Kīpuka as crucibles of evolution
Our data provided strong evidence for our first hypothesis, that we would document a species-area effect relating to community-wide isolation. Lava flows pose a strong dispersal barrier to core forest arthropods (Roderick et al. 2012, Carson et al. 1990), so we did recover a strong species-area effect consistent with previous results from spiders (Vandergast et al 2004), Drosophila flies (Mueller 2015), birds (Flaspohler et al. 2010), and root fungi associated with M. polymorpha (Vannette et al. 2016). The isolation of communities by lava also caused considerably higher community turnover between kīpuka than between equivalently spaced sites in continuous forest. Even kīpuka isolated by only a few 100 m of lava sometimes showed nearly complete community turnover. As further predicted by our first hypothesis, the community-wide isolation between kīpuka also translated into genetic differences between populations of individual species, as has already been shown in a few spiders (Vandergast et al. 2004, Roderick et al. 2012) and flies (Mueller 2015). The regrowth of forest after volcanic eruptions on Hawaiʻi can take from several hundred to several thousand years, providing ample time for genetic differentiation between populations in individual kīpuka by genetic drift (Lourenço et al. 2017). Allopatric isolation, a major driver of speciation for organisms across the tree of life, can happen over long geological time periods and large geographic distances–through continental drift or the emergence of mountain ranges– but we also know that even very short geographic distances can suffice to isolate populations (Johnson & Munshi-South 2017). Our analysis suggests that fragmentation by lava flows could pose a possible means for such microallopatric differentiation for Hawaiian arthropods.
The extreme community turnover between neighboring kīpuka means that individual arthropod species are exposed to very different communities of interacting species in different kīpuka, and hence different selective pressures. For example, a single species present in different kīpuka might interact with different competitors, predators or parasitoids. Interspecific interactions are highly important drivers of speciation (Thompson 2005), so these distinct biotic communities may accelerate the differentiation of taxa between kīpuka. Besides the possibility kīpuka provide for adaptation to different biotic environments, kīpuka may provide opportunities for adaptation along climatic gradients. For instance, adjacent kīpuka are frequently located along pronounced temperature and precipitation gradients in Hawaiʻi, such that the isolation of species in these kīpuka could foster phenotypic differentiation through local climate adaptation. In a continuous forest this adaptation would likely be swamped by gene flow. While here we focused on kīpuka distributed within a relatively similar climatic zone, an old kīpuka system distributed across a climatic gradient would provide an excellent setup for studying the impacts of microallopatry on rates of phenotypic differentiation. A final means by which kīpuka may foster differentiation is through the merging of formerly isolated kīpuka: admixture of the isolated populations they contain could provide the raw material for rapid evolutionary responses, as has been suggested as the underlying cause for many adaptive radiations (Marques et al. 2019).