RESUMO
A recent epidemic of Phytopthora (Sudden Oak Death) in coastal woodlands of California is causing severe mortality in some oak species belonging to the red oak (Lobatae) group. To predict the risks of spread of this disease, an understanding of the relationships among California's red oak species and of their population genetic structure is needed. We focus here on relationships among the four species of red oak. Whereas morphological distinction of Quercus wislizeni and Quercus parvula can pose problems, Quercus kelloggii and Quercus agrifolia in pure forms are easily distinguishable from one another and from Q. wislizeni and Q. parvula in the field. However, hybrids among all species combinations are known to occur in nature and these can confound data from ecological studies. Our results revealed greatest differentiation of the deciduous Q. kelloggii, with only weak AFLP fragment differentiation of the three remaining evergreen species. The molecular data suggest a closer affinity of Q. agrifolia with Q. wislizeni and Q. parvula contrary to earlier suggestions that its origins are likely to have been with northern deciduous oaks probably through a common ancestor with Q. kelloggii. Interior and coastal populations of Q. wislizeni separated in dendrograms based on phenetic and genetic distances suggesting probable isolation in different glacial refugia. The position of Q. parvula remains ambiguous, having a closer affinity with interior populations of Q. wislizeni and with Q. agrifolia, than with coastal populations of Q. wislizeni. Mean population differentiation in Q. wislizeni was 0.18, which is somewhat higher than the average for other oak species, suggesting that range fragmentation has occurred in the past, resulting in a metapopulation structure. Our results provide evidence that introgression among these species may be causing reticulation, further confounding species separation. Whereas Phytopthora has been reported on Q. agrifolia, Q. parvula and Q. kelloggii, it has not yet been detected in natural populations of Q. wislizeni. The species relationships that our molecular data show suggest that this is more likely a result of escape due to ecological tolerances than to genetic differences.