FIGURE 4 in Vertebrate endemism in south-eastern Africa numerically redefines a biodiversity hotspot

FIGURE 4. Proposed zoogeographical regionalisation based on all vertebrate species endemic to south-eastern Africa sensu lato (a) The Phenetic Cluster Analysis (PCA) dendrogram of hierarchical relationships between operational geographic units (OGUs), using Jaccard's dissimilarity index and the UPGMA algorithm. (b) The resulting regionalisation; dominions, provinces and subprovinces are listed with codes in the table c, while all entities labeled on the map are districts, except for Maputaland subprovince (no districts within). (c) the Hierarchy of proposed zoogeographical entities. All "biogeographical taxa" were detected based on phenon lines at arbitrary levels of dissimilarity, in a way to generate maximally contiguous geographical clusters, except in the identification of the Karoo dominion (dash line), where a subjective decision was made considering the separation between relatively mesic and arid areas in almost all published zoogeographical regionalisations. The centres of endemism (BCOEs = broad and NCOEs = narrow) for south-east Africa dominion are also defined (see text for their interpretation). (d) The strict concensus area cladogram of OGUs from Parsimony Analysis of Endemicity (PAE), highlighting similarities to the PCA in the recovery of the South-east Africa dominion and the Greater Maputaland-Pondoland- Albany province (two OGUs shown in red dashed circles detected in the PAE as outliers are slightly enlarging the province compared to the PCA).