Zoological Journal of the Linnean Society
SCELIDOSAURUS HARRISONII OWEN, 1861 FROM THE EARLY JURASSIC OF DORSET, ENGLAND: PART 1. CRANIAL ANATOMY
Journal: Zoological Journal of the Linnean Society
Manuscript ID ZOJ-10-2018-3496.R2 Manuscript Type:ForOriginal Review Article Only monograph < Taxonomy, alpha taxonomy < Taxonomy, Reptilia < Taxa, Dinosauria < Taxa, ontogeny < Growth, vertebrate palaeontology < Keywords: Palaeontology, Jurassic < Palaeontology, English South Coast < Geography, mechanical function < Biomechanics, cranial osteology < Anatomy
The skull of Scelidosaurus harrisonii is described. Scelidosaurus is an early (late Sinemurian) armoured ornithischian dinosaur from Charmouth, Dorset. It is the first largely complete and articulated dinosaur ever discovered. Much of the exterior of the skull has a textured patina of exostotic bone that anchored keratinous scales. A small edentulous beak precedes a row of five premaxillary teeth, and there would have been a minimum of 22 maxillary teeth and 27 dentary teeth in the largest known individuals. Tooth wear is discontinuous along the dentition and jaw action appears to have been tightly constrained. Abstract: There is no physical evidence of a predentary but its existence is inferred from the structure of the dentary. A sclerotic ring was present. A prominent supraorbital brow ridge overhangs the orbital cavity. A crater- shaped osteoderm is attached to the lateral surface of the postorbital. Paired osteodermal horns project dorsally from the occiput, and long bladed styloid bones project obliquely from the posterior of the skull. The nasal chambers are roofed by epivomers that are unique to this taxon. A vertical, conical epipterygoid is attached to the pterygoid. A deep pit on the posterior surface of the quadrate suggests the existence of remnant cranial pneumatism.
Page 1 of 215 Zoological Journal of the Linnean Society
1 2 3 4 SCELIDOSAURUS HARRISONII OWEN, 1861 FROM THE EARLY JURASSIC 5 6 OF DORSET, ENGLAND: PART 1. CRANIAL ANATOMY 7 8 9 10 11 12 DEDICATION 13 14 I wish to offer a personal note of thanks to Ron Croucher (former Head of 15 16 the Palaeontological Preparation-Conservation Laboratory at the Natural 17 History Museum, London) by dedicating this monograph to him. Ron 18 19 committed an enormous amount of time and effort, across several 20 21 decades, to the laboriousFor andReview painstaking Onlypreparation of the lectotype 22 skull and postcranial skeleton of Scelidosaurus (NHMUK R1111). His 23 24 having to watch as numerous researchers subsequently mishandled (and 25 26 damaged) some of these bones, hard-won with such skill and dedication, 27 must have severely tested his patience and goodwill. Ron is indeed a 28 29 wonderfully kind and incredibly patient man. While superintending the 30 31 museum laboratory he could always find time to give advice and 32 assistance to ignorant and cack-handed PhD students who, from time to 33 34 time, ‘messed about’ with fossil specimens in his laboratory … I know this 35 36 all too well, I was one of them. 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 2 of 215
1 2 3 4 EXTENDED ABSTRACT 5 6 Scelidosaurus harrisonii Owen, 1861 is an early (late Sinemurian) 7 armoured ornithischian dinosaur whose remains have, to date, only been 8 recovered from a restricted location on the south coast of Dorset 9 (Charmouth), England. This dinosaur has been known since 1859, but 10 only on the basis of a partial description found in two articles published in 11 12 the early 1860s by Richard Owen. The original material, discovered in 13 1858, comprised the majority of the skull and its associated postcranial 14 skeleton, and represents the first ever more or less complete dinosaur 15 discovered. In addition to the original material, a number of further 16 discoveries have been made at Charmouth; these latter supplement the 17 information that can be gleaned from the original specimen. This article 18 describes the skull of Scelidosaurus. 19 20 The external surface of individual skull bones in ontogenetically relatively 21 mature individualsFor displays Review exostoses, a patina Only of fibrous or granular- 22 textured bone that anchored an external shielding of keratinous scales. 23 There is a small, edentulous rostral beak, behind which is found a row of 24 five heterodont premaxillary teeth. There is a minimum of 22 maxillary 25 teeth and 27 dentary teeth in jaws of the largest well-preserved 26 individuals known to date. Both dentitions (upper and lower) are bowed 27 28 medially and are sinuous longitudinally. Maxillary and dentary crowns are 29 tilted lingually on their roots, trapezoidal in outline and have crenellate 30 (coarsely denticulate) margins. Adjacent crowns of teeth have mesio- 31 distally (anteroposteriorly) expanded bases that overlap slightly and are 32 consequently arranged en echelon. The dentitions are flanked by deep 33 cheek pouches. Tooth abrasion is usually discontinuous along the 34 dentition. In one individual nearly all teeth seem to be fully emerged and 35 there is little evidence of abrasion. There is no physical evidence of a 36 37 predentary, but the presence of this (typically ornithischian) element may 38 be inferred from the structure of the symphyseal region of the dentary. 39 The external narial and antorbital fenestrae are comparatively small, 40 whereas the orbit and temporal fenestrae are large and open. A sclerotic 41 ring was undoubtedly present and supported the eyeball, but it is too 42 poorly preserved to allow it to be reconstructed with accuracy. A 43 prominent supraorbital brow ridge overhangs the orbit. There are three 44 osteoderms: palpebral, middle supraorbital and posterior supraorbital, 45 sutured to the dorsal margin of the orbit. The occiput provides an area for 46 47 attachment of a pair of curved, keratin-sheathed, osteodermal horns. 48 Epistyloid bones project from the ventrolateral region of the braincase; 49 their distal ends flank the anterolateral region of the neck. Rugose facets 50 on either side of the basioccipital are suggested to have provided 51 attachment sites for the epistyloid bones. Internally, the skull has a 52 53 deeply vaulted snout and the nasal chambers are roofed by what are here 54 named epivomer bones that appear to have been sutured to the 55 dorsolateral edges of the vomers. Unusually, among dinosaurs generally, 56 an epipterygoid is preserved attached to the dorsolateral surface of the 57 pterygoid; there is no obvious point of articulation for the epipterygoid 58 against the lateral wall of the braincase. There is also a narrow, slot-like 59 pocket on the medial wall of the quadrate wing of the pterygoid. A deep 60
