Stebbins supported Epling's hypothesis and argued that the inversion system was quite old, while Mayr disagreed with Epling and favored a more recent origin of the distribution pattern. Simpson concluded that the age of the system could not be determined from the available evidence. It thus seems appropriate in this colloquium honoring the 50th anniversary of Simpson's seminal work to address a question about the history of the D. pseudoobscura inversions, not quite the question of age that Simpson considered, but rather the related question of which gene arrangement was the ancestral one.
In D. pseudoobscura, the third chromosome is polymorphic for more than 40 gene arrangements resulting from overlapping paracentric inversions (Olvera et al., 1979), which can be ordered in a phylogeny based on the breakpoints of inversions under the parsimonious assumption that each inversion arose only once (Figure 1). With the single exception of Hypothetical, all of the gene arrangements necessary to reconstruct the complete phylogeny have been observed in nature. Four of these arrangements—Standard (ST), Hypothetical (HY), Santa Cruz (SC), and Tree Line (TL)—are central to the phylogeny, with all the others being their one- or two-step derivatives. The tree in Figure 1 is unrooted, and the question of which of these four gene arrangements is ancestral has remained unanswered for more than 50 years, inasmuch as cytogenetic, biogeographic, and electrophoretic data have not consistently supported a single hypothesis.
ST was the first arrangement proposed as ancestral (Sturtevant and Dobzhansky, 1936), because it is the only one shared by D. pseudoobscura and its sibling species, Drosophila persimilis. HY was suggested (Dobzhansky and Sturtevant, 1938) because its inverted region resembles the banding pattern of the homologous chromosome in Drosophila miranda , a related species more distant from D. pseudoobscura than is D. persimilis. Historically, SC has received less attention than the other arrangements, although it too has been suggested (Dobzhansky and Epling, 1944, Brown, 1989). More recently, TL has been considered a favorite candidate for the ancestral gene arrangement on the basis of its distribution pattern (Wallace, 1966; Olvera et al., 1979; Wallace, 1988) and comparison of alleles at protein loci in D. pseudoobscura, D. persimilis,