utan, and Hylobates, (3) Simiolus, Dendropithecus, and two species each of Proconsul and Ekembo, (4) Equatorius and Afropithecus, and (5) the singleton Morotopithecus. With Spearman correlations the clusters are the same except Morotopithecus joins Equatorius and Afropithecus in a single cluster. With Jaccard distances and Pearson correlations, the clusters are the same as the simple matching and Pearson correlations. But with Spearman correlations and Jaccard distances, the cluster of Equatorius, Morotopithecus, and Afropithecus are joined to the extant ape cluster. Once again, cluster analysis of the Rasmussen et al. (2019) hominoids was quite poor, with average silhouette values <0.3 (Figure 14). Four-cluster medoid partitioning recovered the extant ape cluster with simple matching distances but added Morotopithecus using Jaccard distances. Two additional clusters appear in the four-cluster medoid partitioning using both simple matching and Jaccard distances: (1) Turkanapithecus and the three species of Nyanzapithecus and (2) Equatorius, Afropithecus, and two species each of Proconsul and Ekembo. The fourth cluster produced by medoid partitioning is a heterogeneous collection of the remaining taxa, only one of which has a silhouette width >0 using either simple matching or Jaccard distances. Looking across the four different character matrices used here, there appears to be little consistency in the clustering results. To evaluate this, we selected a subset of 28 taxa that appear in more than one of the four character matrices. We then assessed how often each pair of the 28 taxa co-occur in the same cluster in partitions using different character matrices, distance metrics, distance correlations, and cluster analyses. The results show limited consistency across all clustering partitions with only a few clusters of taxa occurring in most of the methods surveyed (Figure 15). For example, we find that Aegyptopithecus, Ekembo, Proconsul, Dendropithecus, and Simiolus appear in the same cluster in more than 90% of our partitions. This cluster is a curious mix of possibly outgroup taxa. Aegyptopithecus is considered a stem catarrhine, while Ekembo is considered a stem hominoid. Proconsul is sometimes placed in Hominoidea and sometimes in Proconsulidae, while Dendropithecus and Simiolus are thought to be dendropithecids. Similarly, Catopithecus, Epipliopithecus, Laccopithecus, and Lomorupithecus all occur in the same cluster in more than 95% of our partitions. These four may be members of a single family Pliopithecidae, which is also an outgroup to hominoids. Finally, more than 83% of our partitions placed Hispanopithecus, Lufengpithecus, Sivapithecus, Kenyapithecus, Ouranopithecus, Pongo, and Pierolapithecus in the same cluster. These would be a core group of hominoids, including pongines (Pongo, Lufengpithecus, and Sivapithecus) and purported “stem hominoids” Figure 7. Distance correlation results using the Gilbert et al. (2020) characters. Correlations and distance metrics are shown in the diagram. Filled squares indicate significant, positive distance correlation. Open circles indicate significant, negative correlation. BRUMMEL AND WOOD Preliminary Evaluation of Ape Baramins 2023 ICC 156
RkJQdWJsaXNoZXIy MTM4ODY=