly discussed.