Classic theory suggests that environmental variability is bad for population growth. All things equal, populations are expected to benefit from increases in mean population growth rate and from reductions in the variance of growth rates. Context dependent benefits conferred by microbial symbionts often provide benefits to their hosts under harsh abiotic conditions while being neutral or even slightly costly in benign environments. This has posed challenges for predicting the average outcome of a given mutualism, but context dependence may serve to limit variability by reducing the effects of both bad and good years through time. Using 13 years of demographic data from long-term experimental plots of seven species of fungal endophyte-associated grasses (Agrostis perennans, Elymus villosus, Elymus virginicus, Festuca subverticillata, Lolium arundinaceum, Poa alsodes, and Poa sylvestris) in Brown Co. Indiana, we constructed size-structured, stochastic matrix models to quantify the relative contributions of effects on mean fitness and of variance buffering on stochastic population growth rates.
Results/Conclusions
Across species, we find an overall benefit for endophyte infected plants. Variance buffering by the symbiont is a common element of this benefit, but it generally contributes less than effects on mean fitness. Variance buffering and mean effects were particularly common for survival and growth but were less important for other vital rates. Variance buffering is an understudied mechanism within species interactions that we expect may gain increasing importance for populations experiencing warmer temperatures and increased climatic variability as expected under climate change.