During a study of fawn mortality, we investigated proximate factors affecting birth dates of sympatric desert mule deer (Odocoileus hemionus eremicus) and white-tailed deer (O. virginianus texanus) in west-central Texas from 2004 to 2006. We treated this aspect of the study as time-to-event survival (i.e., pregnancy to birth) and modeled the process with accelerated failure-time regression. Our best model included effects from 3 hierarchal levels: within-year variation among individuals within species, because older and heavier females gave birth earlier; among-year variation at the population level, because greater rain during the previous prerut and rut periods resulted in earlier birth dates; and a chronic-cohort effect also at the population level, because even after previous effects were accounted for in regression models, deer gave birth later on more intensely grazed ranches. After accounting for mass, age of females as a significant predictor may have indicated a behavioral phenomenon associated with social dominance. We did not find meaningful relationships between birth dates and either offspring sex or rain during gestation. Overall, Kaplan-Meier product-limit estimates indicated that birthing by white-tailed deer peaked on 20 June (90% range = 31 days) and birthing by mule deer peaked on 21 July (90% range = 45 days). We suggest that the 1-month separation between peak birth dates and breeding periods of these sympatric species of deer was partly due to phylogenetic constraint from parent populations and not localized adaptation with selection against hybridization. Prevention of genetic introgression may be a result by coincidence.
- Accelerated failure-time
- Proximate factors