Aragosaurus

Royo-Torres, Rafael, Upchurch, Paul, Mannion, Philip D., Mas, Ramón, Cobos, Alberto, Gascó, Francisco, Alcalá, Luis & Sanz, José Luis, 2014, The anatomy, phylogenetic relationships, and stratigraphic position of the Tithonian-Berriasian Spanish sauropod dinosaur Aragosaurus ischiaticus, Zoological Journal of the Linnean Society 171 (3), pp. 623-655 : 645

OF ARAGOSAURUS

Although Mannion et al. (2013) incorporated Aragosaurus into a phylogenetic analysis focused on basal macronarians, recovering it as a basal member of this clade, outside Titanosauriformes, this study excluded a number of other contemporaneous Iberian taxa (i.e. Losillasaurus , Lourinhasaurus , and Turiasaurus ) that might have some bearing on its relationships. In order to further understand the phylogenetic position and relationships of Aragosaurus with regard to other Iberian taxa, we have used the updated matrices of Wilson (2002) and Upchurch et al. (2004a), following Wilson & Upchurch (2009), Royo-Torres et al. (2012), and Royo-Torres & Upchurch (2012). In the case of the Upchurch et al. (2004a) data matrix, Lapparentosaurus , Nigersaurus , ‘ Pleurocoelus-tex ’ (a chimera comprising material now referred to Astrophocaudia , Cedarosaurus , and Sauroposeidon ; D’Emic & Foreman, 2012; D’Emic, 2013), Andesaurus , and Argentinosaurus were omitted a priori. For each data matrix, we have: (1) split the Brachiosaurus operational taxonomic unit (OTU) into Brachiosaurus altithorax and Giraffatitan brancai , following Taylor (2009); and (2) added in or revised scores for Cedarosaurus (based on Tidwell, Carpenter & Brooks, 1999 and D’Emic, 2012; Wilson matrix only), Europasaurus (based on Sander et al., 2006), Lourinhasaurus (based on Mocho et al., 2014), Tehuelchesaurus (based on Carballido et al., 2011b), and Venenosaurus (based on Tidwell et al., 2001; D’Emic, 2012). These new codings are provided in the Appendix. Data matrices were analysed using TNT 1.1 ( Goloboff, Farris & Nixon, 2003) in order to find the most parsimonious trees (MPTs). We implemented a heuristic tree search, performing 1000 replications of Wagner trees (using random addition sequences) followed by tree bisection reconnection (TBR) as the swapping algorithm, saving 100 trees per replicate. In order to test the support of the phylogenies, the Bremer support and the bootstrap values were obtained (absolute frequencies based on 5000 replicates; Figs 17 and 18 ).

Figures

Figure 17. The strict consensus tree produced by phylogenetic analysis of the updated Wilson (2002) data set. Bootstrap values are shown as percentages (nodes lacking percentages have bootstrap values of less than 50%). Bremer supports are shown in square brackets.

Figure 18. The strict consensus tree produced by phylogenetic analysis of the updated Upchurch et al. (2004a) data set. Bootstrap values are shown as percentages (nodes lacking percentages have bootstrap values of less than 50%). Bremer supports are shown in square brackets.