2 Page 3 of 215 Zoological Journal of the Linnean Society
1 2 3 pit on the posterior surface of the quadrate of an immature specimen is 4 suggestive of the existence of a remnant of cranial pneumatism; this pit 5 becomes occluded in larger, more mature specimens. 6 7 8 9 NOTE CONCERNING THE LECTOTYPE 10 11 In 1955 Bill (William Elgin) Swinton (1900-1994) approved a programme 12 13 of work for Arthur Rixon, focused upon the acetic acid-mediated 14 15 preparation of a split marlstone (=argillaceous limestone) nodule 16 17 (‘dogger’) that had been collected by J.F. Jackson from the foreshore 18 beneath Black Ven, Charmouth and acquired by the Natural History 19 20 Museum (London). The dogger contained an incomplete articulated 21 For Review Only 22 postcranial skeleton of a small individual of Scelidosaurus. Rixon (1949) 23 had experimented with the use of dilute mineral acids to prepare fossil 24 25 vertebrates and with Harry Toombs (Toombs & Rixon 1950) had 26 27 developed the ‘Transfer Method’ for preparing such fossils in the round 28 (Toombs & Rixon 1959). The scelidosaur project resulted in the successful 29 30 extraction of this small dinosaur skeleton (Rixon 1968) and the 31 32 development of further plans to prepare the lectotype skeleton; these 33 plans were however abandoned after the departure of Swinton. 34 35 Alan Charig (1927-1997) was appointed curator in succession to Bill 36 37 Swinton in 1961. He eventually revisited Swinton’s original project, in the 38 39 light of the success of the technique applied to the small scelidosaur 40 skeleton, and so the extraction of the entire skeleton of the lectotype of 41 42 Scelidosaurus harrisonii (NHMUK R1111) began, under the supervision of 43 44 Ron Croucher. The latter skeleton was preserved in a more-or-less 45 contiguous series of substantial marlstone blocks. As the work continued 46 47 Croucher noticed the fragility of some of the acid-extracted bones and 48 49 modified the original preparation protocol by adding a phosphate buffer to 50 the acid solution. This latter modification reduced the extent of acid- 51 52 induced erosion of the fabric of the fossil bones (and their consequent 53 54 extreme fragility) that has seriously jeopardised the earlier extracted 55 bones – in particular the skull. Charig announced his intention to describe 56 57 Scelidosaurus but he had made little progress by the time of his death. 58 59 60
3 Zoological Journal of the Linnean Society Page 4 of 215
1 2 3 REPOSITORY ABBREVIATIONS: BRSMG - Bristol City Museum UK, CAMSM 4 5 – Sedgwick Museum, University of Cambridge UK; DORCM - Dorset 6 7 County Museum, Dorchester, Dorset UK; FMNH – Field Museum of Natural 8 History, Chicago USA; IVPP – Institute of Vertebrate Palaeontology and 9 10 Palaeoanthropology, Beijing, China; LYMPH – The Lyme Regis (Philpot) 11 12 Museum, Dorset UK; NHMUK – Natural History Museum, London British 13 Museum; MNA – Museum of Northern Arizona, Flagstaff USA; SGDS – St 14 15 George Dinosaur Discovery Site, Utah USA; SGWG – Sektion Geologische 16 17 Wissenshaften der Ernst-Moritz-Arndt-Universität Griefswald, Germany; 18 TMM – Texas Memorial Museum (Vertebrate Paleontology Laboratory of 19 20 the University of Texas at Austin), USA; UCMP – University of California 21 For Review Only 22 Museum of Paleontology, Berkeley, USA; ZDM – Zigong Dinosaur 23 Museum, Zigong, China. 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
4 Page 5 of 215 Zoological Journal of the Linnean Society
1 2 3 4 SCELIDOSAURUS: HISTORICAL INTRODUCTION 5 6 7 8 Scelidosaurus: discovered and named 9 10 The first indication of terrestrial fossil reptiles, rather than the ubiquitous 11 12 specialised marine ichthyosaurs and plesiosaurs (Conybeare & De La 13 14 Beche 1821; Conybeare 1822, 1824; Norman 2000a), in the Liassic strata 15 of West Dorset, was a collection of bones obtained by James Harrison 16 17 (1819-1864) who had retired prematurely to Charmouth due to ill-health 18 19 (Lang 1947). Charmouth was then a hamlet near the mouth of the River 20 Char and situated less than a mile along the coast east of Lyme Regis 21 For Review Only 22 (Fig. 1). The fossil specimens collected from the foreshore and coastal 23 24 cliffs (The Spittles-Black Ven – Fig. 1) at Charmouth were loaned to 25 Richard Owen in 1858. Although Owen made brief reference to this 26 27 material as “Scelidosaurus” in print (Owen 1859, 1860, 1862) it was three 28 29 years after receipt of this material that Owen (1861) formally described 30 and illustrated these fossils. The specimens were collectively given the 31 32 Linnaean binomial Scelidosaurus harrisonii (= ‘Harrison’s limbed reptile’). 33 34 Owen’s choice of this name reflects a wish to differentiate this saurian 35 (with limbs adapted for terrestrial locomotion) from the flippered or 36 37 paddle-limbed aquatic reptiles that were more often discovered in the 38 39 area. 40 41 42 43
5 Zoological Journal of the Linnean Society Page 6 of 215
