Print PageNatural populations consist of phenotypically diverse individuals, due to genetic and/or environmental variation. However, ecological models typically disregard trait variation within populations, focusing instead on the dynamics of populations and species as the unit of interest. Here, we argue that this simplification is unwarranted, and that trait variation can qualitatively alter population and community dynamics and therefore must be systematically incorporated into ecology. The community effects of trait variation can be studied by contrasting species' equilibrium densities and transient dynamics in ecological models which include or exclude intraspecific variation. We synthesize the available theory to distill six general mechanisms by which trait variation can alter community dynamics. These six mechanisms will help us answer the question: When and why will phenotypic variation within populations alter community dynamics?
Results/Conclusions: First, trait variance can alter species' mean demographic parameters when those parameters are non-linear functions of variable traits (Jensen's Inequality). Second, trait variance can alter species niche width and hence food web architecture. Third, fluctuations in population density can be stabilized if different phenotypes within the population experience negatively correlated changes in density (portfolio effect). These first three mechanisms can operate whether or not the trait variation is heritable, whereas the final three mechanisms typically require genetic variance. The fourth process entails the continual recreation of phenotypes via recombination or Mendelian segregation (which we term phenotypic subsidy). This subsidy can prevent the extinction of ecologically unsustainable phenotypes, or drive otherwise fit phenotypes to extinction. As a result, simple Mendelian genetics can cause deviations from the ecologically determined equilibrium densities each phenotype would achieve in isolation. Fifth, heritable trait variance is necessary for eco-evolutionary feedbacks in which interspecific interactions drives natural selection on phenotypes that affect the strength or nature of the interactions. Finally, stochastic sampling effects such as genetic drift can create maladaptive eco-evolutionary feedbacks.
Excluding phenotypic variation from ecological models presumes that all of the above six processes are inapplicable or have weak effects. In fact, we suggest that each mechanism is quite general, and can have large effects on community dynamics which we illustrate with reference to specific models. Consequently, we suggest that phenotypic variation within populations may have ubiquitous and potentially profound ecological effects. We conclude that ecology will benefit by integrating both environmental and genetic trait variation into standard models of interspecific interactions.
