Abstract
Biological invasions have significant ecological and economic impacts. Much attention is therefore focussed on predicting establishment and invasion success. Trait-based approaches are showing much promise, but are mostly restricted to investigations of plants. Although the application of these approaches to animals is growing rapidly, it is rare for arthropods and restricted mostly to investigations of thermal tolerance. Here we study the extent to which desiccation tolerance and its phenotypic plasticity differ between introduced (nine species) and indigenous (seven species) Collembola, specifically testing predictions of the 'ideal weed' and 'phenotypic plasticity' hypotheses of invasion biology. We do so on the F2 generation of adults in a full factorial design across two temperatures, to elicit desiccation responses, for the phenotypic plasticity trials. We also determine whether basal desiccation resistance responds to thermal laboratory natural selection. We first show experimentally that acclimation to different temperatures elicits changes to cuticular structure and function that are typically associated with water balance, justifying our experimental approach. Our main findings reveal that basal desiccation resistance differs, on average, between the indigenous and introduced species, but that this difference is weaker at higher temperatures, and is driven by particular taxa, as revealed by phylogenetic generalised least squares approaches. By contrast, the extent or form of phenotypic plasticity does not differ between the two groups, with a 'hotter is better' response being most common. Beneficial acclimation is characteristic of only a single species. Laboratory natural selection had little influence on desiccation resistance over 8-12 generations, suggesting that environmental filtering rather than adaptation to new environments may be an important factor influencing Collembola invasions.