We evaluated a hypothesis derived from the metabolic theory of ecology (MTE) that the ratio of microzooplankton herbivory (m) to phytoplankton growth (μ) will arise in a warming ocean because of the different temperature dependencies of autotrophic and heterotrophic organisms. Using community‐level growth and grazing data from dilution experiments, generalized additive models (GAMs) were constructed to describe the effects of temperature and chlorophyll on m : μ. At low chlorophyll levels, m : μ decreases with increasing temperature, whereas at high chlorophyll levels, m : μ increases initially with temperature before reaching a peak and then declines. These complex responses of m : μ result from mixed effects of temperature and chlorophyll on microzooplankton biomass (B_z), biomass‐specific microzooplankton grazing rate (m : B_z), and phytoplankton growth rate (μ). B_z decreases with rising temperature and increases with rising chlorophyll. m : B_z increases with temperature and decreases with chlorophyll. Nutrient‐enriched growth rate of phytoplankton (μ_n) and μ increase with increasing temperature and chlorophyll. Holding chlorophyll constant, the calculated activation energies of m : B_z and μ_n are 0.67 ± 0.05 and 0.36 ± 0.05 eV, respectively, both consistent with previous MTE estimates for heterotrophs and autotrophs. Our study indicates that warming may enhance phytoplankton losses to microzooplankton herbivory in eutrophic but not in oligotrophic waters. The GAM analysis also provides important insights into underlying system relationships and reasons why community‐level responses in natural systems may depart from theory based on laboratory data and individual species.

We compile two data sets from ¹⁴C uptake and dilution experiments conducted in surface waters of the global ocean to investigate the relationship between phytoplankton mass‐specific growth rate and cell size. After temperature correction, both data sets suggest that this relationship might be described by a unimodal quadratic curve with the modal size (the size corresponding to the maximal growth rate) being 2.8 and 5.4 µm in the ¹⁴C and dilution data sets, respectively. Nutrient enrichment does not change the qualitative nature of the relationships, and we conclude that inherently low maximal growth rates of picophytoplankton, not ambient nutrient effects, play the major role in determining the positive relationships over the size range where phytoplankton size is below the modal size. Temperature‐corrected phytoplankton grazing mortality rate is positively correlated with phytoplankton average size, but the proportion of daily primary production consumed by microzooplankton is negatively correlated with cell size, suggesting a reduced grazing effect as size increases. The unimodal relationship between phytoplankton growth rate and cell size is consistent with theoretical considerations and might reflect an adaptive response of phytoplankton to varying extents of nutrient limitation and grazing effect in marine systems.