Résumé
Using stable isotope ratios to explore the trophic ecology of freshwater animals requires knowledge about effects of food quality on isotopic incorporation dynamics. The aim of this experimental study was to: (1) estimate carbon and nitrogen isotopic incorporation rates and trophic discrimination factors (TDFs) of a freshwater first‐feeding fish (i.e. salmonid fry) fed three diets that differed only in protein quality (animal or plant or a blend of both); (2) investigate effects of fasting and; (3) evaluate the proportion of each source assimilated when fry were fed a 50:50 animal:plant‐based diet. For each diet, incorporation rates of δ13C and δ15N values were estimated using a time or growth‐dependent isotopic incorporation model. Effects of fasting on isotope ratio values were measured regularly until the death of fry. Bayesian stable‐isotope mixing models were used to estimate the contribution of animal and plant material to fish fed a blend of both food types. Our results show that incorporation rates were lower for fry fed a plant‐based diet than for those fed an animal‐based diet as growth rate decreased. Time‐ and growth‐dependent models indicated that growth was solely responsible for isotopic incorporation in fry fed an animal‐based diet, whereas catabolism increased in fry fed a plant‐based diet. After lipid extraction, carbon TDFs were similar regardless of the diet, whereas nitrogen TDFs increased for fry fed a plant‐based diet. Long‐term fasting induced an increase of 0.63‰ in δ13C values of fry in 23 days, whereas δ15N values did not vary significantly. Proportions of food sources assimilated by fry fed an animal:plant‐based diet were similar to those consumed when using a mixing model with the estimated TDFs, while proportions were unrealistic when using mean TDFs extrapolated from the literature. The results of our study indicate that the quality of food must be considered to use an appropriate timescale to detect changes in fry diets in the field. Moreover, we recommend using different carbon and nitrogen TDFs, one for animal‐derived sources and one for plant‐derived sources, to increase the accuracy of mixing models.