A thorough test of the hypothesis of compensatory mortality is a fundamental requirement for a better understanding of the population dynamics of wildlife species. This knowledge is vital, whether populations are managed for recreational hunting or other purposes. Our research on a pinyon pine (Pinus edulis)-Utah juniper (Juniperus osteosperma) winter range in Piceance Basin, northwest Colorado, from 1981 to 1988 tested for compensatory mortality in the fawn portion of a mule deer (Odocoileus hemionus hemionus) population. Three experimental manipulations were used employing radio-collared deer. In a field study, removing 16-22% of the population from the treatment unit each winter for 2 years had no measurable effect on fawn survival rates as compared to rates on the control unit (P = 0.566). We attributed this mostly to not removing enough deer to immediately affect fawn survival under existing range conditions. In a controlled study, deer removed from the treatment unit were used to stock 3 large pastures at densities of 44, 89, and 133 deer/km2 to simulate hunting removals of 67, 33, and 0%, respectively. Fawn survival rates varied inversely with density (P < 0.001). Starvation was the leading cause of fawn mortality in all pastures indicating a nutritional limitation at all densities. We believe the density-dependent survival response in the pastures demonstrated that a strong compensatory mortality process operated in this mule deer population. In another field study, 49-77% of fawns were killed by predators during 4 winters. We then reduced the coyote (Canis latrans) population for 3 winters while we continued to monitor fawn mortality. Predation rates decreased (P = 0.004) and starvation rates increased (P = 0.042) between pre- and posttreatment periods, but no change in fawn survival was detected (P = 0.842). These results support those from the pastures even though the primary mortality causes differed. Mean fawn weights varied among years, study areas, and trap sites (P < 0.001). Male fawns averaged 2.4 and 3.0 kg heavier than females (P < 0.001) on the 2 field-study areas. Larger fawns had higher survival (P < 0.001), but size was not a significant predictor of whether or not a fawn starved (P = 0.237). In both field-study areas, female fawns had higher survival than males (P < 0.001) when weight was a covariate, but not when weight was excluded (P = 0.697). Adult females had higher survival rates than fawns (P < 0.001) even though adult rates were calculated over 5.5 more months. Vegetation biomass differed among pastures (P < 0.001), but differences were unrelated to fawn survival rates. Biomass estimates indicated adequate forage was available in all pastures. Tame deer in the low density pasture took more bites per 15-minute trial (P < 0.001), had shorter mean times between consecutive bites of Utah serviceberry (Amelanchier utahensis) and true mountainmahogany (Cercocarpus montanus) (P less-than-or-equal-to 0.002), and traveled less distance during afternoon trials (P = 0.015) than tame deer in the high density pasture. These differences were assumed to reflect lower forage quality in the high density pasture. For the Piceance Basin mule deer population, mortality rather than reproduction seemed the major process driving the density-dependent mechanism because the former fluctuated over a much broader range. High survival of adult females, even during severe winters, tended to temper population fluctuations that can occur in harsher environments and allowed density-dependent processes in the fawn segment to continue operating. With density-dependent population regulation, the common management strategy of decreasing harvest when fawn survival is low and increasing harvest when survival is high is counterproductive.
