Summary: Butlin (1995)

The reinforcement hypothesis suggests that natural selection favours an increase in assortative mating (and thus progress towards speciation, if two divergent populations produce low fitness hybrids in the zone of contact. Butlin (1995) briefly reviews a model proposed by Liou & Price (1994) and an empirical study by Noor (1994). Liou & Price's (1994) study considered secondary contact in sympatry of two divergent stickleback populations, with three possible model outcomes (extinction of one population, permanent mixing of the gene pools or reinforcement and speciation). The model suggested  that reinforcement is likely because the diluting effects of gene flow are absent. Butlin (1995) notes, however, that Liou and Price (1995) did not distinguish between the conditions of zero hybrid fitness and reinforcement and argues that the most questionable aspect of this model is the genetic basis of selection against hybrids. Noor's (1994) study addresses the situation where there is a very low level of gene exchange between two species of Drosophila, finding that hybrid males were sterile but hybrid females were fertile. Noor (1994) found evidence for reinforcement, however, Butlin (1995) notes that, while the result is exciting because of testability, the observation is no stronger evidence for reinforcement than other examples. He does agree that this approach could provide more discriminating predictions, but a larger number of localities has to be examined. Finally, Butlin (1995) indicates that reinforcement is potentially important evolutionary process and suggests that further work will identify it's actual importance.