Is lipid correction necessary in the stable isotope analysis of fish tissues?

RATIONALE: Stable isotope analysis (SIA) is a powerful tool for examining diet and food-web dynamics. SIA assumes "you are what you eat" relative to carbon (C) and nitrogen (N). However, fractionation of carbon during lipid synthesis violates this assumption; high-lipid tissues do not reflect δ13C values of diet and therefore have the potential to skew mixing model results and diet interpretations, making corrections necessary. METHODS: Brook Trout (Salvelinus fontinalis) white muscle and liver samples from several fish species representing the temperate North American cold- and warm-water fish community were corrected for lipids via chemical lipid extraction and mathematical lipid normalization. To assess the accuracy of model-predicted lipid-free δ13C values calculated from four normalization models, we compared model-predicted values with those measured after lipid extraction. RESULTS: We found that chemical lipid extraction is unnecessary for Brook Trout white muscle tissue with low initial lipid content. However, in tissues with C:N ratios greater than 3.5, lipid extraction increased δ13C values in fish liver by more than 1.0 ?, indicating that liver lipid content is sufficient to bias δ13C values. We also found that lipids were accurately accounted for with mathematical normalization and recommend that tissues with C:N ratios greater than 3.5 be corrected mathematically. CONCLUSIONS: Our findings indicate that mathematical normalization is sufficient to account for bias in δ13C values associated with lipid content in fish tissues when C:N ratios are above 3.5. C:N ratios below 3.5 indicate that tissues have insufficient levels of lipid to bias the δ13C values. Generally, these findings support the use of more timely and cost-effective processing and analysis methods in future aquatic food-web studies utilizing SIA. [ABSTRACT FROM AUTHOR]