comparison of methods for quantitative analysis of feeding selection of fishes

Predator-prey data were obtained in a study of feeding ecology of land-locked alewife, Alosa pseudoharengus , in Claytor Lake, Virginia, and subjected to two indices of electivity as well as a nonparametric paired-comparisons statistical test. Values obtained from the two indices were often contradictory with each other and the statistical test. Attributes of the statistical procedure render it more appropriate than electivity indices to describe feeding selectivity in fish populations. Results of the study also demonstrated the importance of considering the size composition of zooplankton populations in prey preference studies; measures of selection were often positive for zooplankters greater than or equal to 1.0 mm while negative for those of the same prey group < 1.0 mm length. Because numerous problems are inherent with in situ feeding selectivity studies of planktivorous fishes, selectivity determinations, regardless of methodology, should be construed in a relative rather than absolute context.