The Segregation Distorter system of Drosophila melanogaster consists of two major elements, Sd and Rsp. There are two allelic alternatives of Rsp-sensitive (Rsp(s)) and insensitive (Rsp(i)); a chromosome carrying Rsp(i) is not distorted. According to the model proposed by Hartl (1973), these two elements interact to cause segregation distortion. For a sperm to complete the maturation process, it is assumed that the Rsp locus has to be complexed with the product of the Sd locus. This product is assumed to be a multimetric regulatory protein. Three kinds of regulatory multimers may be distinguished: Sd(+)/Sd(+), which is assumed to complex with both Rsp(s) and Rsp(i); Sd(+)/Sd heteromultimers, which complex preferentially with Rsp(i); and Sd/Sd homomultimers, which complex with neither Rsp(s) nor Rsp(i). Most of the regulatory protein in the Sd(+)/Sd heterozygous male is assumed to be the Sd(+)/Sd heteromultimer.--Some modifications of Hartl's model were made by Ganetzky (1977). Rather than the binding of a product of Sd at the Rsp locus being a necessary condition for normal spermigenesis, this binding causes sperm dysfunction. It is assumed that the product of Sd complexes more readily with Rsp(s) than with Rsp(i) and that the amount of Sd product is limited with respect to the number of binding sites available. No function is ascribed to the Sd(+) locus. In order to explain reduced male fertility of some genotypes, Ganetzky further assumes that the Sd product, when not competed for by an Rsp(s) locus, can bind to an Rsp(i) locus.--Two consequences of these models were critically examined: according to these models (1) an Sd Rsp(s)/Sd(+)Rsp(s) male should not show any segregation distortion, and (2) an Sd Rsp(s)/Sd Rsp(s) male should show much reduced fertility, if not complete sterility.--The results of the present study bear on these two points. (1) Rsp(s) locus seems to consist of multiple alleles, each having a different degree of ability to interact with the product of the Sd locus. An Sd Rsp(s)/Sd(+)Rsp(s) male shows a certain degree of segregation distortion when the two Rsp(s) alleles are different, but it shows a normal Mendelian segregation ratio when the Rsp(s) alleles are homozygous. The first prediction of the models is supported by actual observation when the two Rsp(s) alleles are the same. (2) There is a suggestion of slight reduction in fertility, but generally Sd Rsp(s)/Sd Rsp(s) males are quite fertile. Thus, the second prediction is not supported by actual observation. The mechanism of segregation distortion is still open for future studies.