|Electronic supplementary material for
The phylogenetic affinities of the bizarre Late Cretaceous Romanian theropod Balaur bondoc (Dinosauria, Maniraptora): dromaeosaurid or flightless bird?
Andrea Cau, Tom Brougham, Darren Naish
This file contains:
Details of phylogenetic analyses
Results of implied weighting analyses
Modified dataset of Brusatte et al. (2014) - Data matrix
Modified dataset of Lee et al. (2014) - Data matrix
References for supplementary material
Details of phylogenetic analyses
Our use of phylogenetic taxonomy follows Longrich and Currie (2009), Turner et al. (2012) and Godefroit et al. (2013). Accordingly, Paraves is the most inclusive clade containing crown birds but not Oviraptor philoceratops; Avialae is the most inclusive clade containing crown birds but not Dromaeosaurus albertensis and Troodon formosus; Dromaeosauridae is the most inclusive clade containing Dromaeosaurus albertensis but not crown birds and Troodon formosus; and Eudromaeosauria is the least inclusive clade containing Deinonychus antirrhopus, Dromaeosaurus albertensis, Saurornitholestes langstoni and Velociraptor mongoliensis.
Tree search strategy
Importing each dataset in TNT (Goloboff et al. 2008b), we performed 100 “New Technology” search replicates with paramaters set using default values (i.e., Sectorial Search set on, Tree Fusing set on; both with default parameters). The result of the first search was then explored performing Tree-Bisection-Reconnection (TBR) heuristic searches, and saving all shortest trees found. For each analysis, no more than 99999 trees were saved to reduce computational time. Nodal support was calculated performing 1000 TBR heuristic search replicates and saving all tree up to ten steps longer than the best score found.
Alternative placements and the Templeton's tests were performed in PAUP (Swofford 2002).
Three iterations of the Implied Weighting Analysis were performed with each data matrix in TNT, setting the concavity parameter k (Goloboff et al. 2008a) to 1 (strong downweighting of homoplastic characters), 3 (default value in TNT, Goloboff et al. 2008b), and 9 (moderate downweighting of homoplastic characters) respectively. Concavity parameters greater than 9 were also tested for both matrices, however, because the topologies of the resultant trees converged closely on the unweighted topologies, they were not considered in the final analyses.
Brusatte et al. (2014) matrix
Turner et al. (2012) suggested that Linheraptor is a junior synonym of Tsaagan, an interpretation recently challenged by Xu et al. (2015). We agree with the latter authors that Linheraptor may eventually be a taxon distinct from Tsaagan, and that any Operational Taxonomic Unit (OTU) based on combined information from both Tsaagan and Linheraptor may represent a chimaera (if future analyses using the two taxa separately do not find them as sister taxa relative to other dromaeosaurids). Nevertheless, it is unclear whether some of the character scores of the Tsaagan OTUs present in both the Turner et al. (2012) and Brusatte et al. (2014) datasets include information based on Linheraptor exquisitus holotype. Although Turner et al. (2012) explicitly suggested the synonymy between these two dromaeosaurids, we note that the Tsaagan OTU in both the Turner et al. (2012) and Brusatte et al. (2014) datasets seem as scored solely on the Tsaagan mangas holotype (a specimen which includes a complete skull as well as partial postcranial bones): after checking the scores of the Tsaagan OTU in the Brusatte et al. (2014) dataset, we were unable to find any character state that could be unequivocally based on the well complete holotype skeleton of Linheraptor exquisitus instead of Tsaagan mangas type specimen. As a single example, although character “118.1” of Brusatte et al. (2015: describing the elongation of the distal caudal prezygapophyses – a well-established synapomorphy of non-unenlagiine dromaeosaurids, Turner et al. 2012) is clearly evident in the Linheraptor type specimen based on the published literature (Xu et al. 2010), the Tsaagan OTU in that dataset is cored as “118.?”, which is the state expected if T. mangas is scored using the only holotype of the latter species (a specimen missing the tail). Therefore, we conclude that the Tsaagan OTU in the Turner et al. (2012) and Brusatte et al. (2014) datasets were probably scored exclusively on the type specimen of Tsaagan, with no scores from Linheraptor. This suggests that the Tsaagan OTU in both datasets is not a potential chimaera, regardless of the taxonomic interpretation of Linheraptor discussed by Xu et al. (2015).
The following characters were modified from their original definitions. Comments regarding the nature of the modifications are included in brackets.
Char. 145: Semilunate carpal, placement relative to metacarpals: medially placed, mostly overlapping metacarpals I and II, marginally or not overlapping metacarpal III (0); laterally shifted, marginally or not overlapping metacarpal I, significantly overlapping metacarpal III (1). [Previously, the character included four states that partially overlapped, describing size and placement of the semilunate carpal. Here, the character is simplified to describe exclusively the relative placement of the semilunate carpal instead of the combination of size and position].
