Optimal reproductive allocation in annuals and an informational constraint on plasticity.

In this computational study, we examined optimal reproductive allocation schedules in annual plants whose season lengths vary in predictability. We discuss relationships among season-length predictability, the form of the optimal allocation schedule, the degree of plasticity reflected in the optimal reaction norm, and the competitive consequences of plasticity and bet-hedging. We used an evolutionary algorithm to search the allocation-schedule space for optima, given different distributions of season length. The resulting schedules maximize geometric-mean fecundity under their selecting distributions. We then examined the relative fitness of these schedules in simulated competition among reaction norms optimized for different degrees of season-length predictability. Gradedness of optimal schedules decreases with increasing season-length predictability, and reaction norms comprising highly graded schedules reflect lesser plasticity than norms comprising schedules that are less graded. In simulations, competitively successful genotypes were those that reflected plasticity appropriate to the season-length predictability. Informational constraints in the form of low season-length predictability select for low plasticity and high bet-hedging in allocation. Because an environmental cue must mediate the relationship between environment and fitness, plasticity in reproductive allocation ought to be understood not as a direct response to the selective environment, but rather to cues that are correlated with relevant environmental parameters.