Background/Question/Methods Nearly all mutualisms entail the production of resources by one species that attract and reward the mutualists with which it interacts. As such resource production could otherwise be used for growth or reproduction, mutualists are predicted to minimize these investment costs. Here, we employ optimal defense theory for plant secondary chemicals to examine plant production of extrafloral nectar (EFN) resources to attract and reward ants for protection against herbivores. Through ant exclusion and artificial herbivory experiments, we explore temporal and size-dependent investment in EFN production by buds and fruits of senita cacti (Pachycereus schottii) in the Sonoran Desert. In particular, we test the predictions of optimal defense theory that plants invest more in high value parts (fruits) through constitutive EFN, less in low value parts (buds) through induced EFN, and that constitutive and induced EFN are negatively correlated. Results/Conclusions EFN production by buds and fruits was greatest at night and nearly absent by day. Nocturnal standing stocks of EFN for buds and fruits were lower and nearly absent with ants than without ants. Constitutive levels of EFN were greater in fruits than in buds; only buds showed induced EFN following artificial herbivory. Constitutive EFN scaled positively with fruit size, but not bud size, and was negatively correlated with induced EFN for fruits but not buds. These results indicate that senita may minimize costs by temporally matching EFN resources with that necessary to attract and reward, but not over-supply ant consumers providing the herbivore resistance. Moreover, by shifting from induced to constitutive EFN with the morphogenesis of buds to fruits, our results show that investment in EFN can change as a plant tissue increases from a lower to higher value. Indirect defense through EFN is maintained in accord with direct defenses of plant secondary chemicals, providing additional support for optimal defense theory being a general framework with which to examine investment costs of the many diverse resources produced by plants, lycaenids, and homopterans to attract and reward ants in protection mutualisms.