[Palaeontology, Vol. 52, Part 4, 2009, pp. 681–688] THE AFFINITIES OF THE ENIGMATIC DINOSAUR ESHANOSAURUS DEGUCHIIANUS FROM THE EARLY JURASSIC OF YUNNAN PROVINCE, PEOPLE’S REPUBLIC OF CHINA by PAUL M. BARRETT Department of Palaeontology, The Natural History Museum, Cromwell Road, SW7 5BD London, UK; e-mail: p.barrett@nhm.ac.uk Typescript received 31 July 2008; accepted in revised form 31 October 2008 Abstract: Eshanosaurus deguchiianus is based on a single left dentary from the Lower Lufeng Formation (Lower Jurassic) of Yunnan Province, China. It was originally identified as the earliest known member of Therizinosauroidea (Theropoda: Coelurosauria), a conclusion that results in a significant downward range extension for this clade (>65 million years) and for many other major lineages within Coelurosauria. However, this interpretation has been questioned and several authors have proposed that the anatomical features used to refer Eshanosaurus to Therizinosauroidea are more consistent with attribution to a basal sauropodomorph dinosaur. Detailed consideration of the holotype specimen suggests that several features of the dentary and dentition G host lineages inferred from phylogenetic analyses of theropod dinosaurs suggest that Coelurosauria, the clade that includes birds and their closest relatives, appeared and diversified during the Middle Jurassic (Sereno 1999; Holtz 2000; Rauhut 2003, 2005; Holtz et al. 2004). Direct evidence for the timing of this event is also known, as definitive coelurosaur specimens are now recognised from this interval (Proceratosaurus from the Bathonian of England; Holtz 2000; Rauhut 2003; Holtz et al. 2004; Rauhut and Milner 2008). Reports of pre-Middle Jurassic coelurosaurs, including the Late Triassic genera Protoavis (originally described as the earliest known bird: Chatterjee 1991) and Shuvosaurus (proposed as the earliest known ornithomimosaur: Chatterjee 1993) have been regarded with scepticism (e.g. Rauhut 1997; Witmer 2002; Nesbitt et al. 2007). Recent discoveries and new phylogenetic analyses have demonstrated that Shuvosaurus is a non-dinosaurian archosaur (e.g. Nesbitt 2007). In addition, it is generally accepted that most of the material referred to Protoavis pertains to an indeterminate archosaur; however, there is still a possibility that some of the individual elements assigned to this species are of coelurosaurian (if ª The Palaeontological Association exclude Eshanosaurus from Sauropodomorpha and support its inclusion within Therizinosauroidea. If accepted as an Early Jurassic coelurosaur, Eshanosaurus has important implications for understanding the timing and tempo of early theropod diversification. Moreover, its provenance also suggests that substantial portions of the coelurosaur fossil record may be missing or unsampled. However, the Early Jurassic age of Eshanosaurus requires confirmation if this taxon is to be fully incorporated into broader evolutionary studies. Key words: Therizinosauroidea, Sauropodomorpha, Coelurosauria, systematics. not avian) origin and this taxon requires further investigation (Witmer 2002; Nesbitt et al. 2007). Another potential pre-Middle Jurassic coelurosaur, Eshanosaurus deguchiianus, was described on the basis of an almost complete (but damaged) left dentary from the Lower Lufeng Formation of Yunnan Province, China (IVPP V11579: Zhao and Xu 1998; Xu et al. 2001). The Lower Lufeng Formation was formerly regarded as Late Triassic in age (Young 1951), but subsequent authors now assign it to the Early Jurassic (e.g. Luo and Wu 1994). Several characters present in Eshanosaurus, including the presence of a broad, flat shelf of bone lateral to the tooth row, a high tooth count and various details of the tooth morphology, suggest that this genus may be referable to Therizinosauroidea (Zhao and Xu 1998; Xu et al. 2001). Therizinosauroidea is a clade of unusual, herbivorous coelurosaurian theropods that is known primarily from the Late Cretaceous of North America and eastern Asia (Clark et al. 2004). Although the recent discoveries of Beipiaosaurus (late Barremian, China: Xu et al. 1999) and Falcarius (Barremian, USA: Kirkland et al. 2005) extend doi: 10.1111/j.1475-4983.2009.00887.x 681 682 PALAEONTOLOGY, VOLUME 52 the temporal range of definitive therizinosauroids into the Early Cretaceous, Eshanosaurus remains the only Jurassic representative of the group, and therefore implies the presence of an extensive ghost lineage stretching from the Sinemurian to the Barremian (over 65 mya: Gradstein et al. 2004). If correctly identified, Eshanosaurus pulls the origin of Therizinosauroidea into the earliest Jurassic or Late Triassic. This would also result in major range extensions to many other coelurosaur and tetanuran lineages, most