Experimental Assessment of Several Population Estimation Techniques on an Introduced Population of Eastern Chipmunks

Abstract

An experimental population of eastern chipmunks (Tamias striatus) composed of 85 individuals of known age and sex was released on a 9.4-ha island previously devoid of chipmunks. Three commonly used mark-recapture techniques (Lincoln-Petersen, Schnabel, and Schumacher-Eschmeyer methods) and the Least Squares Removal Method were used to estimate the size of the population. With the exception of equal catchability, the experiment was designed to satisfy the assumptions of the above methods. Point estimates always underestimated the true population size, and only the confidence intervals of the Lincoln-Petersen Method consistently included the actual population value. Least squares regression analyses suggest that the experimental population is composed of two groups of animals: those easily trapped and those hesitant to enter traps. As such, all population estimation methods estimate the easily captured portion of the population and underestimate the true population size. The mathematically simple Lincoln-Petersen Method was more robust than the other methods; when unequal trappability is suspected, it is more accurate than the other methods (although some precision is sacrificed). Caution is suggested in making inter- or intraspecific statistical comparisons of population size when close adherence to the assumptions of the models utilized is not demonstrable; otherwise, statistical differences may be spurious.