The role of interspecific competition as a key factor in the ecology of natural communities where species exploit limited resources is well established, and the study of competition dynamics in snake communities has received much attention in recent years. Twenty years ago, an acclaimed review (Toft 1985) suggested that snakes were atypical among vertebrates because sympatric species usually partition the food niche. Here, I review the articles published in the last two decades with the aim of finding any general geographical or guild patterns and assessing if Toft's main conclusion is still supported by new evidence. Where appropriate, I use Monte Carlo simulations to establish whether observed patterns of niche overlap are real, or if they have occurred by chance. My study shows clear congruence in the patterns of coexistence exhibited by snake communities in different regions of the world, i.e.: (1) cold regions of the northern hemisphere (high latitudes and altitudes) exhibit low species richness and a very low, or even absent, potential for interspecific competition; (2) aquatic snakes that form communities in temperate regions generally partition the food type available and exhibit a broad similarity in habitat use with subtle differences in microhabitat use; (3) terrestrial snake communities in temperate regions are very variable in terms of their coexistence dynamics and show no evidence of generalised patterns; (4) sympatric viperids in Europe, North America and, most interestingly, tropical Asia partition the available habitat but not the prey resource; (5) competition is much stronger in tropical snake communities, and the intensity of this process fluctuates throughout the year being most intense during periods of low food availability; (6) in general, tropical snakes partition the food resource (prey type and/or prey size), but when this resource is not partitioned competitive exclusion can occur.
Prey resource availability is a fundamental variable for all snake communities; this is clearly documented by studies on terrestrial snakes in Australia where, due to a relative scarcity of prey availability in the field, sympatry among species is much rarer than in other continents. I conclude that, although there are several notable exceptions, Toft's main conclusion is still supported by empirical evidence. However, I disagree with Toft's conclusion that most snakes are food specialists, and I contend that interspecific competition is important in structuring many (if not most) of the snake communities around the world.