We determined whether lipid extraction of seabird muscle, liver and feathers is required for δ13C and δ15N analysis, based on a study of Barau's Petrels (Pterodroma baraui) and Whitetailed Tropicbirds (Phaethon lepturus). Samples were analyzed for stable isotopes before and after lipid removal. Lipidfree muscle and liver were significantly enriched in 13C compared to those containing lipids (0.52 ‰ and 0.61 ‰ mean differences in δ13C respectively) implying that lipids should be extracted from these tissues to avoid effects of uncontrolled differential lipid content. However lipid extraction also slightly increased muscle δ15N values in tropicbirds. Researchers should thus run samples separately for δ13C and δ15N analysis, extracting lipids only in aliquots in which δ13C is measured, or apply arithmetic δ13C normalization methods on non lipidextracted samples. Predictive models developed from our data, inferring δ13C values of lipidextracted material from δ13C values and C:N ratios of non lipidextracted material are presented and compared to five normalization methods taken from literature. Our models showed the best prediction efficiencies followed by that developed by McConnaughey and McRoy (1979). In feathers, no difference in stable isotopic measurements were noted between aliquots treated with sodium hydroxide or 2:1 chloroform:methanol. Both protocols seem suitable for preparing feather samples for isotopic analysis.