of which have first occurrences in the Bajocian or Bathonian (Weishampel et al. 2004). Consequences of this would include a reduction in the stratigraphical congruence of all current theropod phylogenies (Wills et al. 2008) and it would also imply that theropod diversification occurred much earlier than would be expected on the basis of other data. The early occurrence of Eshanosaurus has led some authors to question its therizinosauroid affinities, particularly as Eshanosaurus possesses a number of features that are more derived than those present in some Early Cretaceous representatives of the clade (Kirkland et al. 2005). In addition, it has been suggested that the morphology of the type specimen is more consistent with referral to Sauropodomorpha (Lamanna in Kirkland and Wolfe 2001; Rauhut 2003). Here, I review the evidence that has been used to support these conflicting opinions in order to determine if Eshanosaurus is a basal sauropodomorph or a therizinosauroid and to assess the implications of its systematic position and provenance. Institutional abbreviations. AMNH, American Museum of Natural History, New York; BP, Bernard Price Institute for Palaeontological Research, Johannesburg; IVPP, Institute of Vertebrate Paleontology and Paleoanthropology, Beijing; NHM, The Natural History Museum, London; PST, Paleontological and Stratigraphic Section, Geological Institute, Mongolian Academy of Sciences, Ulan Bataar; SAM, Iziko South African Museum, Cape Town. AFFINITIES OF ESHANOSAURUS Eshanosaurus was referred to Therizinosauroidea, and excluded from both Sauropodomorpha and Ornithischia, following consideration of eleven features (Xu et al. 2001). These features are: (1) rostrally positioned teeth larger than more caudally positioned teeth; (2) small size and large number of teeth; (3) recurvature of the dentary teeth; (4) denticles almost perpendicular to the mesial and distal margins of the tooth crown; (5) tooth crowns asymmetrical in mesial or distal view; (6) presence of a constriction between the tooth root and crown; (7) tooth root slightly wider than the crown mesiodistally; (8) root with a sub-circular cross-section; (9) presence of inter- dental plates; (10) presence of a broad, flat shelf lateral to the tooth row; and (11) ventrally deflected rostral end of the dentary. However, Rauhut (2003) noted that several of these features are present in sauropodomorphs (characters 3, 6 and 8–11), one is a dinosaur symplesiomorphy (character 4), two might be under ontogenetic control (characters 1 and 2) and the remaining dental characters might depend on the position in the tooth row that was sampled (characters 5 and 7). Xu et al. (2001) also noted that some of these features were shared by other dinosaur clades, but considered that the combination of characters in Eshanosaurus was characteristic of therizinosauroids and differed from those present in other dinosaurs. Each of these characters is discussed in turn, below. As there is general agreement that Eshanosaurus is not an ornithischian, those features used to distinguish sauropodomorphs from therizinosauroids are emphasised. 1. Therizinosaurs can be distinguished from all other theropods by the reduction in tooth size (both tooth length and width) that occurs from the rostral to the middle part of the dentary tooth row (Russell and Dong 1993; Clark et al. 1994, 2004; Xu et al. 2001). Rauhut (2003) suggested that this character was unreliable, stating that a similar condition may be present in juvenile basal sauropodomorphs. However, although a continuous caudad reduction in dentary tooth size does occur in adult eusauropod sauropodomorphs (Upchurch et al. 2004) this condition is either absent or cannot be determined in basal sauropodomorph taxa. In Massospondylus (e.g. SAM-PK-K1314, BP ⁄ 1 ⁄ 4934: Sues et al. 2004; Barrett and Yates 2006), Plateosaurus (AMNH 6810: see Galton 1984), Pantydraco (NHM P24: Yates 2003) and Jingshanosaurus (Zhang and Yang 1994), the rostralmost dentary teeth (in the first two or three alveoli) are smaller, rather than larger, than those in more caudal tooth positions, so that tooth size initially increases caudally, in contrast to the condition in therizinosauroids. Unfortunately, this character cannot be determined in any juvenile specimens of basal sauropodomorphs as the dentary teeth are often obscured, as occurs in Mussaurus (Bonaparte and Vince 1979; Pol and Powell 2007) and Massospondylus (e.g. BP ⁄ 1 ⁄ 4376: Gow et al. 1990; Sues et al. 2004). Consequently, not only is this feature synapomorphic for Therizinosauroidea within Theropoda, but it can also be used to differentiate therizinosauroid and sauropodomorph dentary dentitions, at least on the basis of current data (contra Rauhut 2003). 