Segregation Distorter (SD) is a naturally occurring system of meiotic drive in Drosophila melanogaster. Males heterozygous for an SD second chromosome and a normal homolog (SD+) transmit predominantly SD-bearing sperm. To accomplish this, the Segregation distorter (Sd) locus induces the dysfunction of those spermatids that receive the SD+ chromosome. Recently, P. A. Powers and B. Ganetzky isolated overlapping DNA clones spanning the region of DNA known to contain the Sd gene and identified a 5-kb tandem duplication that is present on all SD chromosomes examined, but is apparently absent from wild-type chromosomes. Here we report a molecular analysis of two spontaneous revertants from an Australian SD chromosome (SD-Arm28). Both of these revertants have lost the 5-kb tandem duplication along with the ability to distort transmission; the critical observation, however, is that they retain the DNA haplotype in the flanking regions (both proximally and distally) that is characteristic of the original SD-Arm28. We propose unequal sister chromatid exchange between the tandem repeats as the only plausible explanation for loss of a repeat while retaining flanking markers. This provides direct evidence that the tandem duplication is indeed necessary for the Sd phenotype. Further, we examined testes-specific levels of both RNA and protein for the nearby Topoisomerase 2 gene. Neither revealed a consistent difference between SD and SD+ strains. We also measured testes-specific levels of RNA using the tandem duplication itself as probe. Our results suggest that there is strong up-regulation of one or several 2.0-2.3-kb transcripts from the duplicated region in the testes of an SD strain.(ABSTRACT TRUNCATED AT 250 WORDS)