Understanding how temperature affects adaptation of cell size is challenging because cell size mediates numerous physiological and ecological trade-offs. Previous work has identified physiological mechanisms that lead to decreases in cell size with warming (the temperature-size rule; TSR). However, it is unclear how ecological processes (e.g., competition, predation) combine to modify the TSR. Here, we evaluate how ecological interactions affect thermal adaptation of phytoplankton cell size. We perform an eco-evolutionary analysis of a nutrient-phytoplankton-zooplankton model. The model assumes phytoplankton experience size-dependent constraints on resource allocation that cause small cells to sacrifice investment in growth machinery, thereby reducing maximum growth rate but increasing competitive ability. We find that trophic interactions strongly impact the evolutionarily stable cell size across temperatures. Without zooplankton, cell size declines monotonically with temperature, consistent with the TSR. With zooplankton, cell size varies unimodally with temperature, due to temperature-dependent shifts in the grazer's capacity to ease nutrient competition by controlling phytoplankton biomass. Size-selective grazing does not qualitatively alter this result but can facilitate coexistence via a competition-predation trade-off. Trophic interactions therefore can produce temperature-size responses that differ qualitatively from the canonical TSR, and an understanding of how temperature affects cell size is incomplete without this ecological component.

The growth of populations and organisms often depends on their previous history of environmental exposure: a phenomenon referred to as "phenotypic memory." The field of ecology presently lacks a mechanistic theory describing phenotypic memory and, as such, evaluating the ecological consequences of this phenomenon is a major challenge. Here, we show that internal nutrient storage connects past thermal experience to current growth in phytoplankton. We develop a new mechanistic model showing how time lags in the incorporation of stored nutrients into new biomass produces phenotypic memory. By testing this model against experimental data of phytoplankton population growth rates following temperature perturbations, we find general patterns in the population consequences of phenotypic memory: prior exposure to warm temperatures depletes nutrient stores, and, in doing so, slows growth during subsequent temperature exposure and restricts the range of acute temperature exposures yielding a positive growth rate. Our model reveals how phenotypic memory produces temporal variation in critical thermal minima and maxima and predicts that the thermal niche is constricted by long-term exposure to warm temperatures (e.g., during summer months), but that high frequency temperature fluctuations can expand a population's thermal niche. This work provides a general, mechanistic basis for considering the ecological implications of phenotypic memory.