Abstract
In some insect nursery pollination mutualisms, plant hosts impose net costs to uncooperative “cheater” symbionts. These “sanctions” promote mutualism stability but their precise adaptive nature remains unclear. In fig–wasp mutualisms host trees ( Ficus spp.) are only pollinated by female agaonid wasps whose larvae only use galled fig flowers as food. In actively pollinated systems, if wasps fail to pollinate, sanctions can result via fig abortion, killing all wasp offspring, or by increased offspring mortality within un‐aborted figs. These sanctions result from selective investment to pollinated inflorescences, a mechanism present in almost all angiosperms. To more fully understand how selective investment functions as sanctions requires the measurement of variation in their costs and benefits to both hosts and symbionts. Gynodioecious fig‐tree–fig‐wasp mutualisms are particularly suitable for this because pollen and wasps are produced only in the figs of “male” trees and seeds only in the figs of “female” trees. Male and female trees thus incur different net costs of pollen absence, and costs of sanctions to pollen‐free “cheater” wasps only occur in male trees. We used the actively pollinated host tree Ficus hispida and introduced into male and female figs either 1, 3, 5, 7, or 9 all pollen‐laden “cooperative” (P+) or all pollen‐free “cheater” (P−) wasps. Abortion in both male and female trees was highest in P− figs, with P− fig abortion higher in females (~90%) than in males (~40%). Fig abortion was negatively associated with foundress number mainly in P+ figs; in P− figs abortion was only weakly associated with the number of “cheater” wasps, especially in female figs. In un‐aborted male figs, wasp offspring mortality was higher in P− figs than in P+ figs, and in P− figs correlated positively with foundress (cheater) number. Increased offspring mortality was biased against female wasp offspring and likely resulted from reduced larval nutrition in unpollinated flowers. Variation in selective investment to P− figs thus reflects costs and benefits of pollen absence/presence to hosts, variation that translates directly to net costs to cheater wasps.