Measuring the degree of sexual segregation in group-living animals

1. So far, no measure exists to quantify sexual segregation in animal populations. Studies on segregation have relied on ecological measures of overlap and association to estimate the degree of segregation. 2. However, existing ecological measures of overlap and association are stochastically related to sex ratio. population density or group size. These stochastic relations can lead to confounding results, making the existing measures unsuitable for quantitative studies on segregation. 3. In the present paper a new measure of segregation, the 'segregation coefficient', is suggested, which is free of stochastic relations. 4. The segregation coefficient is suitable for quantitative studies on segregation. It also makes possible, for the first time, comparisons between the degree of social segregation and the degree of habitat or spatial segregation in a population of animals.