Parsimony analysis of endemism under the “areas of endemism as individuals” thesis

Cladistic biogeography relies on two operational processes: one that diagnoses fundamental units and another that utilizes them. Discussion of the latter methodology has resulted in numerous protocols (e.g. Nelson and Platnick, 1981; Wiley, 1988; Page, 1990; Nelson and Ladiges, 1991; Brooks et al., 2001; Wojcicki and Brooks, 2005; Eckstut et al., 2011) that hypothesize area relationships based on areas of endemism, which are considered the fundamental units of historical biogeography (Crisci et al., 2003; Crother and Murray, 2011). Methods to diagnose areas of endemism have also been privy to meticulous examination and contested protocol formation (M€ uller, 1973; Rosen, 1988; Axelius, 1991; Crisci et al., 1991; Henderson, 1991; Platnick, 1991; Morrone, 1994). The most notable protocol for diagnosing areas of endemism that emerged from these discussions was parsimony analysis of endemism (PAE). PAE was originally designed by Rosen (1988) and Rosen and Smith (1988) as a method to generate a historical hypothesis of the relationships between the biotas of sampled localities, producing area cladograms (Rosen, 1988). Rosen (1988) also pointed out that PAE results in patterns of endemicity specific to a set study scale comprised of taxonomically incomplete biotas. Morrone (1994) adopted this method for diagnosing areas of endemism using a scale-appropriate, grid-based locality designation. These two functions represent the dynamic (area relationships) versus static (identification of areas of endemism) approaches of the method (Rosen, 1988; Nihei, 2006). Consistent with the method throughout its initial interpretations is the use of areas as operational taxonomic units (OTUs), presence/absence of taxa as characters, and shared biota as synapomorphies between OTUs. Morrone explained the PAE protocol using the distrubution of the beetle genus Sciobius in South Africa. A summary \of this protocol is as follows. 1 Twenty-one quadrats were assigned to the southeastern portion of South Africa in which at least one locality of the focus clade exists (Fig. 1). 2 A data matrix was constructed with species as characters, presence/absence as character states, and quadrats as OTUs with a hypothetical root OTU with all absent taxa. 3 A parsimony analysis of the data matrix was performed (Fig. 2a). Morrone (1994) interpreted his parsimony analysis by recognizing groups of quadrats that are defined by at least two species as areas of endemism. The Sciobius distribution analysis in South Africa recovered three areas of endemism according to Morrone’s (1994) interpretation (Fig. 2a). The distributions of the taxa delimited in the cladogram were then superimposed on a map to delineate the boundaries of the areas of endemism.