HAMILTON’s rule and the principle of inclusive fitness’ provide a theoretical basis for understanding the evolution of social behaviour, and a framework for predicting reproductive characteristics of social insect colonies2-4. Sex allocation in social insects (especially ants) has become a central factor in tests of inclusive fitness theory5-9. The most powerful such test is the analysis of individual colonies where the predicted sex allocation varies depending on variation in worker fitness functions10,11. Recently developed models5,12 predict that workers may enhance their inclusive fitness by biasing sex ratios in response to the degree of relatedness asymmetry in each colony. Here I provide the first empirical evidence of facultative sex ratio biasing in response to relatedness asymmetries caused by inter-colony variations in queen mating frequencies. In a Finnish population of the ant Formica truncorum, colonies have a single queen mated to one or several males. Colonies show a bimodal distribution of sex ratios, with a significantly greater proportion of males in colonies headed by a multiply mated queen.