Concerns about the effects of formalin fixation have largely precluded the use of fluid-preserved museum specimens in stable isotope studies. We sought to determine the magnitude, direction, and causes of formalin-induced change in isotopic values of carbon, nitrogen, and sulfur by examining differences in isotopic signatures between paired fresh-frozen fish tissue samples and tissues subjected to fixation and preservation procedures employed in natural history museums. Several factors were investigated as predictors of formalin-induced change: (1) duration of formalin fixation and ethanol preservation; (2) lipid extraction by formalin and ethanol; and (3) differences in isotopic content among formalin preparations from different manufacturers. Formalin fixation caused a decrease in the δ¹³C-values across samples and treatments (mean decrease in preserved tissues of 1.1‰ for all treatments, SD = 0.8). The magnitude of depletion depended on the isotopic composition of formalin used for fixation, but not duration of fixation or extraction of lipids. δ¹⁵N-values were enriched for all samples and magnitude was independent of brand of formalin and all other factors studied (mean increase of 0.5‰ for all treatments, SD = 0.3). δ³⁴S-values were not systematically affected by formalin under any conditions of the tests. Formalin-induced changes in carbon, nitrogen, and sulfur isotope values of fish tissue were small compared to changes expected from natural fractionation processes that are of interest in ecosystem studies. Thus, specimens held by natural history museums offer a potentially important source of predisturbance material for quantifying historical change in nutrient cycles and food webs in aquatic ecosystems.