A synthetic theory of the ecology of amphibian metamorphosis is founded on the observation that the large variation in length of larval period and body size at metamorphosis typical of a particular species of amphibian cannot be directly explained by differences in dates of hatching or egg sizes. It is proposed that as development proceeds, variation in exponential growth coefficients causes a trend from a normal distribution to a skewed distribution of body sizes. The degree of skewing increases and the median of the distribution decreases with increasing initial densities of populations. The relative advantages of the largest members of a cohort may arise from a variety of mechanisms including the production of growth inhibitors, interference competition, and size-selective feeding behavior. These mechanisms result in a nonnormal distribution of competitive ability, a possible source of the density-dependent competition coefficient found in systems with many species (1). In our model the ranges of body sizes and dates of metamorphosis are determined by a minimum body size that must be obtained and a maximum body size that will not be exceeded at metamorphosis. Between these two size thresholds the endocrinological initiation of metamorphosis is expected to be related to the recent growth history of the individual larva. Species that exploit uncertain environments will have a wide range of possible sizes at metamorphosis. Species exploiting relatively certain environments will have a narrower range. The evolution of neoteny and direct development logically follow from the application of these ideas to the ecological context of the evolution of amphibian life histories. Species that live in constant aquatic habitats surrounded by hostile environments (desert ponds, caves, high-altitude lakes) may evolve permanent larvae genetically incapable of metamorphosis. Other populations may evolve a facultative metamorphosis such that populations are a mixture of neotenes and terrestrial adults. Direct development results from selection to escape the competition, predation, and environmental uncertainty characteristic of some aquatic habitats and is usually accompanied by parental care. The relation between our ecological model and the physiological mechanisms that initiate metamorphosis can only be suggested and it remains an open problem for developmental biologists.