FOUNDER-FLUSH SPECIATION IN DROSOPHILA-PSEUDOOBSCURA - A LARGE-SCALE EXPERIMENT

A founder-flush-crash model of speciation has been proposed that may P articularlY apply to island and other colonizations. Previous laborato ry experiments testing the model have given inconsistent results. We h ave conducted a large experiment with Drosophila pseudoobscura designe d to meet the essential postulates of the model and to separately test some of the postulates. Forty-five experimental and 12 control popula tions have been studied during seven successive founder-flush-crash cy cles, or about 50 generations. Sexual isolation tests yield significan tly positive assortative mating in a few tests between pairs of experi mental populations. Populations with fewer founders (N = 1 or 3) yield more significant instances of assortative mating than those with more founders (n = 5, 7, or 9), and this difference becomes statistically significant for pooled data. Only one of 15 population pairs tested th ree times (generations 25, 32, and 46) shows positive assortative mati ng in all three tests. No significant assortative mating occurs betwee n control populations, including highly inbred ones. We conclude that although founder events may occasionally lead to the evolution of asso rtative mating and hence to speciation, our results do not support the claim that the founder-flush-crash model identifies conditions very l ikely to result in speciation events.