Char. 147: Third manual digit, number of phalanges: (0) four; (1) three; (2) two; (3) one; or (4) splint metacarpal bearing no phalanges. The character is ordered. [States redefined to better describe variation among theropods].
Char. 436: Metatarsal V length: (0) less; or (1) more than 40% of metatarsal III’s length (Brusatte et al. 2013). [Previously, the character did not quantitatively define metatarsal V’s elongation]. The derived state is present exclusively in microraptorine and eudromaeosaurian dromaeosaurids among Coelurosauria. Balaur shows the plesiomorphic state.
The following new characters were included:
Char. 854: Metacarpal I, proximal half: mediolaterally expanded, width comparable to rest of bone (0); narrower than distal half, medial margin sloping proximolaterally (1).
Char. 855: Distally closed intermetacarpal space between metacarpals II and III: (0) absent, metacarpals not contacting distally; (1) present. This character is not redundant with character 391, which describes the extent of the intermetacarpal space among the taxa bearing the closed intermetacarpal space and is scored as ‘inapplicable’ among the other taxa.
Char. 856: Metacarpal III, distal end: (0) bicondylar; or (1) simple convexity.
Char. 857: Dorsal margin of manual unguals: (0) does not; or (1) does arch dorsally above level of articular facet (Senter 2007; Agnolín and Novas 2013).
Char. 858: Interpubic space, width between conjoined pubes: (0) gradually narrowing distally; or (1) wide pubic canal and laterally bowed pubis, followed by an abrupt narrowing at the symphysis.
Char. 859: Length of pedal phalanx I-1: (0) < 66% III-1; or (1) > 66% III-1. The derived state is present exclusively in a subset of avialans among Coelurosauria. Balaur shows the apomorphic state.
Char. 860: Metatarsal II, distal condyles, plantar projection: (0) medial and lateral condyles with comparable projection; (1) medial condyle much further projected ventrally than lateral. (O’Connor et al. 2014). This character describes the marked medial projection of the distal condyles of metatarsal II present in Balaur and some avialans.
Figure S1. Manual ungual I of Microraptor (A), Velociraptor (B) and Balaur (C), in side view, all drawn at the same proximal facet depth, with articular facet dorsoventral axis oriented vertically. Left hand of Deinonychus (D), Archaeopteryx (E), Sapeornis (F), Zhouornis (G), Nothura (H), and Balaur (I), in extensor view, all drawn at the same metacarpal II length. Modified from (A), Senter (2007); (B), Norell and Makovicky (1999); (C, I), Brusatte et al. (2013); (D, E, H), Wagner and Gauthier (1999); (F), Zhou and Zhang (2003); (G), Zhang et al. (2013). Numbers refer to character states in the character list modified from Turner et al. (2012).
Figure S2. Pelvis of Balaur, in lateral (A) and anteroventral (B) views. Pubis of Velociraptor, in posterior view (C). Pelvis of an unnamed enantiornithine from the Maastrichtian of Argentina, in lateral view (D). Pubis of Sapeornis, in anterior view (E). Pelvis of Archaeopteryx (London Specimen), in lateral view (F). Modified from (A, B), Brusatte et al. (2013); (C), Norell and Makovicky (1999); (D), Walker and Dyke (2009); (E), Zhou and Zhang (2003). Numbers refer to character states in the character list modified from Turner et al. (2012).
Figure S3. Tarsometatarsus of Balaur, in medial view (A). Tarsometatarsi of Balaur, Avisaurus archibaldi, Bauxitornis, Yungavolucris and Evgenavis, in extensor view (B-F). Tarsometatarsus of Evgenavis, in medial view (G). All drawn at the same metatarsal III length. Mid-shaft cross section of tarsometatarsus of Avisaurus archibaldi (H). Tarsometatarsus of Evgenavis, in distal view (I). Modified from (A, B), Brusatte et al. (2013); (C, H), Brett-Surman and Paul 1985; (D), based on photograph provided by A. Osi; (E), Chiappe (1993); (F, G, I), O'Connor et al. (2014). Numbers refer to character states in the character list modified from Turner et al. (2012).
Figure S4. Feet of Velociraptor (A), Balaur (B), Patagopteryx (C) and Zhouornis (D), in extensor (A, D), lateroextensor (B), and flexor (C) views. All drawn at the same metatarsal III length. Modified from (A), Norell and Makovicky (1997); (B), Brusatte et al. (2013); (C), Chiappe (2002); (D), Zhang et al. (2013). Numbers refer to character states in the character list modified from Turner et al. (2012).
Lee et al. (2014) matrix
Modifications involved re-definition of character 318 to avoid ambiguity in its interpretation:
Metacarpal III, distal end, medial contact with metacarpal II that is proximodistally extended (metacarpals II-III eventually enclosing an intermetacarpal space): absent (0); present (1).