The evolution of ant pollination systems

The Hymenoptera are the most important pollinators world wide but ants are rarely involved, although abundant in most terrestrial ecosystems, and frequent flower visitors. I review the evidence for the scarcity of ant pollination and the few cases that have been documented. Two very different systems have evolved: The first is limited to one, or possibly two orchid species that occur in southern Australia and exhibit very specialized interactions involving complex floral traits and insect behaviours. By contrast, the second kind involves a variety of widely distributed plant families, is relatively unspecialized, with easily accessible floral rewards that ants share with a variety of small, winged insects. These systems occur chiefly in harsh habitats where plants with low growth forms favours ant visitation. The secretion of surface antimicrobial compounds in ants appears to have been an important factor in limiting the evolution of ant pollination systems. Selection for ant pollination may have occurred at four levels. The first and second levels resulted in the convergent evolution in harsh environments that produced plant communities, plant architectures and floral shapes and rewards that attract ant visitors. The third and fourth levels may involve the selection for particular ant species in pollination systems that otherwise appear to be generalist. This is based on the likelihood that ant species are not functionally equivalent with respect to pollination fidelity and efficiency. A series of traits that may be under selection is suggested. However, investment in ants may be limited since most plant species involved are visited by two functional groups, one winged and one cursorial. Research into cryptic specialization may be rewarding both for ant pollination in particular and plant-pollinator interactions in general.