Carbon, nitrogen and oxygen isotope fractionation during food cooking: Implications for the interpretation of the fossil human record

Objectives: Stable isotope data provide insight into the reconstruction of ancient human diet. However, cooking may alter the original stable isotope compositions of food due to losses and modifications of biochemical and water components. Methods: To address this issue, carbon, nitrogen and oxygen isotope ratios were measured on meat aliquots sampled from various animals such as pork, beef, duck and chicken, and also from the flesh of fishes such as salmon, European seabass, European pilchard, sole, gilt-head bream, and tuna. For each specimen, three pieces were cooked according to the three most commonly-known cooking practices: boiling, frying and roasting on a barbecue. Results: Our data show that cooking produced isotopic shifts up to 1.8 parts per thousand, 3.5 parts per thousand, and 5.2 parts per thousand for delta C-13, delta N-15, and delta O-18 values, respectively. Such variations between raw and cooked food are much greater than previously estimated in the literature; they are more sensitive to the type of food rather than to the cooking process itself, except in the case of boiling. Conclusions: Reconstructions of paleodietary may thus suffer slight bias in cases of populations with undiversified diets that are restrained toward a specific raw or cooked product, or using a specific cooking mode. In cases of oxygen isotope compositions from skeletal remains (bones, teeth), they not only constitute a valuable proxy for reconstructing past climatic conditions, but they could also be used to improve our knowledge of past human diet.