2. Rauhut (2003) posited that the high number of dentary teeth present in Eshanosaurus was not a useful character, as tooth numbers are generally variable among reptiles. However, no known sauropodomorph possesses more than 28 dentary teeth (i.e. Plateosaurus: see Galton and Upchurch 2004) whereas Eshanosaurus possesses a BARRETT: THE ENIGMATIC DINOSAUR ESHANOSAURUS minimum of 34 tooth positions and may have had 37 teeth or more (Xu et al. 2001; Clark et al. 2004; the rostralmost part of the dentary is missing in IVPP V11579). Sauropodomorphs gain additional tooth positions during growth (Galton and Upchurch 2004) and the highest tooth counts occur in the largest individuals of Plateosaurus (which have skull lengths of 25 cm or more: e.g. AMNH 6810). In contrast, the holotype dentary of Eshanosaurus pertains to a much smaller animal; the high tooth count in this species is inconsistent with its interpretation as a juvenile basal sauropodomorph as a similarly-sized ‘prosauropod’ would be expected to have only 15–20 dentary teeth (based on comparisons with Massospondylus: SAM-PK-K1314, BP ⁄ 1 ⁄ 4779). Primitive therizinosauroids, such as Beipiaosaurus and Alxasaurus, have similar numbers of dentary teeth to Eshanosaurus (Russell and Dong 1993; Xu et al. 1999; Clark et al. 2004). Several other theropod clades have members that possess >30 dentary teeth (e.g. basal ornithomimosaurs, troodontids and spinosaurids), but in each of these cases the tooth morphology is clearly distinct from that of Eshanosaurus (Currie 1987; Pérez-Moreno et al. 1994; Charig and Milner 1997). Thus, the combination of small size and high tooth count does provide a useful character combination that allows the dentaries of therizinosauroids to be distinguished from those of basal sauropodomorphs: on this basis Eshanosaurus should be excluded from the latter group. 3. Xu et al. (2001) and Rauhut (2003) pointed out that dentary tooth crowns in both theropods (including therizinosauroids) and basal sauropodomorphs may exhibit slight recurvature; as a result, this feature is ambiguous and does not help to determine the relationships of Eshanosaurus. 4. Rauhut (2003) noted that possession of marginal tooth denticles that extend perpendicular to the crown is a dinosaur symplesiomorphy, which is an assumption that is prevalent among dinosaur workers (see character codings in the following phylogenetic analyses for several examples: Butler et al. 2008; Upchurch et al. 2007; Yates 2007). However, this may not be the case: no sauropodomorphs or ornithischians exhibit this character state and basal dinosaurs that lie outside of the ornithischian ⁄ saurischian dichotomy are currently unknown (Herrerasaurus and Eoraptor have sometimes been considered to occupy this position but are now generally regarded as primitive saurischians: Langer and Benton 2006). Moreover, the situation in dinosaur outgroups is unclear, as teeth are not preserved in the majority of nondinosaurian dinosauromorph specimens (Lagerpeton and Dromomeron, Sereno and Arcucci 1994; Irmis et al. 2007) or have very weak denticles whose morphology is difficult to interpret (Silesaurus, Dzik 2003). The assumption that perpendicular denticles represent the primitive condition is, therefore, based primarily on the 683 presence of teeth bearing perpendicular serrations in a referred specimen of the basal dinosauriform Marasuchus (Bonaparte 1975) and in herrerasaurids (Sereno and Novas 1994). However, given the lack of information on many non-dinosaurian dinosauromorphs and the diversity of tooth morphologies present in the taxa close to the base of Dinosauria, it could be argued that the distribution of this character is ambiguous in basal dinosaurs. Resolution of this problem will be dependent on the discovery of additional basal dinosaur and non-dinosaurian dinosauriform cranial material. On the basis of current data, the possession of tooth serrations oriented perpendicular to the crown should probably be regarded as either a saurischian or theropod symplesiomorphy. If the former, this character cannot be used to resolve the position of Eshanosaurus. The near-perpendicular marginal denticles present in Eshanosaurus appear to be autapomorphic for the taxon whether it is referable to Sauropodomorpha or Therizinosauroidea (Xu et al. 2001), where members of both clades generally have denticles that are oriented apically (e.g. Clark et al. 2004; Galton and Upchurch 2004). However, given the prevalence of perpendicular denticles in theropods, and their absence from all sauropodomorphs, it could be argued that this feature provides additional support for referral of Eshanosaurus to Theropoda. 