Gestation length variation in a wild ungulate

P>1. In species inhabiting a seasonal environment, timing of birth may be highly correlated with fitness. However, we know little about the relative effects of conception date and gestation length in determining birth date in wild mammalian populations. 2. Here, we tested for variation in gestation length using long-term data from a wild ungulate population (red deer, with observations running from 1971 to 2007). High March temperatures during gestation were associated with shorter average gestation length (a decrease of 0 center dot 77 days for every degree warmer), but we found no significant correlations with female age, recent reproductive history or offspring sex. 3. We also tested whether early or late conceiving females adjusted gestation length to give birth closer to an optimum time. Simply fitting conception date as a fixed effect in a model of gestation length is not statistically sound as errors in determining conception date could drive an observed negative covariance; consequently we also approached this problem using a novel technique of bivariate models. We found that late conception was indeed correlated with shorter gestation length, but this was not repeatable within females, suggesting either that the observed association may have been a statistical artefact or that gestation length adjustment may only be apparent on the rare occasions that females conceive outside an optimum time window. 4. Although there were no consistent differences between females in gestation length, there were differences between males in the gestation lengths of females with whom they mated. This possibly reflects a local environment effect, grouping females who were primarily in the same area during conception and gestation. 5. We conclude that observed variation in birth date is predominantly due to systematic variation in conception date as opposed to gestation length. Additionally, we highlight the broad range of reproductive tactics utilized by a number of closely related species, and consequently the difficulties inherent in cross-species comparisons of gestation length.