Phenotypic integration limits the variation in plant phenotypic plasticity among traits: A meta‐analysis

Abstract Phenotypic plasticity, a key mechanism by which plants respond to environmental change, tends to be lower under stressful conditions. Phenotypic integration refers to the degree of trait correlation and is thought to limit phenotypic plasticity and to be higher under stressful conditions, thus potentially explaining the limits to plasticity. Yet, the evidence of whether integrated traits are less plastic is largely mixed. Alternative evidence suggests that phenotypic integration might limit the variation rather than the strength in plastic responses, resulting in correlated plastic responses. We performed a phylogenetically controlled meta‐analysis to evaluate whether phenotypic integration limits the magnitude or the variation in plastic responses under stressful conditions. The meta‐analysis included data from 156 study cases obtained from 41 experimental studies, encompassing 83 different plant species. We also tested whether the association between phenotypic integration and plasticity depended on the level of phenotypic integration in the population, experimental treatment, functional group, or ontogeny. Lastly, to understand the conditions under which limits to plasticity may occur, we evaluated whether phenotypic integration increased in response to biotic or abiotic stress. Overall, more strongly integrated traits were not found to be less plastic. Instead, we found greater phenotypic integration to be associated with lower variation in plastic responses, that is resulting in a greater correlation among plastic responses across traits. The association between plastic responses and phenotypic integration was independent of experimental treatment, plant functional group and ontogeny. Phenotypic integration did not consistently increase with biotic or abiotic stress. Our study reveals a different type of limit imposed on plastic responses by phenotypic integration, reducing the variation rather than the magnitude of plastic responses, which is consistent with previous evidence indicating correlated plasticity among correlated traits. The ecological implications of these limits must be addressed: whether they constrain plant functional responses to the environment or are rather adaptive. Our results could help reconcile the seemingly contradictory evidence in the literature, showing both positive and negative associations between trait‐level phenotypic integration and trait‐level phenotypic plasticity; we suggest that it depends on whether correlated traits are all highly or barely plastic, respectively. Read the free Plain Language Summary for this article on the Journal blog.