5. Ornithischian tooth crowns possess a basal swelling (‘cingulum’) that is present on the lingual surface of maxillary teeth and the labial surface of maxillary teeth, so that in mesial ⁄ distal view the crowns are asymmetrical (Galton 1986). In contrast, the teeth of basal sauropodomorphs lack this structure and have symmetrical crowns. Eshanosaurus teeth possess a swelling at the base of the crown on the lingual surface (Xu et al. 2001, fig. 2F) and therefore differ from those of basal sauropodomorphs. However, it should be noted that sauropod teeth are also asymmetrical in mesial ⁄ distal view due to the development of the ‘labial concavity’ and can also be considered as similar to therizinosaur teeth in other respects (e.g. in the retention of marginal denticles in some non-neosauropods, such as Shunosaurus and Omeisaurus; see Upchurch and Barrett 2000). As a result, this character cannot be used definitively to exclude Eshanosaurus from Sauropodomorpha. Rauhut (2003) was concerned that assessment of this character would depend on the tooth position sampled in the dentary tooth row; however, this is not the case, as the presence ⁄ absence of the basal swelling is consistent along the entire dentary tooth row in both sauropodomorph and ornithischian dinosaurs (Galton 1986; pers. obs.). 6. As originally worded (Xu et al. 2001) this character cannot be used to distinguish between the competing referrals of Eshanosaurus: both therizinosauroids and basal sauropodomorphs possess a constriction of the tooth 684 PALAEONTOLOGY, VOLUME 52 crown adjacent to the root (Rauhut 2003). Although Xu et al. (2001) note that the constriction seen in therizinosauroids differs from that in sauropodomorphs, this difference is actually created by the morphology described by their character 7 (mesiodistal expansion of the tooth root). Characters 6 and 7 essentially describe the same feature and character 6 should be regarded as redundant. 7. In Eshanosaurus the tooth root constricts immediately below the crown and then expands mesiodistally so that the width of the root equals or exceeds that of the crown (IVPP V11579: Xu et al. 2001). This combination of features does not occur in any known ornithischian or sauropodomorph (pers. obs.), but does occur in some theropods (e.g. troodontids, basal ornithomimosaurs), including most therizinosauroids (Clark et al. 1994, 2004; Xu et al. 1999; Zhang et al. 2001; Kirkland et al. 2005). Although Rauhut (2003) has suggested that this feature may vary along the dentary tooth row no evidence was presented in support of this statement. Indeed, this feature does not appear to vary along the dentary tooth rows of either sauropodomorphs or ornithischians (pers. obs.). 8. As noted by Xu et al. (2001) and Rauhut (2003), both sauropodomorphs and therizinosaurs possess tooth roots with circular cross-sections, so this feature does not help to distinguish between isolated teeth from these two clades. 9. Interdental plates are present in the majority of theropods and basal sauropodomorphs and cannot be used as a reliable character for distinguishing members of these two clades (Xu et al. 2001; Rauhut 2003). Inclusion of this character does indicate that Eshanosaurus cannot be referred to Ornithischia, however, as all known members of this clade lack interdental plates. 10. Both basal sauropodomorphs and therizinosauroids possess a ridge on the lateral surface of the dentary that forms the ventral margin of a buccal emargination (Paul 1984; Barrett and Upchurch 2007; Text-fig. 1). As a result, it was suggested that this character could not be used to help determine the affinities of Eshanosaurus (Rauhut 2003). However, the morphology of this ridge (and of the structures adjacent to it) differs markedly in basal sauropodomorphs and therizinosauroids, allowing the dentar- ls A nf lr ?emf d B nf lr C lr ls emf Left mandibles of A, Eshanosaurus (IVPP V11579; reproduced and modified from Xu et al. 2001: ª2001 The Society of Vertebrate Palaeontology. Reprinted with the permission of the Society of Vertebrate Palaeontology); B, the basal sauropodomorph Massospondylus (SAM-PK-K1314); and C, the therizinosauroid Erlikosaurus (PST 100 ⁄ 111). Note that the lateral ridge of Massospondylus is so weak that it is difficult to distinguish in lateral view. It is essentially a smooth change in the orientation of the slope of the lateral surface. In contrast, the lateral ridges of Eshanosaurus and Erlikosaurus are well defined, forming an abrupt, angular change in slope, and clearly demarcate a lateral shelf. The lateral shelf of Erlikosaurus is only visible rostrally in lateral view. The caudal part of this structure faces almost directly dorsally and so is not visible laterally. Abbreviations: d, depression; emf, external mandibular fenestra; lr, lateral ridge; ls, lateral shelf; nf, nutrient