Summary Food chain efficiency (FCE), the proportion of primary production converted to production of the top trophic level, can influence several ecosystem services as well as the biodiversity and productivity of each trophic level. Aquatic FCE is affected by light and nutrient supply, largely via effects on primary producer stoichiometry that propagate to herbivores and then carnivores. Here, we test the hypothesis that the identity of the top carnivore mediates FCE responses to changes in light and nutrient supply. We conducted a large‐scale, 6‐week mesocosm experiment in which we manipulated light and nutrient (nitrogen and phosphorus) supply and the identity of the carnivore in a 2 × 2 × 2 factorial design. We quantified the response of FCE and the biomass and productivity of each trophic level (phytoplankton, zooplankton, and carnivore). We used an invertebrate, Chaoborus americanus, and a vertebrate, bluegill sunfish (Lepomis macrochirus), as the two carnivores in this study because of the large difference in phosphorus requirements between these taxa. We predicted that bluegill would be more likely to experience P‐limitation due to higher P requirements, and hence that FCE would be lower in the bluegill treatments than in the Chaoborus treatments. We also expected the interactive effect of light and nutrients to be stronger in the bluegill treatments. Within a carnivore treatment, we predicted highest FCE under low light and high nutrient supply, as these conditions would produce high‐quality (low C:nutrient) algal resources. In contrast, if food quantity had a stronger effect on carnivore production than food quality, carnivore production would increase proportionally with primary production, thus FCE would be similar across light and nutrient treatments. Carnivore identity mediated the effects of light and nutrients on FCE, and as predicted FCE was higher in food chains with Chaoborus than with bluegill. Also as predicted, FCE in Chaoborus treatments was higher under low light. However, FCE in bluegill treatments was higher at high light supply, opposite to our predictions. In addition, bluegill production increased proportionally with primary production, while Chaoborus production was not correlated with primary production, suggesting that bluegill responded more strongly to food quantity than to food quality. These carnivore taxa differ in traits other than body stoichiometry, for example, feeding selectivity, which may have contributed to the observed differences in FCE between carnivores. Comparison of our results with those from previous experiments showed that FCE responds similarly to light and nutrients in food chains with Chaoborus and larval fish (gizzard shad: Clupeidae), but very differently in food chains with bluegill. These findings warrant further investigation into the mechanisms related to carnivore identity (e.g., developmental stage, feeding selectivity) underlying these responses, and highlight the importance of considering both top‐down and bottom‐up effects when evaluating food chain responses to changing light and nutrient conditions.