In unpredictable environments, one way to increase fitness is to produce offspring with variable dormancy times. Under this strategy, some offspring emerge the year they were born, while others emerge in a future year. In this way, a female hedges her bets against complete reproductive failure. For bet-hedging to be advantageous, geometric mean fitness increases while arithmetic mean fitness decreases, due to a decrease in variance. Bet-hedging has received thorough theoretical treatment, with only a few empirical studies of bet-hedging in taxa such as desert annual plants, weevils, and crustaceans. Here, we ask whether variable dormancy in Colletes validus, a ground-nesting solitary bee, functions as bet-hedging in variable environments. Using structured matrix models parameterized with field-collected data, we tested theoretical predictions that variable dormancy is advantageous as 1) uncorrelated temporal variation increases and 2) when there is a lower cost to remaining dormant. Specifically, we ran stochastic simulations across a range of plausible temporal variances and survival costs to calculate arithmetic mean fitness, variance in fitness, and geometric mean fitness for each dormancy rate between 0 and 1.
Results/Conclusions
We documented, for the first time, that a ground-nesting bee produces offspring with variable dormancy; dormancy rates ranged from 3–44% at two sites in New England in 2019. For some, but not all, environmental conditions, variable dormancy in C. validus met the criteria for bet-hedging: as more two-year offspring were produced, arithmetic mean fitness decreased, while geometric mean fitness increased. For temporal variance greater than 0.95, assuming no cost to dormancy, an intermediate dormancy rate was advantageous. This minimum variance was slightly higher than the median variance in growth rates calculated for 52 insect species (0.69), but well within the observed range (95% limits: 0.14–5.14). As the cost to dormancy increased, both the minimum temporal variance required for bet-hedging increased and the optimal dormancy rate decreased across all environmental conditions. Taken together, our findings suggest that the adaptive value of variable dormancy is likely sensitive to local conditions that affect both overwinter survival and fecundity. Moreover, our work demonstrates the value of natural history studies for detecting cryptic life histories of bees; that solitary bees may be buffered against future environmental change; and that classic predictions regarding optimal life histories in variable environments hold for insect pollinators.