Effect of ants on the structure of plant-herbivore food webs in tropical forests

Lowland tropical forests are characterised by high densities of ants, with prominent roles as predators and mutualists of herbivorous insects. We propose two-pronged approach to the study of the impact of ants on tropical herbivorous communities, comprising with (i) experimental removal of ants from 0.5 ha plots of lowland tropical forests, and (ii) comparative studies of ants and herbivores along a tropical altitudinal gradient, characterised by decreasing ant abundance with altitude.

The experimental approach is appropriate for the study of ecological effects of ants in lowland communities, evolutionarily adapted to high ant abundance. We advocate "whole forest" experimental manipulation of ant densities, i.e. eradication of ants from relatively large, isolated patches of lowland forest comprising diverse vegetation. An removal from tropical forests can be difficult and has not been attempted larger scale. However, our pilot experiment succeeded in the removal of most of the ants from 25 x 25 m plots of primary and secondary forests, using 300 traps placed in the canopy, baited withy poisoned tuna baits.

The comparative approach looking into differences between lowland and montane forests is useful in assessing differences between herbivore communities adapted to respectively high and low ant densities over long time scale. The number of ants rapidly decreases with altitude in the tropics so that montane herbivores are adapted to relatively ant-free environment. However, altitudinal gradients in many other variables, including temperature and forest composition, can make the assessment of the impact of ants difficult. It is therefore important to combine both experimental and comparative approaches.

Photos from pilot ant removal experiment in a lowland forest and proposed study sites along altitudinal gradient

Photo gallery:

[ Ant eradication experiment ]