The stable isotope approach to dietary reconstructions is based on the consistent isotopic fractionation between diet and consumer. Despite the rapidly increasing use of stable isotopes in studies on bivalves, classical fractionation values (0–1‰ for δ13C and 3–4‰ for δ15N) are used in the literature irrespective of the tissue analyzed and of the sample preparation method. We measured δ13C and δ15N for lipid-removed and lipid-containing tissues (adductor muscle, foot, mantle lobe, siphon, labial palp, gill lamella, midgut gland, and male and female gonad) of three suspension feeding bivalves (Crassostrea gigas, Cyclina sinensis and Ruditapes philippinarum) to provide an essential calibration for the diet-tissue fractionation. There were similar tissue-specific patterns in the δ13C and δ15N, which rank adductor muscle>other tissues>midgut gland or female gonad. Removal of lipids raised the δ13C and δ15N for all tissues of the three species. Acid treatment of the total soft tissues of R. philippinarum had no significant effects on the isotopic compositions. Diet-tissue fractionations for R. philippinarum were estimated based on the known fractionation values for the lipid-containing total soft tissues. Lipid removal and associated rinsing raised not only the 13C fractionation by 0.4–1.5‰ but also the 15N fractionation by 0–0.8‰, resulting in large fractionation values outside of the range of currently accepted 13C and 15N fractionation. The present study demonstrates that isotopic analyses are affected by the choice of tissue type and sample preparation method.
2006 The Plankton Society of Japan, The Japanese Association of Benthology