Despite rapid technological advances, most systematists still rely on morphological characters for species recognition. This is especially true for groups such as Quercus, for which little molecular work has been completed. We have a long-term goal of establishing species profiles from molecular data, but our sampling of trees in nature must be based on morphological criteria. In our investigations of the only two taxa (Quercus ellipsoidalis Hill and Q. rubra L.) of Quercus section Lobatae (the red and black oaks) native to the upper midwest (northern Michigan, northern Wisconsin, and western and northern Minnesota), we have relied on a variety of silvic and morphological characters to identify individual trees as members of one species or the other, or as putative hybrids. In this report, we examine the utility of acorn characters for delimiting these two species. Eight quantitative characters were scored for a minimum of ten trees (3-15 acorns per tree) in each of 23 populations. Three populations are sympatric in the Upper Peninsula of Michigan. The other 20 populations are from areas to the west and east, representing 10 populations of each species from sites where no other red or black oaks occur. Both univariate and multivariate analyses indicate that none of the quantitative variables is diagnostic for either species. However, our analyses demonstrate that the species can be recognized by employing a classic phenetic species concept.

Key words: multivariate analyses, phenetics, Quercus, species