foramen. Scale bars equal 20 mm (A, B) and 50 mm (C). TEXT-FIG. 1. BARRETT: THE ENIGMATIC DINOSAUR ESHANOSAURUS ies of these animals to be distinguished from each other. In basal sauropodomorphs, the ridge is a low, rounded eminence that merges into the main body of the dentary ventrally without any distinct break in slope. This ridge extends rostrally for a short distance, terminating within the caudal half of the dentary. It delimits a very shallow, transversely narrow buccal emargination that faces mainly laterally; the tooth row is not strongly inset with respect to the lateral margin of the buccal emargination (Barrett and Upchurch 2007). In contrast, therizinosauroids possess an exceptionally prominent and clearly defined lateral ridge that forms the ventral and lateral boundary of a dorsolaterally oriented and transversely expanded shelf of bone (‘lateral shelf’; Text-fig. 1): there is a clear and sharp change in angle between the lateral shelf and the remainder of the lateral surface of the dentary. The shelf faces dorsolaterally in its rostral part and dorsally in its caudal part, and the tooth row is strongly inset from the lateral margin of the dentary in dorsal view (e.g. Alxasaurus – Russell and Dong 1993; Erlikosaurus – Clark et al. 1994; Beipiaosaurus – Xu et al. 1999). Moreover, the dentary ridge extends into the rostral half of the dentary in therizinosauroids. The morphology of Eshanosaurus is almost identical to that of the therizinosauroids Alxasaurus, Beipiaosaurus and Erlikosaurus and differs considerably from that present in any basal sauropodomorph, thereby providing strong support for referral of Eshanosaurus to Therizinosauroidea (Xu et al. 2001). 11. Rauhut (2003) correctly points out that a ventrally deflected rostral end of the dentary is present not only in therizinosaurs, but also in some basal sauropodomorphs (e.g. Mussaurus: Pol and Powell 2007). Consequently, this feature cannot be used to support the therizinosauroid affinities of Eshanosaurus to the exclusion of other taxa. It has also been suggested that Eshanosaurus cannot be referred to Therizinosauroidea as its dentary teeth bear a distinct, apicobasally extending ridge on the lingual surface: this feature was stated to be ‘present on at least some prosauropod teeth, but … unknown in therizinosaurid teeth’ (Lamanna in Kirkland and Wolfe 2001, p. 412). However, although a lingual ridge is present in eusauropods, it is completely absent in basal sauropodomorphs (Galton and Upchurch 2004; Upchurch et al. 2004; Barrett and Upchurch 2007). Moreover, a similar ridge is present on the dentary teeth of the North American therizinosauroid Falcarius (Kirkland et al. 2005; L. E. Zanno, pers. comm. 2008), demonstrating that this feature is present in at least some members of the clade. Consequently, the presence of the lingual ridge could be regarded as evidence for the referral of Eshanosaurus to Therizinosauroidea and against its assignment to a basal sauropodomorph (contra Lamanna in Kirkland and Wolfe 2001). 685 Finally, Zhao and Xu (1998) proposed one other character in support of the therizinosauroid affinities of Eshanosaurus, which was mentioned but not discussed by Xu et al. (2001): the presence of a small number of large nutrient foramina on the dorsal surface of the lateral shelf. However, although such foramina are rare among theropods, they are common in ornithischians and basal sauropodomorphs (Galton 1973; Paul 1984). As a result, this character is equivocal and could be used to support relationships with either sauropodomorphs or therizinosauroids. In summary, many of the aforementioned characters are insufficient to distinguish the dentaries and dentary dentitions of basal sauropodomorphs and therizinosaurs (Rauhut 2003). This is perhaps unsurprising as Paul (1984) noted a large number of similarities between the skulls of these groups, which he proposed were evidence of a close phylogenetic relationship. However, subsequent work has clearly demonstrated that therizinosauroids are deeply nested within Theropoda and these features are now regarded as homoplasies (Clark et al. 2004). Nevertheless, at least three features – the caudad reduction in dentary tooth size, the high dentary tooth count per unit length of jaw and the detailed morphology of the buccal emargination and ridge – provide positive evidence for referral of Eshanosaurus to Therizinosauroidea. Several other features, such as the morphology of the tooth roots, are also consistent with this interpretation. In contrast, Eshanosaurus possesses no unequivocal sauropodomorph synapomorphies: indeed, several of the character states present in the specimen argue against referral to this clade. On the basis of current evidence, referral