Abstract: Multispecies mutualisms abound in nature, but theory struggles to predict coexistence of multiple partners. Generally, eco-evolutionary theory argues that mutualisms are stabilized by partner specificity, in which high-quality partners are rewarded and low-quality partners are sanctioned. However, many extant mutualisms show high diversity as well as variation in partner quality. Here, we use a mathematical model of the speciose tree-ectomycorrhizal mutualism to explore one possible mechanism for the maintenance of partner diversity. In this mutualism, fungal partners vary in their functional traits, such that an individual fungal partner’s quality may depend upon environmental conditions. Because tree-ectomycorrhizal mutualisms are long-lived, they must endure variable environmental conditions including seasonality, decadal variation, and rare events. Thus, a tree’s optimal fungal partner may vary over time. Therefore, we hypothesize that a tree’s optimal (growth-maximizing) strategy may be a “bet-hedging” one, in which the tree invests in a portfolio of fungi, maintaining a diversity of fungal partners that leaves it better prepared for unpredictable environmental variation. We tested this prediction using a system of ordinary differential equations that describe how a host tree allocates resources to its fungal partners. We used the model to simulate a gradient of investment strategies, from “loyal” strategies that reward only the highest quality partner, to “leaky” ones in which investments spill out to other members of the fungal community. We found that a tree with a diversified (“leaky”) investment strategy has up to 30% higher fitness in variable environmental conditions. However, leaks are only optimal when environmental variation is slow relative to fungal growth dynamics, or when fungal partners have highly divergent environmental responses. For example, when the optimality of two fungal partners switches depending upon environmental conditions, the tree should leak as much as half of its carbon investments to the currently low quality partner in order to be best prepared for future environmental conditions. Our results lead to testable predictions about (1) the maintenance of apparently “low-quality” partners in multispecies mutualisms, and (2) the relationship between diversity and temporal variation in natural systems.
