Spatial and environmental variables structure sponge symbiont communities.

Understanding the maintenance and origin of beta diversity is a central topic in ecology. However, the factors that drive diversity patterns and underlying processes remain unclear, particularly for host-prokaryotic associations. Here, beta diversity patterns were studied in five prokaryotic biotopes, namely, two high microbial abundance (HMA) sponge taxa (Xestospongia spp. and Hyrtios erectus), one low microbial abundance (LMA) sponge taxon (Stylissa carteri), sediment and seawater sampled across thousands of kilometres. Using multiple regression on distance matrices (MRM), spatial (geographic distance) and environmental (sea surface temperature and chlorophyll α concentrations) variables proved significant predictors of beta diversity in all five biotopes and together explained from 54% to 82% of variation in dissimilarity of both HMA species, 27% to 43% of variation in sediment and seawater, but only 20% of variation of the LMA S. carteri. Variance partitioning was subsequently used to partition the variation into purely spatial, purely environmental and spatially-structured environmental components. The amount of variation in dissimilarity explained by the purely spatial component was lowest for S. carteri at 11% and highest for H. erectus at 55%. The purely environmental component, in turn, only explained from 0.15% to 2.83% of variation in all biotopes. In addition to spatial and environmental variables, a matrix of genetic differences between pairs of sponge individuals also proved a significant predictor of variation in prokaryotic dissimilarity of the Xestospongia species complex. We discuss the implications of these results for the HMA-LMA dichotomy and compare the MRM results with results obtained using constrained ordination and zeta diversity.