The applicability of capture-recapture models for estimating tiger Panthera tigris numbers from camera-trap data was investigated in a part of Nagarahole National Park, India during 12 months (387 trap-nights) in 1991-92. Camera-traps were placed along regular travel routes of tigers to obtain 31 photographic 'captures' of individual tigers in a 15 km2 study area, during nine sampling occasions. Tigers could be identified unambiguously from photographs, and capture histories of 10 different animals were obtained. The data were analysed in the conceptual framework of capture-recapture theory, using probabilistic models. The results suggest that the closed capture-recapture model M(h) which allows for heterogeneity of capture probabilities among individual animals was appropriate for estimating tiger population size. The mean tiger numbers and their 95% confidence intervals were estimated at 11 (10-22) and 10 (10-15), respectively, with the M(h) (Jackknife) and M(h) (Chao) estimators. Considering the prey biomass available in the study area, the mean tiger densities of 13.3-14.7 adult/subadult tigers per 100 km2 estimated from the above data appear to be reasonable. Capture-recapture models using comera-trap data offer scope for estimating objectively parameters such as size, density, survival, and recruitment for populations of tigers and other secretive animal species.