1 2 3 Some of the original specimens named Scelidosaurus may have been 4 5 found in water-rolled boulders or as fragments on the foreshore beneath 6 7 Black Ven, while others were most probably exposed during the quarrying 8 of nodules (or slabs) of the carbonate-rich marlstone and shale that crops 9 10 out as distinctive layers (hence “Lias”) along the cliffs of this coastline. 11 12 Quarrying along the Charmouth-Lyme Regis coast was at this time an 13 important commercial enterprise. The purer limestone could be burnt to 14 15 yield quicklime (calcium oxide) an essential ingredient in the production of 16 17 mortar and stucco and also served as a ‘disinfectant’ (for burials) and as a 18 soil conditioner for farming. The cliffs below Charmouth also became of 19 20 particular importance industrially because a new form of mortar: 21 For Review Only 22 ‘hydraulic cement’ (that would harden underwater) could be manufactured 23 from the Black Ven Marl. This marl combined the ideal proportions of 24 25 limestone-to-clay for hydraulic cement production. The present day 26 27 Charmouth Heritage Coast Centre on the seafront has been built on the 28 foundations of the original mid-19th century hydraulic cement works. 29 30 31 32 33
6 Page 7 of 215 Zoological Journal of the Linnean Society
1 2 3
7 Zoological Journal of the Linnean Society Page 8 of 215
1 2 3 It is unfortunate, from a formal nomenclatural perspective, that the 4 5 material described by Owen comprised not only different-sized individuals 6 7 but also different taxa (Newman 1968; Charig & Newman 1992; Norman 8 2000b, 2001). Owen’s first monograph provided the full Linnaean 9 10 binomial: Scelidosaurus harrisonii Owen, 1861 (by monotypy). 11 12 Considerably later, Richard Lydekker (1888) somewhat idiosyncratically 13 (Charig & Newman 1992: 281) selected as the “type” [more correctly, the 14 15 lectotype] the articulated knee joint (NHMUK OR39496) illustrated by 16 17 Owen (1861: Tab. II, 1-3 – see Fig. 2B); consequently, the remaining 18 specimens became a set of paratypes. 19 20 Acid preparation of Lydekker’s lectotype (Fig. 2B) by Arthur Rixon 21 For Review Only 22 enabled Newman (1968) to demonstrate that the knee joint was 23 24 attributable to an indeterminate theropod dinosaur. Additional paratypes: 25 the femoral shaft (NHMUK OR109560: Fig. 2A) and ungual phalanx 26 27 (NHMUK OR109561: Fig. 2B), were also shown to be referable to other 28 29 similarly indeterminable theropod individuals. In marked contrast, the 30 associated bones of the small skeleton (LYMPH 1998 6.1-6.7: Fig. 3) and 31 32 the well-preserved larger skull (NHMUK R1111: Fig. 4) both pertained to 33 34 an ornithischian dinosaur. 35 36 A long record of publications persistently associated the name 37 38 Scelidosaurus with the distinctive skull and its associated articulated 39 skeleton of an armoured ornithischian dinosaur from Charmouth, rather 40 41 than the other poorly understood portions of the hypodigm. Given the 42 43 long-established precedent, an appeal was made to the ICZN (Charig & 44 Newman 1992) for the replacement of the inappropriate lectotype 45 46 designation (by Lydekker) with that of the entire skull and skeleton 47 48 identified as NHMUK R1111. This recommendation was formally approved 49 (BZN 1994). 50 51 52 53 54 ARMOURED DINOSAURS IN THE FOSSIL RECORD: PROBLEMATIC 55 SYSTEMATICS AND PHYLOGENETICS 56 57 58 59 60
8 Page 9 of 215 Zoological Journal of the Linnean Society
1 2 3 The description of Scelidosaurus came at an opportune moment in relation 4 5 to research on armoured dinosaurs (not differentiating between 6 7 ankylosaurs and stegosaurs because this split was not recognised until the 8 early decades of the 20th century). The first discovery and report of 9 10 dinosaur remains suggesting the presence of bony spines and plates 11 12 (osteoderms) covering the body surface was published by Gideon Mantell 13 (1833). He described and illustrated the articulated neck and anterior 14 15 torso of an animal collected from Wealden (Early Cretaceous) rocks in 16 17 Sussex, that he named Hylaeosaurus armatus. The next discovery 18 (unappreciated for what it was at the time) was of an isolated tooth 19 20 collected from Late Cretaceous rocks in Nebraska, and named 21 For Review Only 22 Palaeoscincus costatus by Joseph Leidy (1856). The formal description of 23 Scelidosaurus by Owen (1861, 1863) was followed by the discovery in 24 25 1865 of another partial skeleton, from Wealden rocks on the Isle of Wight, 26 27 of another armoured dinosaur named Polacanthus foxii (Hulke 1881, 28 Nopcsa 1905). Then in 1874 a much larger plated dinosaur was recovered 29 30 from the Kimmeridgian (Late Jurassic) of Wiltshire: Omosaurus 31 32 (=Dacentrurus) armatus (Owen, 1875). Attention moved in the direction 33 of North America following the discovery of the Kimmeridgian-aged 34 35 remains of Stegosaurus and the establishment of a new ‘Order’ 36 37 (Stegosauria) to recognise the existence of a group of closely related 38 plated/armoured dinosaurs (Marsh 1877). By 1889 Marsh was able to 39 40 expand on his “Order Stegosauria” and incorporated all these genera: 41 42 Hylaeosaurus, Palaeoscincus, Scelidosaurus, Polacanthus, Omosaurus and 43 Stegosaurus, along with the later-described Priconodon, Dystrophaeus, 44 45 Euskelosaurus, Anthodon, Acanthopholis and Struthiosaurus (Marsh 46 47 1889a). His Stegosauria thus included both stegosaurs and ankylosaurs, 48 which today are grouped together within the Thyreophora; but the 49 50 ankylosaurs were neither understood nor described as a distinct grouping 51 52 at this time. Marsh further confused matters for a considerable time by 53 incorporating his Ceratopsidae within the Stegosauria (Marsh 1889b) 54 55 because of the mistaken association of some flattened osteodermal plates 56 57 and spikes with one of his ceratopsians (Ceratops montanus) as well as 58 some nodular osteoderms that he named Nodosaurus (also presumed to 59 60 pertain to a ceratopsian).
