Compared to oxygen isotopes, the carbon isotope composition of biogenic carbonates is less commonly used as proxy for palaeoenvironmental reconstructions because shell δ13C is derived from both dissolved inorganic (seawater) and organic carbon sources (food), and interactions between these two pools make it difficult to unambiguously identify any independent effect of either. The main purpose of this study was to demonstrate any direct impact of variable food supply on bivalve shell δ13C signatures, using low/high rations of a 13C-light mixed algal diet fed to 14-month-old (adult) cultured Japanese Crassostrea gigas under otherwise essentially identical in vitro conditions during 3 summer months (May, June and July 2003, seawater temperature means at 16, 18 and 20 °C respectively) in experimental tanks at the Argenton laboratory along the Brittany Atlantic coast of France. At a daily ration of 12% (versus 4%) oyster dry weight, the newly grown part of the shells (hinge region) showed significantly lower δ13C values, by 3.5‰ (high ration: mean of −5.8 ± 1.1‰, n = 10; low ration: mean of −2.3 ± 0.7‰, n = 6; ANOVA Scheffe’s test, p < 0.0001). This can be explained by an enhanced metabolic activity at higher food supply, raising 13C-depleted respiratory CO2 in the extrapallial cavity. Based on these δ13C values and data extracted from the literature, and assuming no carbon isotope fractionation between food and shell, the proportion of shell metabolic carbon would be 26 ± 7 and 5 ± 5% for the high- and low-ration C. gigas shells respectively; with carbon isotope fractionation (arguably more realistic), the corresponding values would be 69 ± 14 and 24 ± 9%. Both groups of cultured shells exhibited lower δ13C values than did wild oysters from Marennes-Ol eron Bay in the study region, which is not inconsistent with an independent influence of diet type. Although there was no significant difference between the two food regimes in terms of δ18O shell values (means of 0.1 ± 0.3 and 0.4 ± 0.2‰ at high and low rations respectively, non-significant Scheffe’s test), a positive δ13C vs. δ18O relationship recorded at high rations supports the interpretation of a progressive temperature-mediated rise in metabolic activity fuelled by higher food supply (in this case reflecting increased energy investment in reproduction), in terms not only of δ13C (metabolic signal) but also of δ18O (seawater temperature signal). Overall, whole-shell δ18O trends faithfully recorded summer/winter variations in seawater temperature experienced by the 17-month-old cultured oysters.