Dominance and species co-occurrence in highly diverse ant communities: a test of the interstitial hypothesis and discovery of a three-tiered competition cascade.

The role of competitive exclusion is problematic in highly diverse ant communities where exceptional species richness occurs in the face of exceptionally high levels of behavioural dominance. A possible non-niche-based explanation is that the abundance of behaviourally dominant ants is highly patchy at fine spatial scales, and subordinate species act as insinuators by preferentially occupying these gaps--we refer to this as the interstitial hypothesis. To test this hypothesis, we examined fine-scale patterns of ant abundance and richness according to a three-tiered competition hierarchy (dominants, subdominants and subordinates) in an Australian tropical savanna using pitfall traps spaced at 2 m intervals. Despite the presence of gaps in the fine-scale abundance of individual species, the combined abundance of dominant ants (species of Iridomyrmex, Papyirus and Oecophylla) was relatively uniform. There was therefore little or no opportunity for subordinate species to preferentially occupy gaps in the foraging ranges of dominant species, and we found no relationship between the abundance of dominant ants and nondominant species richness at fine spatial scales. However, we found a negative relationship between subdominant and subordinate ants, a negative relationship between dominant and subdominant ants, and a positive relationship between dominant and subordinate ants. These results suggest that dominant species actually promote species richness by neutralizing the effects of subdominant species on subordinate species. Such indirect interactions have very close parallels with three-tiered trophic cascades in food webs, and we propose a "competition cascade" where the interactions are through a competition rather than trophic hierarchy.