Evaluating morphology and geographic range extent of genetically delimited species of mushrooms

Species delimitation directly affects interpretation of evolution and biogeography. Following speciation, independently evolving lineages are expected to fix different characters that eventually distinguish them from their closest relatives. However, rates of fixation vary. I delimited species in the mushroom genus Russula based on the fungal nuclear internal transcribed spacer 2 (ITS2) DNA barcode region. I sampled 713 ITS2 sequences of American Pacific Northwest specimens collected by Seattle architect Benjamin Woo (1923-2008). I compared the morphology within and among DNA-delimited species, according to morphological character state data that Woo had recorded for each of specimen. To Woo's data, I added spore measurements for 23 species. The characters in Russula varied within and overlapped between my delimited species. My multivariate analysis showed that the centroids of morphological characters usually differed significantly between pairs of genetically defined species, indicating evolutionary divergence at the level of morphology. However, because of the variation among and within conspecific collections, morphological characters only correctly predicted the identity of ~50% of the individual specimens. Of the delimited species, nine had been collected ten or more times each and were, based on morphology and sequence analysis, undescribed and restricted to North America. I describe the nine as new species, reporting their character variation. I used data from public databases to ask how frequently geographical ranges are intercontinental as opposed to intracontinental among mushroom-forming species. I calculated the ‘range extent’ (maximum geographical distance) recorded for 2324 species world-wide and 341 species from the Pacific Northwest, representing 12 genera. The ranges of most species extended only to ~2000 km (shorter distances than a continent). By permutation of the data, I showed that this pattern vanished if geographical coordinates were randomized with respect to species suggesting the pattern I found in the data was not due to random sampling. More sampling would be needed to resolve whether the pattern arose from sampling bias or a high frequency of regional endemism. However, because it reflects a common pattern seen in the best sampled fungi and in narrower studies of genera and families, I hypothesize that regional endemism is the general pattern in well-studied genera and more generally fungal biogeography.Science, Faculty ofBotany, Department ofGraduat