nimal ecologists sample the populations they study in a variety of ways. Those who study bird populations in Hawaiian forests may count the number of birds seen and/or heard while walking a transect line or standing at a fixed point in the forest. Those who study red and grey kangaroos in the arid rangelands of central Australia may count animals from a fixed-wing airplane flying along transect lines. The fisheries biologist in the northern Pacific may count the number of halibut caught in a commercial fisheries operation, and the mammalogist in oldfield habitat in eastern North America may count the number of meadow voles caught overnight in a grid of small mammal traps. All these ecologists share a common problem. They have obtained a count statistic (animals seen, heard, or caught), but in most cases it is likely that this statistic is smaller than the actual number of animals present in the sampled area. That is, they have sampled some unknown fraction of the population of interest. The count statistic by itself is thus of little use to the ecologist, except to provide a minimum bound on population size. We need to know some-