RATIONALE: The coupled analysis of δ13 C and δ15 N stable isotope values of blubber and skin biopsy samples is widely used to study the diet of free-ranging cetaceans. Differences in the lipid content of these tissues can affect isotopic variability because lipids are depleted in 13 C, reducing the bulk tissue 13 C/12 C. This variability in carbon isotope values can either be accounted for by chemically extracting lipids from the tissue, or by using mathematical lipid normalisation models. METHODS: This study examines: (1) the effects of chemical lipid extraction on δ13 C and δ15 N values in blubber and skin of southern hemisphere humpback whales, (2) whether chemical lipid extraction is more favourable than mathematical lipid correction and (3) which of the two tissues is more appropriate for dietary studies. Strategic comparisons were made between chemical lipid extraction and mathematical lipid correction, and between blubber and skin tissue δ13 C and δ15 N values, as well as C:N ratios. Six existing mathematical normalisation models were tested for their efficacy in estimating lipid-free δ13 C for skin. RESULTS: Both δ13 C and δ15 N values of lipid-extracted skin (δ13 C: -25.57 ‰, δ15 N: 6.83 ‰) were significantly higher than bulk skin (δ13 C: -26.97 ‰, δ15 N: 6.15 ‰). Five of the six tested lipid normalisation models had small error terms for predicting lipid-free δ13 C values. The average C:N ratio of lipid-extracted skin was within the lipid-free range reported in other studies while the average C:N ratio of blubber was higher than previously reported. CONCLUSIONS: These results highlight the need to account for lipids when analysing δ13 C and δ15 N values from the same sample. For optimised dietary assessments using parallel isotope analysis from a single sample, we recommend the use of unextracted skin tissue. δ15 N values should be obtained from unextracted skin, whilst δ13 C values may be adequately lipid corrected by a mathematical correction.