We evaluated a test and cull strategy for lowering chronic wasting disease (CWD) prevalence in a naturally-infected, free-ranging mule deer (Odocoileus hemionus) herd wintering in the town of Estes Park, Colorado, US and in nearby Rocky Mountain National Park. We tested 48−68% of the estimated number of adult (≥1 yr old) deer annually for 5 yr via tonsil biopsy immunohistochemistry (IHC), collecting 1,251 samples from >700 individuals and removing IHC-positive deer. Among males, CWD prevalence during the last 3 yr of selective culling was lower (one-sided Fisher's exact test P=0.014) than in the period prior. In contrast, CWD prevalence among females before culling and after culling were equivalent (P=0.777). Relatively higher annual testing of males (mean 77%) compared to females (mean 51%) might have contributed to differences seen in responses to management. A more intensive and sustained effort or modified spatial approach might have reduced prevalence more consistently in both sexes. Limitations of this technique in wider management application include cost and labor as well as property access and animal tolerance to repeated capture. However, elements of this approach could potentially be used to augment harvest-based disease management.