Little is known about the spatial distribution and abundance of tree holes in New Zealand’s native forests, or the invertebrate communities that they support. I found that tree holes were common on five endemic tree species, belonging to the families Fagaceae and Podocarpaceae in the mixed broadleaf-podocarp rainforest of Orikaka Ecological Area, Buller District, New Zealand. However, tree holes were not uniformly distributed throughout the forest, with more holes found on the three podocarp species, Prumnopitys ferruginea, P. taxifolia and Dacrycarpus dacrydioides, than on Nothofagus fusca or N. menziesii. Nevertheless, Nothofagus fusca had the largest holes of any of the tree species sampled and larger trees generally had larger holes. Large, hole-bearing Nothofagus fusca trees support a specialist hole-dwelling vertebrate fauna in New Zealand and worldwide, tree holes provide habitat for a range of invertebrate species. Using specially-designed emergence traps, I collected invertebrates emerging from naturally-occurring dry tree holes and compared this assemblage with invertebrates inhabiting leaf litter on the forest floor and those dispersing aerially throughout the study area. At the higher taxonomic resolution (i.e., Order or Class), community composition within the tree holes was highly variable, and there was no strong distinction between invertebrates from tree holes, leaf litter or Malaise traps. Moreover, although some beetle species emerging from tree holes were found exclusively in tree holes, most of these were represented by a single individual. Consequently, only minor differences in species composition were detected between beetle assemblages from tree holes, leaf-litter and those aerially dispersing throughout the forest. In contrast, the aquatic invertebrate assemblage within water-filled tree holes was highly distinctive from that in ground-based freshwater ecosystems, with only six aquatic taxa in common between all freshwater habitats. Using experimental water-filled tree-hole microcosms, I found that species richness and community composition within these microcosms were primarily driven by resource concentration, although habitat quality (i.e., water chemistry parameters) was also an important determinant of the identity and composition of colonising species. Overall, my study has shown that tree holes are common in the study area, and are likely to be more abundant in New Zealand’s indigenous forests than previously thought. Moreover, these generally small, discrete forest ecosystems support a diverse array of terrestrial invertebrates as well as a distinctive aquatic invertebrate community that is primarily structured by organic matter resource availability. These findings not only represent an important advance in our knowledge of New Zealand’s freshwater invertebrate biodiversity, but also highlight the need for further investigation into these unique forest canopy habitats which may well be at risk from deforestation and land use change.