Successful recovery of populations of rare and cryptic species requires accurate monitoring of changes in their distribution and densities, which in turn necessitates considering detection rates. Development of population monitoring protocols is needed to aid recovery of the New England cottontail (Sylvilagus transitionalis; currently the top-priority Species of Greatest Conservation Need in the northeastern United States), which lives in dense shrubby habitat and is difficult to detect. To address this need, we conducted repeated, systematic, presence-absence surveys to determine patch-specific detection probabilities and factors influencing detection of the New England cottontail. We surveyed cottontails during 2-6 visits on 30 sites with known occupancy in the northeastern United States during the winters of 2010 and 2011. For each survey visit, we determined whether cottontails were detected by the presence of fecal pellets on fresh fallen snow and subsequent species identification by genetic analysis. Detection probabilities were modeled in Program PRESENCE to explore the influence of snow condition and depth, temperature, wind, number of pellet deposition days, woody stem density, patch size, and search effort. The overall probability of detecting a New England cottontail during a single survey visit was 0.73. The most influential factor in detection was prior knowledge of site-specific cottontail activity. Snow depth <30.5 cm and the number of days without high winds following a snowfall had a positive influence on detection. Patch size had a negative effect on detection when surveys were restricted to 20 minutes. In the absence of prior knowledge, 2-3 surveys conducted with snowpack <30.5 cm and 2-4 days after a snowfall without high wind should yield reliable occupancy status with 95% confidence in detection. Incorporating our recommendations into monitoring programs will improve the accuracy of patch-specific occupancy data for New England cottontail. Our approach and findings may be applicable to monitoring other rare, cryptic, or threatened species that occupy dense habitats, especially where patch-level occupancy knowledge is required. Ó 2014 The Wildlife Society.