Whole-tree silvic identifications and the microsatellite genetic structure of a red oak species complex in an Indiana old-growth forest

The red oaks (Quercus section Lobatae) include important timber species, but we know little about their gene pools. Red oak species can be difficult to identify, possibly because of extensive interspecific hybridization, although most evidence of this is morphological. We used 15 microsatellite loci to examine the genetic composition of a red oak community in 20.6 ha of an Indiana old-growth forest. The community included northern red oak (Quercus rubra L.), Shumard oak (Quercus shumardii Buckl.), and pin oak (Quercus palustris Muenchh.). Species were identified using whole-tree silvic characters, the approach most often implemented by foresters. We found high genetic diversity within species but limited genetic differences between species. Phenetic clustering showed that Q. rubra and Q. shumardii were more genetically similar than either was to Q. palustris, but a neighbor-joining tree revealed that individuals of the different species did not resolve into single-species clusters. We identified four mixed-species subpopulations using Structure, a computer program based on Monte Carlo simulation. The three largest groups are consistent with the following biological interpretations: (i) pure Q. rubra, (ii) Q. rubra, Q. shumardii, and their hybrids, and (iii) Q. rubra, Q. shumardii, Q. palustris, and their hybrids. We discuss the implications of these findings for the whole-tree silvic approach to selection and for management of the red oak gene pool.