Variation in relationships of δ13C and δ15N between lethal and nonlethal samples in fishes

ABSTRACT Objective We summarized variation in muscle–fin isotope relationships in studies of freshwater fishes and evaluated whether season and body condition explained variation in muscle–fin relationships in a field study. Methods We reviewed published relationships of δ13C and δ15N values from muscle and fin tissue samples in fishes and compared them to relationships in a large sample of muscle and fin tissue samples that we collected from two small-bodied fishes, juvenile Brown Trout Salmo trutta and Mottled Sculpin Cottus bairdii. We evaluated the importance of season and body condition (dry matter content and C:N) in our muscle–fin relationships using multiple regression with model selection. Results Correlations between muscle and fin δ13C and 15N values within studies were high, but the variance in relationships across studies was large, even for the same species. Muscle and fin δ13C and δ15N values in our field study were also strongly positively correlated (r ≥ 0.89), with minor muscle–fin differences for Brown Trout (Δδ13C = −0.71 ± 0.42‰, Δδ15N = 0.13 ± 0.38‰) and Mottled Sculpin (Δδ13C = −0.73 ± 0.29‰, Δδ15N = −0.21 ± 0.31‰). Slopes of our muscle–fin relationships for δ13C (β1 = 0.796‒0.911) and δ15N (β1 = 0.826‒0.872) were similar to average literature values (δ13C: β¯1 = 0.824; δ15N: β¯1 = 0.875), but again, the variance in published relationships was large. Incorporating season and body condition in models improved muscle–fin relationships. Conclusions High variance in muscle–fin isotopic relationships makes it difficult to rely on previously established models. We found that body condition improved the fit of muscle–fin relationships, suggesting that the large variance among the previous studies might be due to unmeasured predictors. Incorporating additional predictors, such as energy status, could reduce variation in muscle–fin relationships and increase their applicability across systems.