of Eshanosaurus to Therizinosauroidea is more strongly supported than the alternative hypothesis of referral to Sauropodomorpha. DISCUSSION If accepted as a member of Therizinosauroidea, the provenance of Eshanosaurus implies that coelurosaur diversification occurred much earlier than is generally acknowledged and also indicates that numerous coelurosaur lineages should be represented in the Early and early Middle Jurassic (Xu et al. 2001; Rauhut 2003; Wills et al. 2008). These conclusions conflict with current knowledge of theropod phylogeny and the remainder of the known fossil record of coelurosaurs, which otherwise indicates that Coelurosauria began its radiation sometime in the late Middle Jurassic (Sereno 1999; Holtz 2000; Rauhut 2003, 2005; Holtz et al. 2004). No other Early Jurassic coelurosaurs are known and Eshanosaurus implies that a significant portion of the history of this clade may still remain to be discovered. However, the purported Early Jurassic 686 PALAEONTOLOGY, VOLUME 52 age of the specimen remains a cause for concern. Some of the features present in Eshanosaurus (e.g. the lateral shelf and the ventrally deflected rostral end of the dentary) indicate that this taxon is more derived than the Early Cretaceous therizinosauroid Falcarius (Kirkland et al. 2005). Although this may simply indicate that other basal therizinosaur fossils still await discovery in the Jurassic, the length of time separating Eshanosaurus and Falcarius is anomalous (Kirkland et al. 2005). It may be noteworthy that the type area for Eshanosaurus, the Dianzhong Basin, contains a thick Mesozoic sequence that includes not only basal Jurassic sediments (referred to as the Lower Lufeng Formation or the Fengjiahe Formation by different authors: Ye 1975; Xu et al. 2001) but also Middle Jurassic and Early Cretaceous deposits (Bureau of Geology and Mineral Resources of Yunnan Province 1990). As the precise age of Eshanosaurus has profound implications for understanding the timing of the coelurosaur radiation, additional collecting and stratigraphical work is needed at the type locality in order to confirm or refute its Early Jurassic age. In the meantime, the potential importance of Eshanosaurus should be assessed critically and its unprecedented stratigraphical position should not be used as a criterion for excluding this taxon from analyses of coelurosaurian and theropod evolution. Acknowledgements. Many thanks to Xu Xing for granting access to material in IVPP and to many other colleagues in China for their help and hospitality, including Wang Xiao-Lin, Wang Yuan, Zhang Jiang-Jong and Zhou Zhonghe. Richard Butler, Lindsay Zanno, Xu Xing and Jim Clark are also thanked for discussion and their careful reviews of the manuscript. Thanks also to those curators in the other institutions mentioned: Sheena Kaal (SAM), Michael Raath (BP) and Carl Mehling (AMNH). Phil Crabb (NHM Image Resources) provided the photograph in Text-figure 1B; Angela Milner supplied the photograph of Erlikosaurus in Text-figure 1C, which was taken with the kind permission of Perle Altangerel. Travel to Beijing was funded by the Royal Society of London, the Hodson Award of the Palaeontological Association and the Palaeontological Innovation Fund of the Natural History Museum. REFERENCES B A R R E T T , P. M. and U P C H U R C H , P. 2007. The evolution of feeding mechanisms in early sauropodomorph dinosaurs. 91–112. In B A R R E T T , P. M. and B A T T E N , D. J. (eds). Evolution and palaeobiology of early sauropodomorph dinosaurs. Special Papers in Palaeontology, 77, 289 pp. —— and Y A T E S , A. M. 2006. New information on the palate and lower jaw of Massospondylus (Dinosauria: Sauropodomorpha). Palaeontologia africana, 41, 123–130. B O N A P A R T E , J. F. 1975. Nuevos materials de Lagosuchus talampayensis Romer (Thecodontia – Pseudosuchia) y su significado en el origen de los Saurischia. Chañarense inferior, Triásico medio de Argentina. Acta Geologica Lilloana, 13, 1–90. —— and V I N C E , M. 1979. El hallazgo del primer nido de dinosaurios Triasicos, (Saurischia, Prosauropoda), Triasico Superior de Patagonia, Argentina. Ameghiniana, 16, 173–182. BUREAU OF GEOLOGY AND MINERAL RESOURCES O F Y U N N A N P R O V I N C E 1990. Regional geology of Yunnan Province. Geological Publishing House, Beijing, x + 728 pp. [In Chinese with English summary]. B U T L E R , R. J., U P C H U R C H , P. and N O R M A N , D. B. 2008. The phylogeny of the ornithischian dinosaurs. Journal of Systematic Palaeontology, 6, 1–40. C H A R I G , A. J. and M I L N E R , A. C. 1997. Baryonyx walkeri, a fish-eating dinosaur from the Wealden of Surrey. Bulletin of the Natural History Museum, London (Geology Series), 53, 11–70. C H A T T E R J E E , S. 1991. Cranial anatomy and relationships of a new Triassic bird from Texas. Philosophical Transactions of the Royal Society of London, Series B (Biological Sciences), 332, 277–342. —— 1993. Shuvosaurus, a new theropod. National Geographic Research and Exploration, 9, 284–285. C L A R K , J. M., M A R Y A N S K A , T. and B A R S B O L D , R. 2004. Therizinosauroidea. 