Table 1.

Methods and hypothesis.

Our approach proceeded in two steps. (Step 1) For each substrate, immigration from the seawater species (circles labeled as “SA”, etc.) onto the beads (colored circles) could be stochastic or driven by selection. We started classifying the beads' samples on the basis of their β−diversity similarity. If we did not find clusters (scenario 1), we performed a test of neutrality considering all the samples of that substrate. If we found clusters “community-classes” (scenario 2), we investigated if it was a signature of distinct selection forces acting in different classes (“metacommunity selection”). Under this scenario, neutrality was tested considering each subset of samples determining each class independently, hence creating a different metacommunity for each of them. The interpretation depending on the different outcomes is indicated, with the results we found highlighted in bold in the table. (Step 2) Since selection at the metacommunity and local levels was suggested, we investigated its mechanisms considering the phylogenetic similarity of the samples. The combination of results from Mantel correlograms and β−NTI comparisons led to three cases compatible with our results in Step 1. The analysis of these cases suggested that the dominant processes were those indicated in case C and, to a lesser extent, case B. Very few comparisons supported case A. Note that these cases were restricted to those compatible with our results in Step 1 and other possibilities (e.g., homogenizing dispersal) were excluded.