Field Tests of Strip Census Methods

Twenty strip censuses were conducted in areas containing known populations of inanimate objects or, in two instances, of live animals. Each census was based on sighting or perpendicular dis- tances from straight-line transects. These data were then converted to population estimates according to 10 different strip census methods, and the results were compared with the known populations. Of the 10 methods tested, only three provided estimates sufficiently close to the actual for serious con- sideration. The two most promising methods were the belt transects of Kelker (1945) and Anderson and Pospahala (1970), both of which require a frequency distribution of perpendicular distances of the sighted objects from the census lines. The third method was King's (Leopold 1933), which utilizes the mean of sighting or flushing distances to establish the width of the census strip. This method ap- pears subject to some biases but less so than others based on sighting distances.