9 Zoological Journal of the Linnean Society Page 10 of 215
1 2 3 With the passing of Marsh, other workers started trying to resolve 4 5 the affinities and relationships between these armour-plated dinosaurs. 6 7 Nopcsa (1902) created a new family Acanthopholidae for Hylaeosaurus, 8 Polacanthus and Acanthopholis, and in doing so removed them from the 9 10 family Scelidosauridae (the latter having been created by Huxley in 1870). 11 12 This proved to be an important step in the gradual process of recognition 13 of ankylosaurs and stegosaurs as distinct groups of dinosaurs, and the 14 15 placing of Scelidosaurus as a little apart from these other two groups. 16 17 Better quality material began to be recovered from the North American 18 Late Cretaceous after the turn of the century. Brown (1908) described a 19 20 new taxon Ankylosaurus and erected a new family Ankylosauridae to 21 For Review Only 22 recognise the unusual and unique characters associated with this familial 23 grouping; however, he still retained his new family within Marsh’s Order 24 25 Stegosauria. Gilmore (1914), in what became a classic monograph on 26 27 Stegosaurus, reviewed the classification of plated and armoured dinosaurs 28 and clustered what he referred to as the “ridge-scuted” dinosaurs into a 29 30 grouping that resembles the ankylosaurs as presently conceived. Hennig 31 32 (1915) summarised the then current understanding in his Fossilium 33 Catalogus. Within the “suborder” Stegosauria Hennig recognized three 34 35 families of armoured dinosaurs in a format that seems much more familiar 36 37 to modern eyes: Scelidosauridae for Scelidosaurus and the “Purbeckian” 38 Echinodon. The latter taxon had similar shaped teeth to those seen in 39 40 Scelidosaurus and was reported to have osteoderms “granicones” 41 42 associated with its remains (later disproved: Norman & Barrett 2002); 43 Stegosauridae for animals that are mostly recognised as stegosaurs 44 45 today; and Nodosauridae for animals that are mostly recognised as 46 47 ankylosaurs. 48 49 The constituents (and even names) of the higher-level general 50 51 groupings of armoured and plated dinosaurs continued to chop and 52 change throughout much of the 20th century, reflecting in large measure 53 54 the fragmentary nature of their remains. One voice of clarity across this 55 56 period to time proved to be that of Al (Alfred Sherwood) Romer (1927). 57 He elevated the ankylosaurs to subordinal rank and articulated the case 58 59 for distinguishing ankylosaurs from stegosaurs (and all other 60
10 Page 11 of 215 Zoological Journal of the Linnean Society
1 2 3 ornithischians) on the basis of their distinctive pelvic structure and the 4 5 form of their body armour. This suggestion was not followed by other 6 7 workers, but was reaffirmed by Romer (1956, 1966). Indeed it was not 8 until the work of Coombs (1971, 1978) that a greater measure of clarity 9 10 was established for the ankylosaurs, as well as Scelidosaurus. 11 12 With the advent of a cladistic approach to systematics, the 13 14 classification and implied phylogenetic relationships of the armoured and 15 plated dinosaurs has become more stable. Nopcsa’s (1915) suprafamilial 16 17 category Thyreophora (introduced by Nopcsa to replace the older term 18 19 Orthopoda) was re-instated by Norman (1984) for a clade of armoured 20 quadrupedal ornithischians that included ankylosaurids and nodosaurids 21 For Review Only 22 and he positioned Scelidosaurus as the basal sister-taxon to the 23 24 Ankylosauridae. This topology was opposed by Sereno (1984) who 25 proposed that the same clade name Thyreophora be used to include the 26 27 Ankylosauridae, Nodosauridae, Stegosauridae, Pachycephalosauridae and 28 29 Ceratopsia – much closer to Nopcsa’s original conception of the 30 Thyreophora. In response to the clear and obvious disagreement between 31 32 these two analyses Sereno (1986) tacitly accepted, without 33 34 acknowledgement, the topology generated by Norman (1984) and 35 36 formalised a classification and phylogeny of the Ornithischia, albeit 37 without a character matrix or numerical phylogenetic analysis. This 38 39 schema posited Scelidosaurus as a basal sister-taxon to the clade named 40 41 Eurypoda (Ankylosauria + Stegosauria: Sereno 1986: fig.3). This topology 42 has remained largely unchallenged since the early 1990s. 43 44 45 46 47 SCELIDOSAURUS: HISTORICAL CONTEXT 48 49 50 51 The coincidence in timing of the discovery of the near complete 52 53 scelidosaur skeleton (NHMUK R1111) in relation to the gradual unfolding 54 of an understanding of the anatomy and relationships of dinosaurs was 55 56 remarkable (Norman 2000b, 2001). The year 1858 coincided with Joseph 57 58 Leidy’s preliminary study of the partial remains of the dinosaur 59 Hadrosaurus foulkii (Leidy 1859a,b), which was to challenge the early 60
11 Zoological Journal of the Linnean Society Page 12 of 215
1 2 3 Victorian (Owen inspired) conception of the appearance of dinosaurs 4 5 (Desmond, 1975). Until Leidy’s work was published, the dominant 6 7 interpretation of the appearance and characteristics of dinosaurs was that 8 created by Richard Owen (1842). Initially his concept of pillar-limbed, 9 10 giant land reptiles was a purely intellectual exercise but, somewhat later, it 11 12 was given presence and form (and popularized) through the creation of 13 life-sized iron-framed, concrete and tiled models for the landscaped 14 15 gardens of the Crystal Palace Park that was created in 1853-4 (Desmond 16 17 1975; Wilford 1985; Norman 1985, 1991, 2017; Rudwick 1992; Cadbury 18 2000). 19 20 Leidy described Hadrosaurus as an altogether taller, more upright 21 For Review Only 22 (bipedal) animal with a somewhat kangaroo-like stance and appearance. 