The effects of formalin and ethanol preservation on the δ13C and δ15N isotope signatures of Arctic charr Salvelinus alpinus muscle tissue were examined. The lipid content of the tissue samples studied ranged from 3·6 to 6·1% and was not correlated with the magnitude of observed isotopic shifts in preserved samples. Ethanol and formalin significantly depleted and enriched, respectively, the δ13C isotope signatures of preserved tissues when compared to control samples. Ethanol did not significantly enrich δ15N signatures in comparison to controls, whereas formalin did. A meta-analysis of multiple species effects further demonstrated significant preservation effects in fish tissue. Statistical analysis of data obtained by correcting preserved tissue isotope signatures with literature, bootstrapped or meta-analysis derived correction factors demonstrated significant differences between corrected and control sample isotope signatures or failure to produce a unity slope when the data sets were regressed against one another. Species-specific, bootstrapped linear correction models resulted in no such errors. Results suggest that species-specific correction methods should be used for fishes because of the known wide variation in fish tissue lipid content and composition. Accordingly, the use of pilot studies will be required to develop correction factors that properly adjust for preservation effects when interpreting temporal patterns in historic analyses of food webs.