Most multicellular species alive are tropical arthropods associated with plants. Hence, the host-specificity of these species, and their diversity at different scales, are keys to understanding the assembly structure of global biodiversity. We present a comprehensive scheme in which tropical herbivore megadiversity can be partitioned into the following components: (A) more host plant species per se, (B) more arthropod species per plant species, (C) higher host specificity of herbivores, or (D) higher species turnover (beta diversity) in the tropics than in the temperate zone. We scrutinize recent studies addressing each component and identify methodological differences among them. We find substantial support for the importance of component A, more tropical host species. A meta-analysis of published results reveals intermediate to high correlations between plant and herbivore diversity, accounting for up to 60% of the variation in insect species richness. Support for other factors is mixed, with studies too scarce and approaches too uneven to allow for quantitative summaries. More research on individual components is unlikely to resolve their relative contribution to overall herbivore diversity. Instead, we call for the adoption of more coherent methods that avoid pitfalls for larger-scale comparisons, for studies assessing different components together rather than singly, and for studies that investigate herbivore beta-diversity (component D) in a more comprehensive perspective.