Previous work on transposable element distribution in colonizing populations of Drosophila buzzatii revealed a high frequency of occupancy in several chromosomal sites. Two explanatory hypotheses were advanced: the founder hypothesis, by which founder genetic drift was responsible, and the unstable hypothesis that assigns this unusual distribution to bursts of transposition toward some chromosomal sites. Here, we study the molecular structure of three euchromatic Osvaldo clones isolated from sites occupied at high (A4 and B9) and low frequency (B4) in colonizing populations, to test these hypotheses. Large insertions, duplications, and indels in the Osvaldo coding region and LTR were detected in the A4 clone and a truncated Osvaldo with many substitutions was found in the B9 clone. These altered sequences indicate that the two copies of this retro-element are precolonization insertions. Interestingly, the LTR of the A4 clone and the reverse transcriptase region of B9 show identical sequences in all colonizing populations indicating, most probably, that they are identical by descent. Moreover, Osvaldo is inserted at the same nucleotide site in all colonizing populations. On the other hand an almost identical LTR sequence, except by 1 base deletion, was found in the B4 clone compared to the canonical active Osvaldo element. These results suggest that Osvaldo copies in highly occupied sites are,mostprobably, identicalbydescent andstrongly favor the founderhypothesis. On the other hand, low-insertion-frequency sites could represent recent transposition events. This work emphasizes the importance of molecular population studies to disentangle the effects of genetic drift and transposition in colonization. Copyright © 2007 by the Genetics Society of America.