Aims Plants with extrafloral nectaries (EFNs) are common in tropical and subtropical habitats and, despite many other arthropods also forage for EFN, most studies solely focused on the defense mutualisms between EFN plants and ants. This study aims at a quantitative assessment of the entire arthropod community that visits EFN trees to compare visitor communities between different tree species to disentangle the mechanisms that may drive EFN visitor community composition. We also test for tree diversity effects on EFN visitors, as it is unknown if local tree species richness relates to the abundance and species richness of arthropods foraging for EFN. Methods We sampled EFN-visiting arthropods in the experimental tree species richness gradient of the BEF-China Experiment, the currently largest forest diversity experiment in the world, and tested if tree species richness affects EFN visitors and if visitor community composition differs between EFN tree species. In a second step, we analyzed the EFN of Ailanthus altissima and Triadica cochinchinensis, the two EFN tree species with highest visitor abundance, for sugars and amino acids (AA) to test if tree species-specific differences in nectar chemistry translate to differing visitor communities. Lastly, we conducted a choice experiment using different artificial nectar solutions to test if nectar quality affects foraging decisions of ants, the most frequent EFN visitors in our study sites. Important Findings EFN trees in young successional forests in subtropical South-East China are visited by a diverse assemblage of arthropods including ants, beetles, flies, and spiders. Albeit ants accounted for about 75% of all individuals, non-Ant visitors were by far more species rich. Visitor abundance and species richness declined with increasing tree species richness, suggesting a resource dilution effect, because plots with more tree species had proportionally less EFN tree individuals and thus lower nectar availability. Ailanthus altissima and T. cochinchinensis were visited by different arthropods and their nectar had species-specific AA composition and sugar concentration, indicating that differences in visitors may, at least partly, be explained by differences in nectar chemistry. These findings are supported by the choice experiment, in which artificial nectars containing sugar solutions supplemented with essential AAs attracted more ants than pure sugar solutions or sugar solutions supplemented with non-essential AAs. Our results improve the understanding of the complex ecology of EFN trees, a plant life form that might be crucial for understanding how tree diversity influences patterns of tree growth in young successional tropical and subtropical forests.