Community-level effects of experimental control of invasive ants on offshore islets in the Hawai`ian archipelago

Invasive ants are a significant conservation concern and can have far-reaching effects in ecosystems they invade. I used the experimental control of ant numbers on two pairs of small (<5 ha) offshore islets dominated by either the big-headed ant, Pheidole megacephala, or tropical fire ant, Solenopsis geminata, to investigate the influence of ant control on arthropods, plants and seabirds. I also measured effects of application of the formicide hydramethylnon on non-target arthropods. Island-wide hydramethylnon treatments resulted in eradication of P. megacephala and temporary reduction of S. geminata on treated vs. untreated islets. Non-target effects appeared limited to alien cockroaches (Order Blattodea). The eradication of P. megacephala was followed by significant changes in members of the arthropod community including amphipods (Order Amphipoda), spiders (Order Araneae), cockroaches (Order Blattodea), beetles (Order Coleoptera) and crickets (Orthoptera: Gryllidae). Additionally, dynamic changes in the ant fauna were observed including the apparent colonization by three previously undetected species (S. geminata, Tetramorium bicarinatum and Anoplolepis gracilipes). One of these, A. gracilipes, underwent a rapid range expansion during 2006--2008 which corresponded with reduced Wedge-tailed Shearwater (Puffinus pacificus) nest success. In the two years following eradication of P. megacephala, but prior to the A. gracilipes invasion, there was no difference in hatching success, growth, or fledging success of shearwaters.;Alternatively, control of S. geminata on the second islet pair resulted in a temporary increase in shearwater fledging success and decrease in injury frequency. Severely injured chicks (i.e., chicks that lost >20% of tissue on their feet) weighed significantly less than uninjured chicks and did not fledge. In addition, control of S. geminata indirectly caused increased leaf cover in a dominant plant species due to a reduction in sap-sucking insects (Hemiptera: Stenorrhyncha). I conclude that eradication and control of some ant species is possible, and can result in community-level changes including shifts in arthropod community composition, increased fledging success in seabirds, and increased leaf cover of some plants. However ant eradications can also have long-term detrimental effects on ecosystems due to the potential for subsequent colonization by other ant species that are more harmful and more difficult to eradicate.