Sex ratio variation and mixed pairs in roe deer: evidence for control of sex allocation?

Sex allocation has provided rich ground for the development of evolutionary theory. The dominant models in vertebrates have provided predictions of sex ratio based on asymmetry in variance in breeding success between sexes in the breeding system, and the relative effect of local competition. In ungulates, empirical work has provided some support for these models, but has also generated apparently contradictory observations. Recent models have provided some predictions for both individual and population sex ratio, showing that the availability of high-quality habitat patches and dispersal rates can critically affect both population- and individual-level expectations. We explore patterns in offspring sex ratio among a large sample of roe deer, a species with some interesting aspects of the breeding system. We found that in singleton pregnancies (the minority) there was an excess of male embryos, which was more marked in does in good condition at most sites. While the sex ratio among twins was close to parity, and unaffected either by average condition among populations or by individual condition within populations, we observed a clear excess of mixed pregnancies (sub-binomial variance, SBV). The excess was greater where the average condition was high. Within sites, population changes in average condition were associated with changes in SBV: in populations declining in condition, SBV also declined. There was no tendency for mixed twins to be more likely in individuals in better condition within a population. We conclude that condition-dependent allocation of sex does occur in roe deer. We suggest that in these "income breeders", mixed pairs might offer the greatest opportunity to optimize maternal investment during lactation, when conditions will be unpredictable, and that this is increasingly the case when average condition is poor.