We have analyzed the structures of 19 mutant alleles at the white locus of Drosophila melanogaster. Thirteen of the mutant alleles in our selected sample arose spontaneously, and of these, seven are associated with insertions of non-white-region DNA sequence elements. Several lines of evidence strongly suggest that these insertions are responsible for their associated mutant alleles, and further suggest that most or all of these insertions are transposons. Moreover, the white locus DNA sequences can be divided into two nonoverlapping domains on the basis of the properties of the two domains as mutational targets. One of these domains behaves, in this regard, in the manner expected of functional coding sequences, whereas the other does not. We propose a model for the nature and function of the presumptive noncoding white locus genetic elements. The two domains of the white locus defined by our studies are approximately coextensive with the functionally distinct subintervals of the locus defined by previous genetic analysis. Lastly, our results strongly suggest that the dominant, mutable wDZL allele results from the insertion of a transposon outside of, but near, the white locus. This putative transposon apparently carries genetic elements that act at a distance to repress expression of the white locus.