Evolution 2014: Are island mutualist communities more likely to be nested because they are inherently more unstable?

The interactions in ecological communities can be structured in a variety of ways, and recently there has been a push to categorize these networks along the spectrum between modular (smaller clusters of more specialized interactions) and nested (unclustered networks with more generalist species). Theoretically it is understood that the nested communities are more stable, so why aren’t all interactions networks more towards the nested end of the spectrum? David Hembry of Kyoto University has made a discovery that suggests an answer. In his talk he showed that island mutualistic networks of leafflower plant/leafflower moth species are much more nested than continental networks of similar species. It is well understood that islands are more unstable habitats, both for particular species populations and for overall ecological communities. Are some islands producing ecological conditions that can only support nested communities? Do modular communities blink out frequently enough to be rare?