Expansion of the black bear (Ursus americanus) population outside White River National Wildlife Refuge, Arkansas, has caused local citizens to request reduction of the population. An estimate of black bear population size is necessary before any management recommendations can be made. We used 2 census methods to estimate population size. Hunter surveys of bear sightings during the 2-day firearm deer hunt were used in 1994 to estimate population size with a Petersen-Lincoln estimator. The survey yielded a population estimate of 213 bears (95% CI = 129-297). We used cameras triggered by infrared monitors to photograph bears visiting bait sites over a 7-day period during August 1995. Infrared monitors recorded 176 events, resulting in 87 photographs of 23 different bears. We used 2 separate models to calculate population sizes using data from cameras. Program CAPTURE calculated a population size of 348 (95% CI = 300-396) bears for the refuge. A Minta-Mangel mark-resight model calculated a population size of 464 (95% CI = 348-638) bears for the refuge. The hunter survey population estimate was low even with the probable violation of the assumption that tags were not lost. Population estimates based on the photographs were different because the actual number of bears on the study area of sampling is unknown. Problems with the hunter survey and Minta-Mangel population estimates provided evidence that the program CAPTURE population estimate is the most reliable.