Three gene systems have been shown to exhibit proximity-dependent phenotypes in Drosophila melanogaster: bithorax (BX-C), decapentaplegic (DPP-C) and white (w). In structurally homozygous genotypes, specific allelic combinations at these loci exhibit one phenotype, while in certain rearrangement heterozygotes the same allelic combinations exhibit dramatically different phenotypes. These observations have led to the suggestion that, through the process of somatic chromosome pairing, such loci are brought into sufficient proximity to permit effective passage of molecular information between homologues; rearrangement heterozygosity would then displace the homologues relative to one another such that this trans-communication is obviated. The genetic properties of the proximity-dependent allelic complementation (termed transvection effects) at the BX-C and DPP-C, are quite similar. Chromosomal rearrangements which disrupt transvection possess a breakpoint in a particular segment of the chromosome arm bearing the transvection-sensitive gene (arm 2L for the DDP-C and 3R for the BX-C); this segment of each arm has been termed the critical region by Lewis (1954). As determined by cytogenetic analysis of transvection-disrupting rearrangements, the critical regions for the BX-C and DDP-C transvection effects extend proximally from these loci for several hundred polytene chromosome bands (Lewis 1954; Gelbart 1982). The interaction between the zeste and white loci appears to depend upon the proximity of the two w+ alleles. By use of insertional duplications, displacement of w+ homologues has been shown to interfere with the zeste-white interaction. In contrast to transvection at bithorax and decapentaplegic, however, only breakpoints in the immediate vicinity of the white locus can disrupt the zeste-white interaction (Gans 1953; Green 1967; Gelbart 1971; this report). In this report, we investigate the basis for the difference in the size of the BX-C and DPP-C critical regions from that of white. We test and eliminate the possibility that the difference is due to the presence near the white locus of a site which mediates somatic chromosome pairing. Rather, our evidence strongly suggests that the zeste-white interaction is, at the phenotypic level, much less sensitive to displacement of the homologous genes than is transvection at either the BX-C or DPP-C. We also show that many of the breakpoints in the vicinity of the white locus do not behave as if they are disrupting a critical region for somatic chromosome pairing. Given these results, we suggest that the zeste-white interaction and transvection are two different proximity-dependent phenomena.