SummaryA discussion of speciation subsumes that we know what a species is, which operationally is true. Populations that are physiologically incapable of gene exchange and are phenotypically distinguishable pose no problem at the species level. When barriers are not absolute or when they are complete but the phenotypic differences are inconspicuous one must consider the quantitative and qualitative nature of the barriers to gene exchange relative to the number and conspicuousness of the phenotypic differences. In this context barriers to gene exchange and their origin are discussed with special emphasis placed on ecological differentiation and its role in the maintenance of the distinction between species. Speciation involving chromosomal alteration differs markedly from that resulting from genetic divergence alone and the products of the two modes are compared. The argument is made that an understanding of the mode of speciation can be of help in the interpretation of patterns of variation.