SummaryTheories on the evolution of litter size among organisms were reexamined. The competition theories, including that based on the r−K‐selection hypothesis, could not explain well why low‐fecundity strategies have often evolved in stressful environments such as mountain streams, deep sea and the antarctic, where interspecies competition is considered to be lax. The theory, based on Itô's (1980) concept of the procurability of food by the young, was considered to have greater generality because it could explain not only the above‐mentioned cases but also those where small litter size is observed in habitats with high species diversity (where interspecific competition may be keen), such as tropical rain forest. Examination of the process of selection of high‐fecundity and low‐fecundity genotypes also suggested that the procurability of food by the young can best explain the evolution of low‐fecundity. The concept of density‐induced dispersal and a distinction between density‐dependent and density‐independent predation pressures should be incorporated into our discussions on the evolution of reproductive rates.