23 24 This alternative view of dinosaurs became prevalent in the USA and 25 contrasted strikingly with Owen’s original models of dinosaurs: it in effect 26 27 constituted a profound ‘Old World’ vs ‘New World’ divergence in 28 29 interpretation and understanding (Norman 2000b, 2001). However, just 30 31 as Leidy was describing his ‘New World’ Hadrosaurus, Richard Owen was 32 presented with the nearly complete, articulated skeleton of a large, 33 34 armoured (and perforce scaly), pillar-limbed and apparently quadrupedal 35 36 ‘Old World’ dinosaur. Scelidosaurus represented (or so it seems from the 37 perspective of the present day) an indisputable vindication of Owen’s 38 39 (1842) inductive method (based as it was upon incomplete and extremely 40 41 fragmentary fossil evidence), insofar as dinosaur anatomy and possible 42 appearance were concerned. It is interesting to observe that neither 43 44 Owen’s 1861 nor his 1863 monograph dwells on the posture or 45 46 reconstruction of this dinosaur. Indeed, Owen’s remarks show 47 considerable ambivalence with regard to the habits and mode of life of 48 49 this new dinosaur. It seems probable that Owen had been strongly 50 51 influenced (albeit unacknowledged) by Leidy’s reasoning in respect of the 52 habits of Hadrosaurus (Norman 2000b, 2001). Leidy interpreted his New 53 54 Jersey dinosaur as amphibious in habit because its bones had been 55 56 collected from marine rocks, and because it had short forelimbs, long hind 57 limbs and a deep (scull-like) tail. 58 59 60
12 Page 13 of 215 Zoological Journal of the Linnean Society
1 2 3 In the first of the two monographs, Owen constructed a convoluted 4 5 argument concerning the likely habits of scelidosaurs from comparisons 6 7 with what he knew of the skeleton of Iguanodon and the small skeletal 8 bones of the supposed juvenile Scelidosaurus. He noted the comparatively 9 10 small size of the forelimb of Iguanodon (as had been argued by Mantell in 11 12 1852) and related this to the swimming abilities of “the living land lizard 13 of the Gallopagos [sic] Islands, called Amblyrhynchus” (Owen 1861: 6). 14 15 These squamates hold their shorter forelimbs against the sides of their 16 17 bodies, while swimming by undulating their flattened tail. Owen also 18 posited that the collection of small skeletal bones of the scelidosaur may 19 20 have belonged to a “neonate”. The “parent” of the smaller individual, he 21 For Review Only 22 suggested, might have been represented by the larger skeleton whose 23 skull he had just described. He envisaged these two individuals swimming 24 25 in shallow water before perishing and becoming buried on the sea floor. 26 27 Owen also used the known habits of living armoured crocodiles in order to 28 subtly reinforce his line of reasoning. Having established that scelidosaurs 29 30 had an “aptitude for swimming” (Owen 1863: 12), he allowed himself to 31 32 conclude, in his second monograph, that the skeleton was designed “for 33 terrestrial rather than aquatic life, or at least for amphibious habits on the 34 35 margins of a river rather than for pursuit of food in the open sea” (Owen 36 37 1863: 26). It would seem that the combination of the crocodile-like 38 armour plating that covered the body of the scelidosaur, combined with 39 40 the longer and more dinosaurian hind limb and sacrum of this animal, 41 42 persuaded Owen that it was on balance terrestrial in habit but could, at 43 times, swim like a crocodile. A little later in the same concluding 44 45 paragraph he, disconcertingly, refers to the forelimbs of Scelidosaurus as 46 47 “paddles”. 48 49 Scientifically and historically, Owen’s reporting of Scelidosaurus 50 51 represents a curious mixture of intensely technical anatomical description 52 mixed with classically Owen-style sophistry. It is surprising (in retrospect) 53 54 that Owen did not consider it necessary to demonstrate the validity of his 55 56 inductive reasoning of 1842 (regarding the anatomical construction of 57 dinosaurs) and its ‘proof’ in the form of Scelidosaurus (Norman, 2000b, 58 59 2001). The similarity between the new, nearly complete dinosaur skeleton 60
13 Zoological Journal of the Linnean Society Page 14 of 215
1 2 3 from Dorset and Owen’s model of Hylaeosaurus (Fig. 6) – based as it was 4 5 upon the incomplete armoured dinosaur skeleton (Mantell 1833) – for the 6 7 new Crystal Palace Park, is striking indeed (Norman 2000b, 2001). Owen 8 allied Scelidosaurus and Hylaeosaurus in both structure and inferred 9 10 habits (Owen 1863: 12) but left it to the reader to draw the necessary 11 12 inferences. What seems particularly unfortunate, from both anatomical 13 and systematic perspectives, is that since the time of Owen’s monographs 14 15 there has never been a serious review of, or supplement to, his work. 16 17 18 19
14 Page 15 of 215 Zoological Journal of the Linnean Society
1 2 3 the anatomy of this discovery could not be described, given the 4 5 techniques available at the time. 6 7 Since 1863, there has been almost no new research on the 8 9 osteology, functional morphology or systematics of Scelidosaurus. Brief 10 commentaries, with low-resolution photographs of the specimen collected 11 12 by J.F. Jackson (NHMUK R6704) were published by Rixon (1968, 1976), 13 14 Charig & Horsfield (1975) and Charig (1979). Newman (1968) 15 photographed some of the limb bones of Owen’s hypodigm and in doing 16 17 so observed that the “… bones of the young individual [LYMPH 1998 6.1- 18 19 6.7] … are … possibly related to the Wealden genus Hypsilophodon. 20 Comparison with the skeleton (R1111), suggests they are not of the same 21 For Review Only 22 species …” (Newman 1968: 42). Swinton (quoted in Rixon) identified the 23 24 acid-prepared similarly small specimen (NHMUK R6704) as a 25 Scelidosaurus (?) juvenile. However, other authors challenged Swinton’s 26 27 identification, prompted largely by Newman’s commentary. 28 29 Thulborn (1977: 738) used limb proportion measurements and 30 31 some general anatomical comparisons to claim that “[t]his dolichopodous 32 33 ornithopod is quite distinct from Scelidosaurus harrisonii and is most 34 conveniently included in the family Fabrosauridae owing to its many 35 36 resemblances to the Triassic Fabrosaurus australis. The Liassic fabrosaurid 37 38 may well be identical with the specimen of ‘Hypsilophodon or some allied 39 form’ identified by Newman (1968) in the Charmouth Lias.” Coombs, 40 41 Weishampel & Witmer (1990: 431) referred to NHMUK R6704 as an 42 43 “unnamed thyreophoran”. Now that the prepared skeleton of the lectotype 44 is available for detailed comparison, it is clear that NHMUK R6704 is, as 45 46 Swinton suspected, an ontogenetically immature (juvenile) specimen of S. 47 48 harrisonii. The other small, partial skeleton (LYMPH 1998 6.1-6.7) is 49 nearly identical to NHMUK R6704 but appears to be slightly different 50 51 anatomically because all of its limb bones are crushed flat. The lectotype 52 53 (NHMUK R1111) is approximately five times larger than NHMUK R6704, 54 but despite the size difference the individual elements of these two 55 56 skeletons resemble each other remarkably closely. Observable differences, 57 58 most notably the profile of the preacetabular process of the ilium, can be 59 attributed most plausibly to the ontogenetically immature status of 60
