Sexual segregation is common in ungulates. We show, in a high latitude population of feral goats where behavioural synchrony and fission rates have been shown to be the best explanation for segregation, that it is differences explicitly in the feeding time requirements of the two sexes (but not those for other activities) that best explains the variations in monthly frequencies of segregation. However, this effect is less marked during winter months when short day length forces the time budgets of the two sexes to converge. We argue that the various explanations for segregation can best be interpreted as separate factors in a multivariate model in which species- and habitat-specific weightings influence the relative importance of these variables, and thus the likelihood that segregation will occur.