Estimating density and relative abundance of sloth bears

Estimates of abundance based on capturing, marking, and recapturing a small sample of bears are likely to be biased and imprecise, and indices of abundance are of little value if not verified with reliable population estimates. We captured and radiocollared 17 sloth bears (Melursus ursinus) in Royal Chitwan National Park, Nepal, but recapture rates were too low to derive a meaningful mark-recapture population estimate. However, we frequently observed these bears while radiotracking them, and in 47 instances saw another bear (not offspring) near the one that we were tracking; we used these sightings of associated bears as our "recapture" sample, thereby providing a means of estimating abundance. We divided estimates of abundance by the area of composite seasonal home ranges to obtain estimates of density. Densities varied by season and habitat (25-72 bears/100 km²). We extrapolated density estimates to obtain a population estimate for the entire park (=250 bears, excluding dependent young). We also evaluated incidence of bear-excavated termite colonies (holes in mounds or in the ground) as a potential index of bear density. This proposed index did not reflect differences in estimated bear density between habitats within the park, probably because of habitat-related differences in bear diets and in the persistence of their diggings. However, incidence of digging reflected apparent differences in bear density for areas with similar habitats. Incidence of bear diggings outside the park, where local people reported low bear densities, was only 10-20% of that in similar habitat within the park, despite equal densities of termite mounds. No bear diggings were found where local people indicated that bears had been extirpated. Thus, diggings appear to be a means of detecting the presence or absence and relative abundance of sloth bears.