Staurikosaurus pricei as a herrerasaurid

Prior to the work of Novas (1992), who identified several characters thought to unite S. pricei and H. ischigualastensis in a single clade (Fig. 34b), the hypothesis of a close relationship between both species forming the clade Herrerasauridae was contested by several authors (Galton 1977; van Heerden 1978; Brinkman & Sues 1987; Sues 1990). Herrerasauridae is nowadays widely accepted as monophyletic and wellsupported by a suite of postcranial characters (Novas 1993, 1996, 1997; Sereno & Novas 1993; Sereno 1999; Kellner & Campos 2000; Rauhut 2003; Bittencourt & Kellner 2004; Langer 2004; Langer & Benton 2006). Reanalysis of the matrix of Langer & Benton (2006) resulted in the following synapomorphies for Herrerasauridae: 1) centra and neural spines of caudal dorsal vertebrae axially shortened; 2) ribs of the first primordial sacral vertebra are deeper than half the depth of the ilium; 3) neural spines of proximal caudal vertebrae dorsally directed; 4) prezygapophyses of distal caudal vertebrae overlap more than a quarter of the adjacent centrum; 5) ventral portion of the postacetabular process does not bear a deep fossa in its caudal portion for the origin of M. caudofemoralis brevis (= reduction of the brevis fossa); 6) lateral margin of the pubis caudally folded at its distal portion, with the pubic pair showing a U-shaped transverse section; 7) caudolateral flange of distal tibia short and does not project to the fibula. Characters 1–3 are seen only in herrerasaurids, but character 4 is also observed in theropods (Gauthier 1986; Novas 1992), and was formerly used to support a phylogenetic position of herrerasaurids within Theropoda (e.g. Sereno 1999). In the phylogenetic hypothesis of Langer & Benton (2006) this character is interpreted as independently acquired in theropods and herrerasaurids. Characters 5 and 7 are reversals to the ancestral state for Dinosauria; usage of these characters to diagnose Herrerasauridae is unproblematic because the position of this clade within dinosaurs is well-supported (see above). The presence of plesiomorphic character states in taxa of a given ingroup should not be surprising, because evolution is prolific in every kind of homoplasy, i.e. convergences, parallelisms and reversals (Amorim 1997). In the reanalyses of the dataset of Rauhut (2003) conducted here, herrerasaurids do not appear as a clade in the strict consensus tree, possibly due to the non-inclusion of characters considered synapomorphic for H. ischigualastensis and S. pricei (Langer & Benton 2006).

Among the characters presented above, Chindesaurus bryansmalli and Caseosaurus crosbyensis only definitely possess the reduction of the brevis fossa, but the materials attributed to these taxa are too incomplete to yield accurate conclusions as to their phylogenetic position. The shortening of the caudal dorsal vertebrae reported for Chindesaurus bryansmalli (Novas 1997) was contested by Nesbitt et al. (2007). According to Irmis et al. (2007a), the presence of two sacral vertebrae unites H. ischigualastensis and Chindesaurus bryansmalli. However, as stated by Sereno (2007), H. ischigualastensis bears three sacral vertebrae (two sacrals plus one dorsosacral), as does the basal dinosauriform Silesaurus opolensis (Dzik & Sulej 2007). In S. pricei, there are also three sacrals, although this differs from H. ischigualastensis by the incorporation of a caudal element, rather than a dorsal element, into the sacrum. Therefore the presence of two sacral vertebrae is not a synapomorphy of Herrerasauridae, and Chindesaurus bryansmalli and Caseosaurus crosbyensis cannot be unequivocally assigned to Herrerasauridae on the basis of available evidence.