The collection of Drosophila melanogaster from Zimbabwe and nearby regions (the Z-type) yield females who would not mate with the cosmopolitan D. melanogaster males (the M-type). To dissect the genetic basis of this sexual isolation, we constructed 16 whole-chromosome substitution lines between two standard Z- and M-lines. The results were as follows: (1) All substitution lines appear normal in viability and fertility in both sexes, indicating no strong postmating isolation. (2) The genes for the behaviors are mapped to all three major chromosomes with the same ranking and comparable magnitude of effects for both sexes: III > II >> X > or = 0 (III, II and X designate the effects of the three chromosomes). The results suggest less evolution on the X than on autosomes at loci of sexual behavior. (3) The genes for "Z-maleness" are many and somewhat redundant. Whole-chromosome effects for Z-maleness appear nearly additive and show little dominance. (4) In contrast, "Z-femaleness" has less redundancy as partial genotypes never exhibit full phenotypic effects. Epistatic interactions and incomplete dominance can sometimes be detected. (5) The extensive genetic divergence underlying sexual isolation has evolved in the absence of detectable reduction in hybrid fitnesses. Sexual selection has apparently been a driving force of multiple facets of speciation at the nascent stage without reinforcement.