The results of our compatibility studies will be evaluated in light of the application of several "phylogenetic" species concepts (Cracraft, 1983; Donoghue, 1985; de Queiroz and Donoghue, 1988, 1991; Nixon and Wheeler, 1990; Davis and Nixon, 1992; Baum and Shaw, 1995; Baum and Donoghue, 1995). Although such concepts have been discussed at great length, and have engendered heated debate, there are few concrete examples where any of these have been applied (but see, e.g., Crisp and Weston, 1993; Davis and Manos, 1991; Freudenstein and Doyle, 1994; Theriot, 1992), and virtually none in which the results of applying several different concepts are compared in detail.
We intend to contrast two main approaches. The first approach (phylogenetic species sensu Donoghue, 1985; see de Queiroz and Donoghue, 1988) views species as historical groups (Baum and Donoghue, 1995). Individual specimens, representing natural populations, will be scored for a variety of characteristics, and the resulting data matrix will be analyzed under parsimony. The smallest clusters of terminals for which there exists evidence of monophyly (e.g., one or more presumed apomorphies), will then be delimited as species. The smallest unresolved sets of terminals that are not marked by characters as being monophyletic (but which may, in fact, be so), will also be recognized as species, but their equivocal status will be recognized using a standard convention (e.g., "metaspecies," with names marked by an asterisk; Donoghue, 1985). Species delimited in this way will be compared with those derived from a second approach, wherein species are "the smallest aggregation of populations (sexual) or lineages (asexual) diagnosable by a unique combination of character states in comparable individuals (semaphoronts)" (Nixon and Wheeler, 1990). To delimit such "diagnostic" species we will employ the method described by Davis and Manos (1991; also see Davis and Nixon, 1992; Freudenstein and Doyle, 1994), wherein terminal units are delimited through a process of "population aggregation", grouping together those that are not distinct on the basis of characters. If we obtain significant levels of variation in several genes, we may also be able to apply the "genealogical" species concept of Baum and Shaw (1995; also see Baum and Donoghue, 1995). This approach centers on the coalescence of genes, and in practice relies on assessing the congruence of separate gene trees.
These analyses will allow us to compare and contrast the results of different approaches. Although it is sometimes assumed that there will be few instances in which species limits will differ significantly under these different approaches, this conjecture has not been subjected to significant tests in particular cases. Our primary aim in considering the alternatives is not to select one species concept over another, but rather to explore where there are areas of agreement, and to try to clarify the general circumstances under which there may be significant differences. In this case we have the added advantage of being able to compare information on breeding barriers, and we will have assembled gene trees from both nuclear and mitochondrial genomes. Perhaps most significantly, these studies will serve to bring fungi to the attention of those carrying on theoretical work on the species problem. These issues are also clearly of central importance in regard to conservation issues (Hibbett et al., in review).