Are new species formed in rare catastrophes, distinct from the normal processes of phyletic evolution? Or does reproductive isolation evolve gradually, as a by-product of the divergence of gene pools? Mayr (120-124) has argued the former, holding that speciation usually results from genetic revolutions triggered by founder effects: An isolated population, small in numbers and in geographic extent, colonizes a new area. Both changes in selection pressures and genetic drift result in the rapid shift of many genes to a new, coadapted combination, which is reproductively isolated from the ancestral population. Carson (27, 29, 3 1) and Templeton (I175-180), among others, have put forward similar models. This cluster of theories is woven from many strands; we will try to tease these apart in order to find out precisely which processes may be involved in speciation by founder effect. By placing them in the context of other models, we will argue that, although founder effects may cause speciation under sufficiently stringent conditions, they are only one extreme of a continuous range of possibilities. Complete geographic isolation is unnecessary; absolute coadaptation between closed systems of alleles is unlikely; and divergence may be driven in a variety of ways, without the need for drastic external changes. Reproductive isolation is most likely to be built up gradually, in a