15 Zoological Journal of the Linnean Society Page 16 of 215
1 2 3 NHMUK R6704, compounded by a degree of distortion (during compaction 4 5 and diagenesis) that has affected the lectotype and other referred 6 7 specimens. 8 9 Charig (1972) provided photographs and drawings of the pelvis and 10 femur of NHMUK R6704 (although the bones were incorrectly articulated). 11 12 Barrett (2001) described aspects of tooth anatomy, wear patterns and jaw 13 14 action in NHMUK R1111. Carpenter et al. (2013) provided several 15 photographs of the pelvic bones and sacrum of Scelidosaurus (using the 16 17 lectotype and the referred cast specimen SGDS 1311). Unfortunately, 18 19 while manipulating the fragile bones of the lectotype, for photographic 20 purposes, breakages occurred. The pelvis-sacrum images were used for 21 For Review Only 22 comparative purposes (rather than description), but further errors were 23 24 made in their attempted reconstruction (Norman [Part 2]). 25 26 Most other references to this taxon have tended to be in the form of 27 28 taxonomic listings or brief mentions within the context of more general 29 reviews (e.g. Romer 1956, 1966, 1968; Steel 1969; Norman 1985; 30 31 Coombs, Weishampel & Witmer 1990; Norman, Witmer & Weishampel 32 33 2004b). In recent years, Scelidosaurus has been used as a reference 34 taxon for systematic analyses of ornithischian dinosaurs (Butler et al. 35 36 2008, Baron et al. 2017b) and more particularly those of the armoured 37 38 and plated or thyreophoran groupings (Maidment et al. 2008, Mateus et 39 al. 2009, Arbour & Currie, 2016, Raven & Maidment 2017, Brown et al. 40 41 2017). Unfortunately, many of the anatomical characters assigned to 42 43 Scelidosaurus have proved to be incorrect because they were obtained 44 either from Owen’s (incomplete) original descriptions or a cursory 45 46 examination of the original material. 47 48 Our knowledge of ornithischians in the Early Jurassic has 49 50 progressed substantially since the time of Owen following the discovery 51 52 and description of a number of new taxa, a few of which include well- 53 preserved and nearly complete skeletons. These include 54 55 Heterodontosaurus (Crompton & Charig 1962); Santa Luca, et al. 1976, 56 57 Santa Luca 1980; Butler et al. 2008, Norman et al. 2011); Eocursor 58 (Butler et al. 2007, Butler 2010); Lesothosaurus (Thulborn 1970, 1972; 59 60 Porro et al. 2015, Baron et al. 2017a); Scutellosaurus (Colbert 1981,
16 Page 17 of 215 Zoological Journal of the Linnean Society
1 2 3 Breeden 2016); and Emausaurus (Haubold 1990). There are a number of 4 5 more fragmentary and consequently less well-understood forms. Most 6 7 notable, and contentious, among these is Pisanosaurus (Casamiquela, 8 1967) which has consistently been claimed to be the earliest (Norian) 9 10 ornithischian, following the early descriptive work of Casamiquela (e.g. 11 12 Sereno 1991, 2012). However, this taxon has been doubted as an 13 ornithischian (Norman et al. 2004a, Padian 2013, Baron et al. 2017c) and 14 15 recently Agnolin & Rosadilla (2018) have established a strong case for this 16 17 taxon being a silesaurid dinosauromorph (as reasoned independently by 18 Baron et al. 2017c). 19 20 Understanding the osteology and early diversification of 21 For Review Only 22 ornithischians has become an increasingly important topic in the context 23 24 of dinosaur systematics and the relationship between the clades 25 Theropoda and Ornithischia (Baron et al. 2017b,c). Stratigraphically, 26 27 Scelidosaurus has always deserved attention because it is represented by 28 29 comparatively good quality skeletal material from the earliest phase of 30 diversification within the clade Ornithischia, which is now considered to be 31 32 the Early Jurassic, given the absence of substantiated ornithischian 33 34 material from the Late Triassic. 35 36 Caveat. Owen’s description (1861, 1863) was meticulous, but necessarily 37 38 limited in scope. Over a century later, the specimen was developed as 39 part of a project that spanned a period in excess of thirty years (1965- 40 41 1997). Consequently, the osteology of the lectotype of Scelidosaurus can 42 43 now be described more completely than was possible in Owen’s time. 44 Unfortunately, the prolonged acid exposure of the permineralised skeleton 45 46 (made necessary because of the sheer bulk of the marlstone blocks in 47 48 which the skeletal remains were embedded) resulted in chemical and 49 associated physical damage to individual bones and made many of them 50 51 extremely fragile. The excessive fragility (especially of the cranial, 52 53 presacral vertebrae and pelvic bones) can be attributed to a combination 54 of some compressional distortion, progressive leaching of minerals from 55 56 within individual bones and the pernicious release of gas (under pressure) 57 58 within the medullary region of individual bones. 59 60
17 Zoological Journal of the Linnean Society Page 18 of 215
