A population has been examined in which an overall parity between the sexes hides considerable between-family variation in sex ratio. A proportion of families show highly distorted sex ratios, with either an excess of females or an excess of males. Distorted sex ratios are invariably associated with mortality in the immature stages at a level appropriate to the action of recessive lethal genes. It has been shown that 26 ~ of M-bearing (Y) chromosomes and at least 24 ~ of m-bearing (X) chromosomes carry a recessive lethal gene. Two such genes have been investigated, l kills males and, in a cross between two heterozygotes, gives rise to a sex ratio close to 25:13 ("excess ~" families), k kills females and, in a cross between two heterozygotes, gives rise to a sex ratio close to 17:23 ("excess 3" families). Selection for "excess ~" or "excess 3" did not increase the level of sex ratio distortion.
No crossing over occurs between k and the M/m locus whereas l shows 5-I0~o recombination with M/m. A test for aUelism confirmed that l and k are not allelic.
The penetrance of k is complete whereas l shows somewhat less than full penetrance. The penetrance of l has been improved by selection. The high frequency of lethals remained in the population during the two year period of study. There was evidence for heterosis preserving this frequency, the heterozygotes living longer and producing more progeny. However lethals were no longer to be found after four further years of laboratory culture.