Species traits and habitats in springtail communities: A regional scale study

Abstract

Abstract Although considerable work has been done investigating the factors that underlie species trait assemblages in emblematic groups such as plants and vertebrates, more remains to be done in belowground invertebrate species. For this group, in particular, relationships between species traits and habitat preferences are still a matter of debate. To eludicate upon the relationships, springtails were sampled in a heterogeneous landscape centered on the Senart forest, near Paris (northern France), which represented the largest possible array of five environmental gradients (humus forms, vegetation, moisture, vertical strata, and seasons) over which Collembola are known to be distributed. The distance between samples varied from a few cm to several km. Canonical correspondence analysis using species (128) as observations and species trait attributes (30) and habitat indicators (82) as dependent and independent variables, respectively, allowed us to discern whether species habitats and species trait assemblages were related and which trends could be found in trait/environment relationships. It was concluded that, within the studied area, species habitats were significantly associated with species trait assemblages. The main gradient explaining the distribution of species traits combined the vertical distribution of habitats (from the mineral soil to plant aerial parts), and the openness of the environment, suggesting that a complex of many ecological factors is at play. For the ecological traits of Collembola, this gradient corresponded to an increasing contribution of sensory and locomotory organs, bright color patterns, size and sexual reproduction, all attributes associated with aboveground life under herbaceous cover. Another important, although secondary contrast concerned traits associated with habitats far from soil but that were concealed (corticolous vs. all other habitats). Soil acidity and water did not contribute significantly to trait distribution, at least within the limits of our database.