1 2 3 Significant gaps in our knowledge of the osteology of the lectotype 4 5 cranium include mechanical damage to the tip of the snout and erosion of 6 7 the skull’s posterodorsal, posterolateral and ventral margins. It is also 8 extremely unfortunate that severe damage to the cheek, palate and 9 10 dermal skull roof resulted from mishandling of the fragile, completely 11 12 acid-prepared, three-dimensional skull. Postcranially (Norman [Part 2]), 13 most of the cervical series is missing and may have been lost through 14 15 erosion after the skeleton began to be exposed. What remains of the 16 17 cervical vertebral series has also been damaged by the drilling of holes for 18 mounting rods during preparation for mounting the unprepared skeleton 19 20 as an exhibit at the Natural History Museum. The entire right forelimb is 21 For Review Only 22 missing, whereas the left forelimb, just distal to the humerus, is also 23 missing. Judging from the condition of more recently discovered 24 25 specimens it seems likely that the forelimb eroded away because either 26 27 that part of the skeleton was not completely enclosed within a marlstone 28 nodule (carbonate precipitation expands outward from nucleation sites 29 30 represented by skeletal elements), or the forelimb-containing nodule 31 32 broke away as the skeleton was exposed on the cliff face and was lost. 33 The distal end of the tail is also incomplete, which is perhaps less 34 35 surprising. It has also become apparent, in the light of the discovery of 36 37 additional skeletal remains, that some of the osteodermal skeleton was 38 lost or had weathered away at the time of collection. We are fortunate 39 40 that other material collected at the same locality, and apparently referable 41 42 to this taxon, provides anatomical information that supplements nearly all 43 of the lectotype’s missing portions (Norman [Parts 2 and 3]). 44 45 46 47 48 Important additional referred specimens 49 50 Additional specimens have been recovered from the cliffs and foreshore at 51 52 Charmouth that supplement the osteological information afforded by the 53 lectotype. Nearly all specimens for which there is reasonable evidence of 54 55 provenance have come from either the Topstones or Stonebarrow 56 57 Flatstones Beds (upper Sinemurian) in the Black Ven Marls on The 58 Spittles-Black Ven cliff (Figs 1, 7). Many isolated bones retrieved from cliff 59 60 debris or on the foreshore at Charmouth are in the collections of the
18 Page 19 of 215 Zoological Journal of the Linnean Society
1 2 3 Natural History Museum (documented below). Several additional 4 5 specimens are held in the collections of The Bristol City Museum and the 6 7 Sedgwick Museum, Cambridge. A few specimens collected while 8 beachcombing or prospecting on the cliffs are known to be in private 9 10 collections in the Lyme Regis/Charmouth area. 11 12 13 14
19 Zoological Journal of the Linnean Society Page 20 of 215
1 2 3 The Charmouth Specimen 1 (BRSMG LEGL 0005). During the winter of 4 5 1994-1995 David Sole collected another, intermediate-sized partial 6 7 skeleton of a scelidosaur from the beach, about midway between 8 Charmouth and Lyme Regis. Its lithology suggests that it was derived 9 10 from the Topstones Bed (Obtusum zone). This specimen was prepared for 11 12 exhibit by the late David Costin and is now on permanent display at the 13 Bristol City Museum, where it is recorded as BRSMG LEGL 0005. 14 15 The Charmouth Specimen 2 (BRSMG LEGL 0004). This is a well-preserved 16 17 sub-adult skeleton (smaller than the lectotype – contra Carpenter et al. 18 19 2013). The first portions of this skeleton were found by David Sole, on the 20 beach west of the Charmouth beach car park following a cliff fall on Black 21 For Review Only 22 Ven in December 2000. Further portions of this skeleton were found 23 24 sporadically over the next decade. The marlstone nodules containing the 25 skeleton came from the Topstones Bed (Obtusum zone) on Black Ven. 26 27 Additional blocks containing parts of this skeleton were recovered by Peter 28 29 Langham, Jo Anderson, Andrew Sole, Christine Endecott, Rick Taylor and 30 Bernie Abbott. The skeleton was reassembled under the stewardship of 31 32 David Sole and was prepared for display using a combination of dilute 33 34 acid, lapidarist’s tools and low-frequency vibropens by the late David 35 36 Costin. The original specimen (BRSMG LEGL 0004) is on permanent 37 display to Bristol City Museum. A cast of this specimen can be seen at the 38 39 Charmouth Heritage Coast Centre on the seafront at Charmouth and 40 41 another copy exists as a cast at the St George Discovery Site, Utah (SGDS 42 1311). 43 44 Note. BRSMG LEGL 0004 & 0005 remain in private ownership. Describing 45 46 these additional referred specimens may appear to some to violate the 47 48 ethical guidelines adopted by the Society of Vertebrate Paleontology. 49 However, the original specimens are held in the collections of the Bristol 50 51 City Museum and are on permanent loan with the agreement of David 52 53 Sole; they have their own collection reference numbers and are available 54 to all researchers (and have been so for over a decade). 55 56 57 58 59 Programme of work 60
20 Page 21 of 215 Zoological Journal of the Linnean Society
1 2 3 This article is one of three linked, descriptive accounts: the first focused 4 5 upon the skull [Part 1], a second on the postcranial endoskeleton [Part 2] 6 7 and a third on the dermal skeleton [Part 3]. These monographs provide 8 the foundation for a fourth [Part 4] on the palaeobiology and systematic 9 10 position of this important early ornithischian dinosaur. MicroCT scans of 11 12 some of the cranial material of the lectotype (NHMUK R1111) and the 13 smaller referred specimen (BRSMG Ce12785) are being visualized and 14 15 segmented in order to reveal some additional details concerning the 16 17 internal cranial and dental anatomy of this taxon (Norman & Porro, in 18 preparation); and, finally, a MicroCT and thin-section-based survey of the 19 20 internal gross anatomy, growth and development of scelidosaur 21 For Review Only 22 osteoderms is being completed (Norman & Baker, in preparation). 23 24 25 26 SYSTEMATIC PALAEONTOLOGY 27 28 Indented hierarchy 29 30 DINOSAURIA Owen, 1842 (sensu Gauthier 1986) 31 32 ORNITHOSCELIDA Huxley, 1870 (sensu Baron et al. 2017b) 33 34 ORNITHISCHIA Seeley, 1887 35 36 THYREOPHORA Nopcsa, 1915 (sensu Norman 1984) 37 38 39 Type-genus: Scelidosaurus Owen, 1859 40 41 42 43 Type species. S. harrisonii Owen, 1861. Early Jurassic, Dorset, England. 44 45 Diagnosis. Monotypic genus, therefore as for species S. harrisonii. 46 47 48 49 50 Scelidosaurus harrisonii Owen, 1861 51 52 53 54 Lectotype. NHMUK R1111 (Charig & Newman 1992; BZN 1994). The 55 56 remains of a medium-sized (5-6 metres long) armoured ornithischian 57 dinosaur, comprising the skull, lower jaws (missing only the rostral 58 59 portion of the snout), and most of the articulated postcranial skeleton 60
