The first invasive ants were detected in Hawai`i Volcanoes National Park (HAVO) more than 80 years ago. Ecological impacts of these ants are largely unknown, but studies in Hawai`i and elsewhere increasingly show that invasive ants can reduce abundance and diversity of native arthropod communities as well as disrupt pollination and food webs. Prior to the present study, knowledge of ant distributions in HAVO has primarily been restricted to road- and trail-side surveys of the Kīlauea and Mauna Loa Strip sections of the park. Due to the risks that ants pose to HAVO resources, understanding their distributions and identifying tools to eradicate or control populations of the most aggressive species is an important objective of park managers. We mapped ant distributions in two of the most intensively managed sections of the park, Mauna Loa Strip and Kahuku. We also tested the efficacy of baits to control the Argentine ant (Linepithema humile) and the big-headed ant (Pheidole megacephala), two of the most aggressive and ecologically destructive species in Hawai`i. Efficacy testing of formicidal bait was designed to provide park managers with options for eradicating small populations or controlling populations that occur at levels beyond which they can be eradicated.

Within the Mauna Loa Strip and Kahuku sections of HAVO we conducted systematic surveys of ant distributions at 1625 stations covering nearly 200 km of roads, fences, and transects between August 2008 and April 2010. Overall, 15 ant species were collected in the two areas, with 12 being found on Mauna Loa Strip and 11 at Kahuku. Cardiocondyla kagutsuchi was most widespread at both sites, ranging in elevation from 920 to 2014 m, and was the only species found above 1530 m. Argentine ants and big-headed ants were also found in both areas, but their distributions did not overlap. Surveys of Argentine ants identified areas of infestation covering 560 ha at Mauna Loa Strip and 585 ha at Kahuku. At both sites, upper boundaries of big-headed ants coincided with lower boundaries of Argentine ants. Significantly, Wasmannia auropunctata (little fire ant) was not detected during our surveys.

Formicidal baits tested for controlling Argentine ants included Xstinguish^TM (containing fipronil at 0.01%), Maxforce^® (hydramethylnon 1.0%), and Australian Distance^® (pyriproxyfen 0.5%). Each bait was distributed evenly over four 2500 m² replicate plots. Applications were repeated approximately four weeks after the initial treatment. Plots were subdivided into 25 subplots and ants monitored within each subplot using paper cards containing tuna bait at approximately one week intervals for about 14 weeks. All treatments reduced ant numbers, but none eradicated ants on any of the plots. Xstinguish^TM produced a strong and lasting effect, depressing ant abundance below 1% of control plot levels within the first week and for about eight weeks afterward. Maxforce^® was slower to attain maximum effectiveness, reducing ants to 8% of control levels after one week and 3% after six weeks. Australian Distance^® was least effective, decreasing ant abundance to 19% of control levels after one week with numbers subsequently rebounding to 40% of controls at four weeks and 72% at 10 weeks. In measurements of the proportion of bait cards at which ants were detected, Xstinguish^TM clearly out-performed Maxforce^®, reaching a minimum detection rate of 3% of bait cards at one week compared to a low of 19% for Maxforce^® two weeks following the second treatment. Although ant abundances were dramatically reduced on Xstinguish^TM plots, it is not currently registered for use in the USA. Our results suggest that ant abundance can be greatly reduced using registered baits, but further research is needed before even small-scale eradication of Argentine ants can be achieved.