Robust estimation from line transect data

Line transect sampling often provides a practical way to approach estimation of wildlife population density (Seber 1973, Eberhardt 1978). Burnham and Anderson (1976) provided a general framework for the estimation of animal density from line transect data, and many specific analytical methods have been proposed in the literature. Numerous persons are researching methods of statistical analysis for line transect data; however, little consideration has been given to specifying criteria for "good" methods of line transect estimation. We present here some criteria which we believe line transect estimators should satisfy. We do not deal with the case where objects may move from their initial location before being detected.