Summary
1. We tested thehypotheses that feeding guild structure ofbeetle assemblages changedwith different
arboreal microhabitats and that thsse differsnces were consistent across rainforest tree species.
2. Hand collection and beating techniques were used from the gondola of the Australian Canopy
Crane to collect beetles from five microhabitats (mature leaves, flush leaves, flowers, fruit and
suspended dead wood) within the rainforst €nopy. A sirnple randomization procedure was
implemented to test whether the abundances of each feeding guild on each microhabitat were
different from that expected based on a null hypothesis of random distribution of individuals
across microhabitats.
3. Beetles from different feeding guilds were not randomly distributed, but congregated on those
microhabitats that are likely to provide the highest concentrations of their preferred food sources.
Herbivorous beetles, in particular, were oyer-represented on flowers and flush foliage and underrepresented
on mature leaves and dead wood. Proportional numbers of species within each feeding
guild were remarkably uniform across tree spcies for each microhabitat, but proportional abundances
of feeding guilds were all sigrrificantly non-uniforrrly distributed between host tree species,
regardless ofmicrohabitat, confirmingpatternspreviously found for arthropods iu trees in temperate
and tropical forests.
4. These results show that the canopy beetle community is partitioned into discrete assemblages
between microhabitats and that this partitioning arises beeuse of differences in feeding guild
structure as a functior of the diversity and the temporal and spatial availability of resources found
oneachmicrohabitat.