1. A before/after-control/impact experiment (BACI) was used to examine the
shortterm effects of canopy openness on insect herbivores foraging on the s
eedlings of five rain forest species in an unlogged forest in Guyana. Durin
g the first and second study years, insects were collected monthly from 250
sites, representing 9750 seedlings spread over a plot of 1 km(2). At the o
nset of the second collecting year, half of the parent trees at the sites w
ere felled, creating an average canopy openness of 26%.
2. Comparisons of the control and impact sites before and after the felling
showed that overall insect abundance, particularly of psyllids, and specie
s richness increased at the impact sites, whereas evenness decreased marked
ly. These responses were consistent with the prediction that intermediate d
isturbance promotes high diversity in tropical systems.
3. Twenty-nine per cent of insect species analysed showed a significant res
ponse to felling, with some populations doubling or halving at the impact s
ites after felling. The abundance of most species increased, although respo
nses varied even between congeners.
4. A canonical correspondence analysis at the impact sites after felling in
dicated that most species responded to an increase in leaf production by se
edlings, not to canopy openness per se. Species declining at the impact sit
es were more likely to feed indiscriminately on young and mature foliage an
d probably reacted to changes in microclimate.
5. The study illustrates the difficulty in identifying higher taxa, umbrell
a or indicator species that adequately integrate the responses of insect he
rbivores to forest disturbance. Because the integrity of insect communities
in the understorey of this forest depends more on minimizing damage to pla
nt resources than on canopy openness, we found some support for the conserv
ation value of reduced impact logging in tropical wet forests.