The Optimal Defense Theory (ODT) postulates that reproductive structures should be more heavily defended because they contribute the most to a plant's fitness and have the highest probability of being attacked by herbivores. Ants can provide indirect defense to plants through their mutualistic interactions with hemipteran insects. In this well-studied interaction, ants provide protection to hemipterans against their natural enemies (e.g., predators and parasitic wasps) in exchange for the sugar-rich honeydew secreted by hemipterans. In turn, ants attending hemipterans can indirectly benefit plants by suppressing other harmful herbivores. Despite the numerous investigations of this ant-hemipteran mutualism, patterns of optimal allocation to indirect plant defenses mediated by this mutualistic interaction have not been previously investigated. In this study, we evaluated whether allocation to indirect plant defenses mediated by ant-hemipteran interactions and the effectiveness of such indirect defenses differ between vegetative branches (with only leaves) and reproductive branches (with leaves and inflorescences) of the tropical shrub Solanum lycocarpum. For this, we selected plants with aggregations of the ant-tended hemipteran Enchenopa brasiliensis in both reproductive and vegetative branches. We then estimated indirect defenses (measured as the production of sugar in the plant sap and hemipteran honeydew, as well as ant attraction), and their effectiveness in terms of plant protection by ants (measured as the damage and survival of leaf-chewing herbivores). Supporting ODT predictions, we found that the sugar concentration in the plant sap, and consequently in hemipteran honeydew, was higher in reproductive than in vegetative branches. However, the increase in sugar concentration in hemipteran honeydew did not result in greater ant attraction to reproductive branches. Additionally, contrary to ODT predictions, we found that ants attending hemipterans did not enhance plant protection against leaf-chewing herbivores on reproductive branches. Overall, our study demonstrated that the patterns of allocation to indirect defenses mediated by ant-hemipteran interactions in S. lycocarpum plants did not support the predictions of the ODT.