Data from: Elevation and leaf litter interact in determining the structure of ant communities on a tropical mountain

Tropical mountains encompass a wide range of environmental conditions and are useful models for studying drivers of community structure. Invertebrate species richness and abundance show various elevational patterns. However, the drivers of these differences are not well understood, although microhabitat complexity is potentially important. We studied ground-dwelling ants using pitfall trapping and hand collection on Mt. Wilhelm (Papua New Guinea) from 169 to 3,795 m a.s.l. We tested for the effects of elevation and leaf litter depth (as a measure of microhabitat complexity) on ant abundance, species richness and composition. We sampled 118 species, with ants present up to 2,331 m a.s.l. Species richness peaked at mid-elevation (~700 m), but the elevational pattern for abundance varied depending on sampling scale. Leaf litter depth negatively affected abundance once elevation had been accounted for, while elevation and litter depth had an interactive effect on species richness. Species richness was positively related to litter depth at lower elevations, but negatively above ~700 m. Species composition varied with elevation and less strongly with leaf litter depth. We speculate that in the lowlands, litter depth rather than temperature limits ant communities. At high elevations, the deeper litter decreases temperature of the litter layer, and temperature becomes limiting. At mid elevations, temperature is not yet too low, and litter is still relatively deep, hence generating a mid-elevation peak in ant richness. Our results may explain differing richness-elevation patterns of litter arthropods around the world, and provide testable predictions for future studies on this topic. Data on the ground-dwelling ant communities and the associated environmental variables from Mt. Wilhelm Papua New Guinea. See the Metadata sheet in the xlsx data file for complete description of variables in datasheets. In addition, a complete R code for the study analyses and Figs is available in the second file.