Interpretation of Relative Weight in Three Populations of Wild Bluegills: A Cautionary Tale

Abstract

Physiological status of wild fish is inherently variable in time and space, and investigators need to understand how this variation affects the interpretation of condition indices. Our objective was to describe the relationship of relative weight (W_r) to several indicators of nutritional status (lipid, protein, and water content, and weights of internal organs) in two populations of bluegill Lepomis macrochirus and see how well this relationship performed when tested against an independent data set, a third population. We used four samples from the third population to evaluate the relationship of W_r to nutritional status through the annual cycle. The regression model relating W_r to the component variables was highly significant (R² = 0.573, F = 14.97, P < 0.001). The most influential variables in the model were liver–-somatic index and viscerosomatic index. In the testing data, mean bias was significantly greater than zero in three of four cases, and unexplained variance was much greater than in the model-building data in two of four cases. Most prediction errors were noticeably positive and tended to be greater at lower predicted W_r. The model-testing exercise showed that the relationships of condition indices to nutritional variables in wild fish are imprecise and qualitative. Bluegill W_r should be interpreted cautiously because of its imprecise relationship to compositional variables and the many potentially confounding factors. In this regard, we offer several recommendations for the use and interpretation of W_r in wild fishes.