The colonization of the New World by the Palearctic species Drosophila subobscura was first detected in 1978 in South America and around 1982 in western North America. The ensuing dramatic expansion of the species, in territory as well as numbers, provides an opportunity for studying evolution in a scale rarely possible. We have used 10 restriction endonucleases to analyze the mitochondrial DNA (mtDNA) of individuals from 23 widely dispersed localities. Only two mtDNA composite morphs have been detected in the Americas. None of the two morphs has been found in Africa, and only one in the Atlantic islands; but both are widespread in Europe, which provides no clue of the precise geographic origin of the colonizers. The amount of nucleotide-substitution polymorphism detected in D. subobscura is typical for animals, but it is greater in the Old than in the New World, presumably due to the recent colonization by a limited number of colonizers. Assuming standard evolutionary rates of mtDNA base substitution, the mtDNA morphs found in D. subobscura can be traced to a single one that existed no less than one million years ago. We argue against the inference that the D. subobscura flies now living descend from only one or a few females that lived at that time. This type of inference, which we call the "Mother Eve hypothesis," has been made to conclude that the human population went through a severe constriction about 200,000 years ago, so that all living humans descend from only one or a few women who lived at that time. The Mother Eve hypothesis is fallacious.