151–164. In W E I S H A M P E L , D. B., D O D S O N , P. and O S M Ó L S K A , H. (eds). The Dinosauria, 2nd edn. University of California Press, Berkeley, xviii + 861 pp. —— P E R L E A. and N O R E L L , M. A. 1994. The skull of Erlicosaurus andrewsi, a Late Cretaceous ‘segnosaur’ (Theropoda: Therizinosauridae) from Mongolia. American Museum Novitates, 3115, 1–39. C U R R I E , P. J. 1987. Bird-like characteristics of the jaws and teeth of troodontid theropods (Dinosauria, Saurischia). Journal of Vertebrate Paleontology, 7, 72–81. D Z I K , J. 2003. A beaked herbivorous archosaur with dinosaur affinities from the early Late Triassic of Poland. Journal of Vertebrate Paleontology, 23, 556–574. G A L T O N , P. M. 1973. The cheeks of ornithischian dinosaurs. Lethaia, 6, 67–89. —— 1984. Cranial anatomy of the prosauropod dinosaur Plateosaurus, from the Knollenmergel (Middle Keuper, Upper Triassic) of Germany. I. Two complete skulls from Trossingen ⁄ Württ. with comments on the diet. Geologica et Palaeontologica, 18, 139–171. —— 1986. Herbivorous adaptations of Late Triassic and Early Jurassic dinosaurs. 203–221. In P A D I A N , K. (ed.). The beginning of the age of the dinosaurs. Cambridge University Press, Cambridge, 378 pp. —— and U P C H U R C H , P. 2004. Prosauropoda. 232–258. In W E I S H A M P E L , D. B., D O D S O N , P. and O S M Ó L S K A , H. (eds). The Dinosauria, 2nd edn. University of California Press, Berkeley, xviii + 861 pp. G O W , C. E., K I T C H I N G , J. W. and R A A T H , M. A. 1990. Skulls of the prosauropod dinosaur Massospondylus carinatus Owen in the collections of the Bernard Price Institute for Palaeontological Research. Palaeontologia africana, 27, 45–58. G R A D S T E I N , F. M., O G G , J. G. and S M I T H , A. G. (eds) 2004. A geologic timescale 2004. Cambridge University Press, Cambridge, 589 pp. BARRETT: THE ENIGMATIC DINOSAUR ESHANOSAURUS H O L T Z , T. R. 2000. A new phylogeny of the carnivorous dinosaurs. Gaia, 15, 5–61. —— M O L N A R , R. E. and C U R R I E , P. J. 2004. Basal Tetanurae. 71–110. In W E I S H A M P E L , D. B., D O D S O N , P. and O S M Ó L S K A , H. (eds). The Dinosauria, 2nd edn. University of California Press, Berkeley, xviii + 861 pp. I R M I S , R. B., N E S B I T T , S. J., P A D I A N , K., S M I T H , N. D., T U R N E R , A. H., W O O D Y , D. and D O W N S , A. 2007. A Late Triassic dinosauromorph assemblage from New Mexico and the rise of dinosaurs. Science, 317, 358–361. K I R K L A N D , J. I. and W O L F E , D. G. 2001. First definitive therizinosaurid dinosaur (Dinosauria: Theropoda) from North America. Journal of Vertebrate Paleontology, 21, 410–414. —— Z A N N O , L. E., S A M P S O N , S. D., C L A R K , J. M. and D E B L I E U X , D. D. 2005. A primitive therizinosauroid dinosaur from the Early Cretaceous of Utah. Nature, 435, 84–85. L A N G E R , M. C. and B E N T O N , M. J. 2006. Early dinosaurs: a phylogenetic study. Journal of Systematic Palaeontology, 4, 309–358. L U O Z H E - X I and W U X I A O - C H U N 1994. The small tetrapods of the Lower Lufeng Formation, Yunnan, China. 251–270. In F R A S E R , N. C. and S U E S , H.-D. (eds). In the shadow of the dinosaurs: early Mesozoic tetrapods. Cambridge University Press, Cambridge, 435 pp. N E S B I T T , S. J. 2007. The anatomy of Effigia okeeffeae (Archosauria, Suchia), theropod-like convergence, and the distribution of related taxa. Bulletin of the American Museum of Natural History, 302, 1–84. —— I R M I S , R. B. and P A R K E R , W. G. 2007. A critical re-evaluation of the Late Triassic dinosaur taxa of North America. Journal of Systematic Palaeontology, 5, 209–243. P A U L , G. S. 1984. The segnosaurian dinosaurs: relics of the prosauropod-ornithischian transition? Journal of Vertebrate Paleontology, 4, 507–515. P É R E Z - M O R E N O , B. P., S A N Z , J. L., B U S C A L I O N I , A. D., M O R A T A L L A , J. J., O R T E G A , F. and R A S S K I N G U T M A N , D. 1994. A unique multi-toothed ornithomimosaur dinosaur from the Lower Cretaceous of Spain. Nature, 370, 363–367. P O L , D. and P O W E L L , J. E. 2007. Skull anatomy of Mussaurus patagonicus (Dinosauria: Sauropodomorpha) from the Late Triassic of Patagonia. Historical Biology, 19, 125–144. R A U H U T , O. W. M. 1997. Zur Schädelanatomie von Shuvosaurus inexpectatus (Dinosauria; Theropoda). Terra Nostra, 1997 (7), 17–21. —— 2003. The interrelationships and evolution of basal theropod dinosaurs. Special Papers in Palaeontology, 69, 1–213. —— 2005. Osteology and relationships of a new theropod dinosaur from the Middle Jurassic of Patagonia. Palaeontology, 48, 87–110. —— and M I L N E R , A. C. 2008. Cranial anatomy and systematic position of the Middle Jurassic theropod dinosaur Proceratosaurus from England. Journal of Vertebrate Paleontology, 28 (Suppl. 