Distinct impacts of food restriction and warming on life history traits affect population fitness in vertebrate ectotherms

The reduction of body size with warming has been proposed as the third universal response to global warming, besides geographical and phenological shifts. Observed body size shifts in ectotherms are mostly attributed to the temperature size rule (TSR) stating that warming speeds up initial growth rate but leads to smaller size when food availability does not limit growth. Nevertheless, climate warming can decrease food availability by modifying biochemical cycles and primary production. Food availability can also influence growth, fecundity and survival and thus potentially modulate the effect of temperature on life history strategies and fitness. However, the interactive effects of temperature and food availability on life history traits have been mostly studied in small invertebrate species where life history traits have been mainly considered in isolation. In contrast, we have limited information on (1) how temperature and food availability jointly influence life history traits in vertebrate predators and (2) how changes in different life history traits combines to influence fitness and population growth. To fill this gap, we investigated under laboratory conditions the independent and interactive effects of temperature (20 or 30 °C) and food availability (restricted or ad libitum) on the growth, fecundity and survival of the medaka fish Oryzias latipes. We next used our empirical estimates of vital rates as input parameters of an Integral Projection model (IPM) to predict how modifications in vital rates translate into generation time and population growth rate (i.e. mean fitness). Our results confirm that warming leads to a higher initial growth rate and lower size leading to crossed growth curves between the two temperatures. Food-restricted fish were smaller than ad libitum fed fish throughout the experiment, leading to nested growth curves. Fish reared at 30 °C matured younger, had smaller size at maturity, had a higher fecundity but had a shorter life span than fish reared at 20 °C. Food restriction increased survival probabilities under both temperature conditions corresponding to a "eat little die old" strategy. According to the IPM, warming reduces generation time and increases mean fitness in comparison to the cold treatments. Food restriction increased generation time and fitness in the cold treatment but had no effect in the warm treatment. Our results highlight the importance of accounting for the interaction between temperature and food availability to understand how body size shifts can affects vital rates and population demography. This is of importance in the context of global warming as resources (e.g., phytoplankton and zooplankton communities in aquatic ecosystems) are predicted to change in size structure and total abundance with increasing temperatures. Interestingly, our results suggest that food restriction has a weaker effect on fish mean fitness under warming