Certain recent models of sex determination in mammals, Drosophila melanogaster, Caenorhabditis ekgans, and snakes are examined in the light of the hypothesis that the relevant genetic regulatory mechanisms are similar and interrelated. The proposed key element in each of these instances is a noncoding DNA sequence, which serves as a high-affinity binding site for a repressor-like molecule regulating the activity of a major "sex-determining" gene. On this basis it is argued that, in several eukaryotes, (i) certain DNA sequences that are sex-determining are noncoding, in the sense that they are not the structural genes of a sex-determining protein; (it) in some species these noncoding sequences are present in one sex and absent in the other, while in others their copy number or accessibility to regulatory molecules is significantly unequal between the two sexes; and (fii) this inequality determines whether the embryo develops into a male or a female.Sex determination in Drosophila melanogaster (1-4) and Caenorhabditis elegans (5) appears to be based on interactions among a small number of genes, with the level of activity of one gene being of primary importance. Consequently, it becomes feasible to look upon the choice of sexual phenotype in development as being mediated through a kind of genetic switch that opens up one or the other of two alternative pathways. If so, one is then led to inquire about the nature of the regulatory elements that might constitute such a switch.If sex determination is based on a small number of critical genes and on an "either/or" mechanism, it is reasonable to expect both single-gene mutations and, occasionally, environmental perturbations to shift sexual development from one pathway to the other. Mutations that transform sex in this manner are known in D. melanogaster, in C. elegans, and in mammals. In some reptiles sex is determined by the temperature at which egg incubation occurs. The sharp temperature thresholds seen in these experiments (6, 7) and the absence of intersexes or significant egg mortality suggest that not only a small number of genes might be involved but also a high degree of cooperativity in the relevant regulatory processes.On the basis of a recent model for mammalian sex determination (8, 9) and related results (2), I wish to suggest that certain noncoding DNA sequences could be responsible for the choice of sexual phenotype during development. The term "noncoding" sequence is used in this paper to mean that the DNA sequence is not the structural gene for a sexdetermining protein. Rather, its role in sex determination is to bind a repressor-like molecule and, thereby, to regulate indirectly the activity of a "sex-determining" gene. "Repressor" denotes a regulatory molecule that has high affinity for the noncoding sequence(s) but whose primary role is in controlling the activity of a major sex-determining gene.THE MODELS Mammals. There are at least three observations that must be faced by any model of primary sex determination in mammals: (i) the failure...