It is well accepted that the shape of the dispersal kernel, especially its tail, has a substantial effect on the genetic structure of species. Theory predicts that dispersal by fat-tailed kernels reshuffles genetic material, and thus, preserves genetic diversity during colonization. Moreover, if efficient long-distance dispersal is coupled with random colonization, an inverse isolation effect is predicted to develop in which increasing genetic diversity per colonizer is expected with increasing distance from a genetically variable source. By contrast, increasing isolation leads to decreasing genetic diversity when dispersal is via thin-tailed kernels. Here, we use a well-established model group for dispersal biology (peat mosses: genus Sphagnum) with a fat-tailed dispersal kernel, and the natural laboratory of the Stockholm archipelago to study the validity of the inverse isolation hypothesis in spore-dispersed plants in island colonization. Population genetic structure of three species (Sphagnum fallax, Sphagnum fimbriatum and Sphagnum palustre) with contrasting life histories and ploidy levels were investigated on a set of islands using microsatellites. Our data show (, amova, IBD) that dispersal of the two most abundant species can be well approximated by a random colonization model. We find that genetic diversity per colonizer on islands increases with distance from the mainland for S. fallax and S. fimbriatum. By contrast, S. palustre deviates from this pattern, owing to its restricted distribution in the region, affecting its source pool strength. Therefore, the inverse isolation effect appears to hold in natural populations of peat mosses and, likely, in other organisms with small diaspores.