After the publication of this work , other researchers independently discovered that some of the data deposited in the NCBI Trace Archive was labeled erroneously. In particular, the sequencing center responsible for two of the Drosophila genome projects (Agencourt Biosciences) mistakenly deposited 20,000 sequences from D. ananassae and labeled them as D. mojavensis. The center recently corrected the mistake by removing the mislabeled sequences from the Trace Archive. We then searched through the newly updated D. mojavensis sequences for the 114 Wolbachia sequences that we had originally reported, and found that all had been removed. Thus our article should be corrected to report that new Wolbachia genome sequences were discovered in D. ananassae and D. simulans, but not in D. mojavensis.
While searching the Trace Archive to verify this correction, however, one of us (S.L.S.) found that the traces for a new fly sequencing project, that of D. willistoni, had just been deposited. On searching the D. willistoni traces, a substantial Wolbachia infection in this species was discovered and 2,291 sequences belonging to Wolbachia were found. They were assembled into 485 contigs using the comparative assembler AMOS-Cmp  and the methods described in . These sequences and assemblies are freely available for download from .
We thank Therese Markow of the University of Arizona for bringing this error in the Trace Archive data to our attention, and Jack Werren of the University of Rochester for suggesting that D. willistoni might have a Wolbachia infection.