News & Comment

breeding-opportunity interval is also likely to promote coexistence by repeatedly changing the balance between the competitive ability of fast larvae and the greater probability that slow adults have of finding patches. Such mechanisms would also produce correlations between patch frequency and the life histories of emerging species, and lead to the observed differences between the distribution of life histories among adults and recruits*. Nevertheless, this field evidence fails to falsify the life-history model and suggests that life-history differences might facilitate species coexistence in nature whether or not environmental heterogeneity is high. quently the model does not imply partitioning of the habitat nor of the niche. Coexistence relies entirely on the balance between the life-histories of the two species. It is unlikely that the attributes of more than two species could be balanced in this way and, given this form, the model predicts saturated communities with very few coexisting species. In the initial form of the model’, however, the number of coexisting species allowed may well be determined by the size of the variance attached to breeding opportunity interval, and several species might be accommodated. Such a community would be oversaturated (in the terminology of Cornell and Lawtons) since, if the variance were removed, it would collapse to the pair of coexisting species implied by the laboratory test and the non-stochastic model. Acknowledgements Ian Hardy and Andrew Davis wrote this article in the Drosophila Population Biology Unit, Dept of Pure and Applied Biology, The University of Leeds, Leeds UK LS2 9JT.