In eusocial Hymenoptera, relatedness asymmetries lead to conflict between parents and offspring over sex investment; workers strive for a 3:1 female—biased ratio of sex allocation, and queens strive for a 1:1 ratio. Many studies support this genetic relatedness hypothesis, but the variation in allocation ratios in natural populations is great and remains mostly unexplained. In this paper, we examined whether food supply determines sex investment of the ant Formica podzolica. We compare reproductive parameters of colonies and populations across habitats, and compare sex allocation of fed and unfed colonies. Nest density, worker size, sex ratio, and sex allocation were all greater along forest edges than in meadows, and these patterns were associated with natural food abundance. Furthermore, there was a strong tendency for individual colonies to produce either all—male or all—female sexuals, and worker size was greatest in colonies producing all females and smallest in those producing all males. Most important, sex investment was greatly affected by a supplemented diet, as the population investment ratio, R (i.e., ratio of males to males plus females), was female biased (0.36) for fed colonies and male biased (0.62) for unfed, control colonies. Finally, investment ratios were more male biased in polygynous than in monogynous colonies as predicted by a genetic relatedness hypothesis. These results demonstrate that food supply has an important proximate influence on sex investment, and may explain much of the natural variation in sex investment in populations of eusocial Hymenoptera.