I investigated whether soil moisture affects relative fitness of females and hermaphrodites and sex ratio in a gynodioecious plant with nuclear-cytoplasmic sex inheritance. I contrast these results with those from species with strictly nuclear sex inheritance. I performed a manipulative watering experiment on seed fitness of the two sexes, and field studies measuring seed fitness and sex ratio as a function of soil moisture. In the dry site, watered hermaphrodites produced approximately twice as many seeds as unwatered hermaphrodites, with little treatment effect on female seed production. Over a natural soil moisture gradient, the ratio of female to hermaphrodite seed production was higher in dry than in wet sites. These data show that the seed fitness advantage of females is a function of soil moisture. Despite this, regression of soil moisture on the sex ratio of 23 populations was not significant. These results indicate a sex-dependent effect of soil moisture on resource allocation to seeds that does not translate into a strong effect on sex ratio. This is consistent with theory based on genomic conflict in which sex ratios are predicted to be only partly determined by fitness differences of the sexes.