3), 130A. R U S S E L L , D. A. and D O N G Z H I - M I N G 1993. The affinities of a new theropod from the Alxa Desert, Inner Mongolia, People’s Republic of China. Canadian Journal of Earth Sciences, 30, 2107–2127. 687 S E R E N O , P. C. 1999. The evolution of dinosaurs. Science, 284, 1237–1247. —— and A R C U C C I , A. B. 1994. Dinosaurian precursors from the Middle Triassic of Argentina: Lagerpeton chanarensis. Journal of Vertebrate Paleontology, 13, 385–399. —— and N O V A S , F. 1994. The skull and neck of the basal theropod Herrerasaurus ischigualastensis. Journal of Vertebrate Paleontology, 13, 451–476. S U E S , H.-D., R E I S Z , R. R., H I N I C , S. and R A A T H , M. A. 2004. On the skull of Massospondylus carinatus Owen, 1854 (Dinosauria: Sauropodomorpha) from the Stormberg Group (Lower Jurassic) of South Africa. Annals of the Carnegie Museum, 73, 239–257. U P C H U R C H , P. and B A R R E T T , P. M. 2000. The evolution of sauropod feeding mechanisms. 79–122. In S U E S , H.-D. (ed.). Evolution of herbivory in terrestrial vertebrates: perspectives from the fossil record. Cambridge University Press, Cambridge, 256 pp. —— —— and D O D S O N , P. 2004. Sauropoda. 259–322. In W E I S H A M P E L , D. B., D O D S O N , P. and O S M Ó L S K A , H. (eds). The Dinosauria, 2nd edn. University of California Press, Berkeley, xviii + 861 pp. —— —— and G A L T O N , P. M. 2007. A phylogenetic analysis of basal sauropodomorph relationships: implications for the origin of sauropod dinosaurs. 57–90. In B A R R E T T , P. M. and B A T T E N , D. J. (eds). Evolution and palaeobiology of early sauropodomorph dinosaurs. Special Papers in Palaeontology, 77, 289 pp. W E I S H A M P E L , D. B., B A R R E T T , P. M., C O R I A , R. A., L E L O E U F F , J., X U X I N G , Z H A O X I J I N , S A H N I , A., G O M A N I , E. M. P. and N O T O , C. R. 2004. Dinosaur distribution. 517–606. In W E I S H A M P E L , D. B., D O D S O N , P. and O S M Ó L S K A , H. (eds). The Dinosauria, 2nd edn. University of California Press, Berkeley, 861 pp. W I L L S , M. A., B A R R E T T , P. M. and H E A T H C O T E , J. F. 2008. The stratigraphic congruence of dinosaur phylogenies and the Modified Gap Excess Ration (GER*). Systematic Biology, 57, 891–907. W I T M E R , L. M. 2002. The debate on avian ancestry: phylogeny, function, and fossils. 3–30. In C H I A P P E , L. M. and W I T M E R , L. M. (eds). Mesozoic birds: above the heads of dinosaurs. University of California Press, Berkeley, xii + 520. X U X I N G , T A N G Z H I - L U and W A N G X I A O - L I N 1999. A therizinosauroid dinosaur with integumentary structures from China. Nature, 399, 350–354. —— Z H A O X I - J I N and C L A R K , J. M. 2001. A new therizinosaur from the Lower Jurassic Lower Lufeng Formation of Yunnan, China. Journal of Vertebrate Paleontology, 21, 477–483. Y A T E S , A. M. 2003. A new species of the primitive dinosaur Thecodontosaurus (Saurischia: Sauropodomorpha) and its implications for the systematics of early dinosaurs. Journal of Systematic Palaeontology, 1, 1–42. —— 2007. The first complete skull of the Triassic dinosaur Melanorosaurus Haughton (Sauropodomorpha: Anchisauria). 9–55. In B A R R E T T , P. M. and B A T T E N , D. J. (eds). Evolution and palaeobiology of early sauropodomorph dinosaurs. Special Papers in Palaeontology, 77, 289 pp. 688 PALAEONTOLOGY, VOLUME 52 Y E H I A N G - K U I 1975. Jurassic system. 11–30. In NANJING INSTITUTE OF GEOLOGY AND PALAEONTOLOGY (ed.) Mesozoic redbeds of Yunnan. Geological Publishing House, Beijing, 214 pp. [In Chinese]. Y O U N G C H U N G - C H I E N 1951. The Lufeng saurischian fauna. Palaeontologica Sinica, Series C, 13, 1–96. pls 1–12. Z H A N G X I A O - H O N G , X U XING, Z H A O X I - J I N , S E R E N O , P. C., K U A N G X U E - W E N and T A N LIN 2001. A long-necked therizinosauroid dinosaur from the Upper Cretaceous Iren Dabasu Formation of Nei Mongol, People’s Republic of China. Vertebrata PalAsiatica, 39, 282– 290+ pls 1–3. Z H A N G Y I - H O N G and Y A N G Z H A O - L O N G 1994. A new complete osteology of Prosauropoda in Lufeng Basin, Yunnan, China. Yunnan Publishing House of Science and Technology, Kunming, 100 pp. + 24 pls [In Chinese with English summary]. Z H A O , X I - J I N and X U XING 1998. The oldest coelurosaurian. Nature, 394, 234–235.