Selection on offspring size among environments: the roles of environmental quality and variability

P>1. How mothers balance the trade-off between offspring size and number to maximize maternal fitness has long been of interest to ecologists seeking to understand the evolution of offspring size. Predictions of the optimal offspring size depend fundamentally on the relationship between offspring size and offspring performance, which may in turn vary with environmental conditions. 2. Selection for larger offspring is expected to intensify as environmental quality deteriorates. Models also predict that variable selection on offspring size may favour the evolution of larger offspring than those favoured when selection is constant, or of strategies of variable offspring provisioning (e.g. bet-hedging, plasticity). To date, there is mixed empirical support for the first expectation and few tests of the second. Given, however, that offspring size effects are often estimated under controlled laboratory conditions that presumably downplay their strength and variability, we may not yet understand how selection shapes offspring size in nature. 3. We examined several relationships between offspring size and performance in controlled (laboratory) and natural (field) environments over time for a colonial marine invertebrate, Bugula neritina, and assessed the variability of these relationships by doing so for replicate cohorts. We further developed a simple optimality model to examine whether predictions of the optimal offspring size were similar (or similarly variable) across environments. 4. We found that selection on offspring size varied substantially among laboratory and field environments, and among cohorts in the latter. In the laboratory, our model consistently predicted that mothers should maximize their fecundity by producing the smallest possible offspring. In the field, however, the predicted optimal offspring size varied from the smallest possible size to the largest possible size for different cohorts. 5. Our study suggests that laboratory estimates of offspring size effects, though often necessary, may not always reflect the direction or variability of selection on offspring size under natural conditions. The optimal offspring size for mothers in nature may be an ever-shifting target that shapes provisioning strategies such as bet-hedging or plasticity in offspring size.