Despite countless biogeographic, ecological and molecular data, the origin, age and palaeogeographic trajectory of Drosophila melanogaster or D. simulans have remained highly debatable. Even the most widely accepted views, like the place and antiquity of the wild-to-domestic behavior shift in D. melanogaster, and which of the two species arose first appear questionable. Here, we present a critical review of the conflicting hypotheses, and make testable suggestions. Although both species are cosmopolitan human commensals, they have a contrasted biogeography, ecology and history. The disappearance of the wild behavior in some or all D. melanogaster populations may date back to the 18,000 years BP post-Aterian hyperarid phase. Alternatively, D. melanogaster populations where the wild behavior would still prevail may still exist in Central Africa. The 600-fold reduction in surface that affected the Seychelles Bank 10,000 years ago, undoubtedly affected D. simulans and D. sechellia population sizes dramatically. We stress the good match between the geographical ranges of D. simulans and Morinda citrifolia, and the mismatch between D. sechellia and this rubiaceous plant, yet assumed to be its restricted host-plant. We suggest that the ecological status of the four species of the melanogaster complex may represent four steps of the same process indicating a gradual shift from specialization and local confinement to opportunism and human commensalisms.