We evaluated naive density estimates and density estimates corrected for “edge effect” using mean maximum distance moved (MMDM) for small mammal populations by Monte Carlo methods. Two densities, 25 and 100/ha, were generated in random or slightly clumped spatial patterns within a 4-ha area and populations had average capture probabilities of either 0.16 or 0.24 allowing variation in time, behavior, and heterogeneity. Animals were assumed to have a bivariate normal utilization distribution of either 0.25, 0.5, or 0.75 ha. An 18 by 18 trapping grid with 7 m trap spacing was simulated with trapping over 6 or 8 occasions. Evaluation of 1,393 repetitions divided among 8 different cases revealed a large positive bias (69–89%) for the naive density estimates, and density estimates by using the MMDM had less than 22% bias. A robustness to home range size was demonstrated by the MMDM. Difficulties with both methods are indicated.