Abstract Alterations at the X‐linked Hmr gene of Drosophila melanogaster can fully restore viability and partially restore fertility in hybrid flies from crosses between D. melanogaster and any of its three most closely related species. Although more than one gene is expected to be involved in these barriers to reproduction, a single DNA‐binding protein was recently identified as HMR. The Hmr gene was shown to evolve unusually fast, a feature that supports its role in causing genetic incompatibility in a hybrid genotype. The current treatment of hybrid genetics focuses not only on Hmr but also on the Rab9D gene, which lies only 1kb from Hmr. Rab9D is proposed also to influence hybrid viability. This gene has remained tightly linked to Hmr for about 10 million years, but it has diverged even more than Hmr with regard to D. melanogaster and its most closely related species. Furthermore, the 197‐amino acid RAB9D protein contains four amino acid substitutions in the D. melanogaster‐rescuing mutant Hmr1. Rab9D is shown to have evolved under very strong positive selection and to be the most recent member of a cluster of six paralogs that encode small RAB GTPases. Four of the six paralogs are unique to D. melanogaster in which they have diverged considerably, their encoded proteins sharing less than 50% amino acid identities with proteins from their orthologs in the closest species. Only two Rab orthologs are present in these sibling species and none is present in the genomes of more distantly related Drosophila species. Rapidly evolving Rab paralogs near the Hmr locus probably developed functional specialization of redundant proteins involved in trafficking macromolecules between cytoplasm and nucleus. Positive selection acting on duplicates of these Rab genes appears to participate in reproductive isolation.