The recent colonization of North America by Drosophila subobscura has provided a great opportunity to analyze a colonization process from the beginning. A comparative study using 10 microsatellite loci was conducted for five European and two North American populations. No genetic differentiation between European populations was detected, indicating that gene flow is high among them and that the microsatellites used in the present work represent neutral markers not subject to differentiation due to selection. Extensive reduction in the number of alleles and a significant decrease in heterozygosity in colonizing populations were detected that could be explained by the founder effect and a subsequent quick but not infinite expansion. Assuming that all alleles present in the colonized area were carried by the sample of colonizers, we estimated that most probably 4-11 individuals expanded in the new area. F(ST) and the chord distance measures reflect the colonization process more accurately, since drift has been the major force in differentiating the Old and New World populations, and thus other measures considering allele size differences, such as Rho(ST) and deltamu2, are less reliable for studying nonequilibrium populations. Finally, our results were consistent with the two-phase microsatellite mutational model, indicating that most alleles are generated by gain or loss of a repeat unit, while some alleles originate by more complex mutations.