ure 25) because they had the highest average silhouette values (0.51 and 0.45, respectively). In the PAM analysis for two groups (Figure 25A), all of the phytosaurs (including Diandongosuchus) are grouped together (red), whereas the two outgroup taxa are in green. In the PAM analysis for three groups (Figure 25B), Diandongosuchus was grouped with the parasuchids (red), although it has a negative silhouette value. Lotosaurus (green) and Prestosuchus (blue) were placed in their own separate groups. In the FANNY analysis we ran the dataset in three, four, and five clusters (Figure 26). The average silhouette width for each analysis is as follows: three groups (0.26), four groups (0.21), and five groups (0.24). The FANNY analysis for three groups (Figure 26A) splits up the phytosaurs into two groups (red and blue), with the outgroup and Diandongosuchus forming their own group (green), although Diandongosuchus has a negative silhouette width. The FANNY analysis for four groups (Figure 26B) splits up the larger phytosaur group into two (red and yellow). The FANNY analysis for five groups (Figure 26C) places the outgroup taxon Lotosaurus in its own group, and puts together Diandongosuchus and Prestosuchus (although Diandongosuchus has a negative silhouette width). Additionally, Pal. sawini has a negative silhouette width in its phytosaur group (red). H. Pseudosuchia The Pearson and Spearman BDC results (Figure 27) for the Pseudosuchia subset analysis were uninformative except that they showed Lotosaurus as very distinct, sharing no positive correlation with any other taxon. We suspected that evidence for discontinuity was masked by these results, given that Pseudosuchia is almost certainly made up of many created kinds. As such, we decided to split up the taxa according to the major division of Paracrocodylomorpha vs. non-paracrocodylomorph Pseudosuchia. As Ticinosuchus is difficult to classify and falls right between the two groups, we included it in both subset analyses. I. Non-paracrocodylomorph Pseudosuchia The Pearson and Spearman BDC results for the non-paracrocodylomorph pseudosuchian subset analysis are almost identical (Figure 28), both showing three blocks of positive correlation: Aetosauria (3 taxa), Ornithosuchidae (4 taxa), and Gracilisuchidae + Nundasuchus + Ticinosuchus. The Spearman results include Parringtonia with the gracilisuchid block of positive correlation (Figure 28B), whereas the Pearson BDC shows it on its own (Figure 28A). Both Pearson and Spearman BDC plots show negative correlation between Aetosauria and Ornithosuchidae. The MDS results (Figure 29) agree well with the BDC plots, showing three different clusters: Aetosauria, Ornithosuchidae, and Gracilisuchidae. Nundasuchus clusters closely with the gracilisuchids in a single linear trajectory. Parringtonia and Ticinosuchus are closest to the Gracilisuchidae + Nundasuchus cluster, but they are in a different trajectory, one on each side of the gracilisuchid Yonghesuchus. The PAM and FANNY results agree with the BDC and MDS results Figure 14: Fuzzy analysis (FANNY) of the Rhynchosauria subset dataset: A) FANNY analysis at four groups including problematic taxa. (b) FANNY analysis at four groups excluding problematic taxa. MCLAIN, CLAUSEN, PEREZ, BEEBE, AND AHTEN Archosauromorph Baraminology 2023 ICC 502
RkJQdWJsaXNoZXIy MTM4ODY=