Nitrogen stable isotope natural abundance data are often used in trophodynamic research. The assumed nitrogen diet-tissue fractionation (Δδ 15 N) determines conclusions about trophic level, potential food sources and ontogenetic diet shifts. Δδ 15 N is usually assumed to be 3.0–3.4‰ per trophic level and unaffected by the size or age of animals or their environment. To assess the effects of body size, experimental duration and environmental conditions on fish tissue Δδ 15 N, two populations of European sea bass (Dicentrarchus labrax) were reared on constant diets of dab (Limanda limanda) muscle or sandeel (Ammodytes marinus) for 2 years under natural light and temperature regimes. Bass were sampled at approximately monthly intervals to determine Δδ 15 N for muscle, heart and liver tissue. Mean values of Δδ 15 N were 3.83‰, 3.54‰, 2.05‰ (sandeel diet) and 3.98‰, 3.32‰, 1.95‰ (dab diet) for muscle, heart and liver tissue respectively. The assumption that fractionation was independent of body mass was upheld for muscle and heart tissue, but not for liver. Time effects on muscle Δδ 15 N were explainable by a sinusoidal function with a period of 1 year and wave height ∼0.3‰. Time resulted in increases in heart δ 15 N and decreases in liver δ 15 N which were small compared to background variation, equating to 1/6 of a trophic level over 2 years, and unlikely to have great significance in ecological studies. Heart and liver δ 15 N were also affected by temperature probably reflecting the metabolic functions of these tissues and their associated rates of turnover. However in heart the explanatory power of temperature appeared tied to that of time. Although the Δδ 15 N for bass muscle on both diets approached 4‰, the Δδ 15 N values from this study, when combined with those from the literature, suggest that where fish species specific data are not available, a mean Δδ 15 N for fish muscle of 3.2‰ should be applied (mean white muscle Δδ 15 N=3.15). The literature based mean Δδ 15 N for whole fish was lower than that of white muscle suggesting that a separate Δδ 15 N (2.9‰) should be applied when sampling whole fish.