Methods Study sites Five middle to high elevation mesic shrubland or savannah Hydroxylase inhibitor ecosystem sites were chosen on the islands of Maui
and Hawaii, such that each represented a homogeneous habitat undergoing invasion by an expanding unicolonial population of invasive ants. The five sites were all located in natural areas supporting mostly native vegetation; none represented an invasion from a habitat edge. Habitat homogeneity within each site was judged by consistency of vegetative community type and species composition, as well as by the lack of apparent changes in substrate type or levels of disturbance. There were differences between sites, Acalabrutinib in vitro however, in substrate age, annual rainfall and vegetative type and composition, and hence arthropod density and diversity. The five sites were: Puu O Ili, at 2360 m elevation on the west slope of Haleakala volcano, Haleakala National Park, Maui; Kalahaku, upslope from Puu O Ili at 2800 m elevation in Haleakala National Park; Ahumoa, at 1880 m on the southwestern slope of Mauna Kea, Hawaii Island; Pohakuloa, at 2060 m elevation
on the south slope of Mauna Kea, Hawaii Island; Huluhulu site, at 2040 m elevation in the saddle between Mauna Kea and Mauna Loa, Hawaii Island. These sites are described more fully in Krushelnycky and Gillespie (2008). The Ahumoa site is being invaded by the big-headed ant (P. megacephala), while the other four sites are all being invaded by the Argentine ant (L. humile). These two species are among the most dominant invasive ATM Kinase Inhibitor Galactosylceramidase ants worldwide, and are primarily generalist predators and scavengers, but can also engage in extensive tending of honeydew-producing Hemiptera (Holway et al. 2002). We chose to examine correlates of species vulnerability at the five sites together, combining the effects of the two ant species, for several reasons. In addition to their similar generalist diets, the two ant species are similar in size, and at our sites the big-headed ant occurred at densities and exerted impacts that were intermediate to those
of the Argentine ant (Supplementary Table 1). Furthermore, big-headed ants did not influence rates of variability in population-level impacts differently than did Argentine ants (see “Results”), and separate laboratory behavioral studies indicated that the two ant species exhibited similar aggression towards the same groups of herbivore species (Krushelnycky 2007). Sampling design As in most studies examining the impacts of invasive ants on arthropod communities, we assessed ant effects by comparing arthropod communities in invaded areas with adjacent uninvaded areas. Our sites were carefully selected so as to minimize confounding factors that might be associated with static ant distributional limits, habitat gradients, or with invasions from habitat edges.