Background/Question/Methods The foremost answer to the famous question “Why is the world green?” is that plants have evolved a battery of defenses against herbivores. However, the mechanisms by which this evolutionary arms race progresses are not well understood. It is generally assumed that secondary metabolism evolves via the elaboration of an ancestral pathway, such that selection for novel catalytic activity permits the synthesis of a new, more toxic metabolite. Additionally, it is suggested that trade-offs occur between investment in secondary metabolites and investment in extra-floral nectaries (ant defense). We tested these ideas by characterizing the defenses of many species in the tropical tree genus, Inga (Leguminosae). Inga is highly speciose (>300 species) and ecologically important, typically being the most diverse and abundant genus in rainforests of South and Central America. Results/Conclusions In Panama, species with greater ant-visitation also had more toxic secondary metabolites. The reverse was true in Ecuador. We found examples of novel compounds derived through embellishment of ancestral backbones, as predicted by co-evolutionary theory. However, our results suggest that other mechanisms may be more important. For example, the major mode of evolution appears to proceed by reshuffling standard building blocks in novel combinations, presumably via mutations in regulatory genes or promoters. Consistent with this is the fact that closely related plant species are not necessarily more similar chemically, going against a standard assumption of coevolutionary theory. Regulatory changes would lead to large and rapid phenotypic shifts, permitting a long-lived tree to evolve new traits as fast as its insect herbivores.