21 Zoological Journal of the Linnean Society Page 22 of 215
1 2 3 (missing portions of the cervical series and distal portions of both 4 5 forelimbs). 6 7 8 9 Derivation of name. Skelos (Gr) = limb/leg + Sauros (Gr) = lizard/reptile 10 11 + harrisonii = belonging to James Harrison. 12 13 The generic name was probably chosen to distinguish this terrestrial (or 14 15 amphibious) taxon from the paddle-limbed reptiles that predominate in 16 17 the Liassic beds of the Charmouth/Lyme Regis area. 18 19 20 21 Stratigraphic rangeFor. Charmouth Review Mudstone Only Formation (upper Sinemurian), 22 23 Black Ven Marl Member, Asteroceras obtusum Zone, Obtusum Subzone 24 (Hesselbo & Jenkyns 1995). The nodule-bearing beds that occasionally 25 26 reveal these remains are referred to locally as the Topstones Bed and 27 28 there have been reports of scelidosaur remains occurring in the 29 Stonebarrow Flatstones Bed (see Fig. 7). 30 31 An additional report (Ensom 1987) records scelidosaur remains as 32 33 having been collected from the lower Pliensbachian (Belemnite Marl 34 35 Member, Uptonia jamesoni Zone) at Seatown (on the coast east of 36 Charmouth). These specimens are considered likely to have been derived 37 38 from Charmouth. 39 40 41 42 43 Locality. Cliff exposures on The Spittles-Black Ven between Charmouth 44 and Lyme Regis, Dorset. Weathered nodules containing bones and isolated 45 46 eroded bones are found occasionally on the foreshore beneath these cliffs 47 48 (Fig. 1). 49 50 51 52 Diagnosis (Cranial only – the equivalent Diagnosis in Part 2 will add 53 54 autapomorphies identified in the postcranial endoskeleton. Part 3 will 55 provide a synthetic Diagnosis based upon the total anatomy of this 56 57 taxon). Rugose, mound-shaped projection on the premaxillae supports a 58 59 small, keratinous beak; facet for the posteromedial process of the 60
22 Page 23 of 215 Zoological Journal of the Linnean Society
1 2 3 premaxilla positioned on the laterodorsal surface of the nasal; small, 4 5 bean-shaped, shallow antorbital fossa delimited by a sharp edge ventrally; 6 7 the sagittal crest comprises two parallel crests separated by a narrow 8 midline trough; localised exostosis restricted largely to the angular on the 9 10 mandible (no osteoderm present); epivomer plates form part of the roof 11 12 of the nasal chambers; narrow pocket-like slot on the medioventral 13 surface of the quadrate wing of the pterygoid; epipterygoid bone forms a 14 15 small, vertically oriented, conical structure with a laterally flattened base 16 17 applied to the dorsolateral surface of the pterygoid; large oblique facets 18 on the ventrolateral wall of the basioccipital; expanded and ventrally 19 20 faceted pedicle on the opisthotic; elongate epistyloid bones project 21 For Review Only 22 obliquely (posteroventrally) from the skull; spur-like structure on the 23 dorsal edge of the paroccipital encloses the post-temporal fenestra; a pair 24 25 of large curved osteodermal horns are sutured to the posterodorsal 26 27 surface of the occiput. 28 29 30 31 INVENTORY OF MATERIAL 32 33 34 35 Lectotype (Owen 1861, 1863; vide Charig & Newman 1992; BZN 1994) 36 37 NHMUK R1111. Major part of the skull (including the lower jaws) and 38 39 partial articulated skeleton of a large individual (est. 5+ metres long). 40 Missing elements: the rostral portion of the skull, parts of the skull table 41 42 and occiput; most of the cervical vertebral series and some caudals; all of 43 44 the forearm and manus of both sides and portions of the osteoderm 45 covering. 46 47 48 Owen (1863: 1) reported that the skull and skeleton were 49 “discovered during the quarrying operations on the face of the cliff of 50 51 Lower Lias at Charmouth, Dorsetshire, with liberal encouragement to the 52 53 workmen.” A lack of experience in fossil extraction on the part of the 54 quarrymen may have played a rôle in the loss of (and damage to) various 55 56 portions of what may well have been a more or less complete, articulated, 57 58 skeleton. However, despite Owen’s words, it should also be noted that 59 skeletal remains within the mudstones of the Lias tend to form foci for 60
23 Zoological Journal of the Linnean Society Page 24 of 215
1 2 3 ‘nucleation’ (precipitation of carbonate and other minerals from ground 4 5 water). This cumulative process leads to the creation of discrete marlstone 6 7 nodules (locally named ‘doggers’); these may gradually coalesce over 8 time as adjacent nodules enlarge and merge with one another. Landslips 9 10 on The Spittles-Black Ven are frequent and nodular layers are often 11 12 exposed as projecting ‘benches’ that become fractured by spring-water 13 lubricated slippage and rotational faulting. Irregular chunks break off and 14 15 form part of the scree that eventually accumulates on the foreshore, 16 17 where they are subjected to wave action. Some of the lectotype blocks, 18 such as the one containing the skull, may well have suffered this fate 19 20 before they were collected. 21 For Review Only 22 23 24 Paratypes (vide Newman 1968; BZN 1994) 25 26 27 These specimens including the lectotype (above) formerly constituted the 28 syntypic series from which the lectotype (NHMUK R1111) was selected. 29 30 Following lectotype designation (BZN 1994) the syntypes became para- 31 32 lectotypes and no longer eligible for selection as type specimens of the 33 nominal genus and species. 34 35 NHMUK OR109560 - (Owen 1861: see Fig. 2A). Femur, damaged upper 36 37 (proximal) portion of a large, left femoral shaft. Theropoda indet. 38 39 NHMUK OR109561 – (Owen 1861: see Fig. 2B). Ungual phalanx, eroded. 40 41 Theropoda indet. 42 43 NHMUK OR39496 - (Owen 1861: see Fig. 2B). Distal end of large right 44 45 femur and the proximal end of the associated fibula. Theropoda indet. 46 47 NHMUK OR39497. (Owen 1861: see Fig. 2B) Tibia, large proximal portion. 48 49 Theropoda indet. (probable renumbered association with OR39496 – 50 above) 51 52 LYMPH 1998 6.1-6.7 – (Owen 1861: see Fig. 3). Scelidosaurus. Left 53 54 femur, distal portion left tibia, calcaneum, 2 metatarsals, 2 phalanges of 55 56 the pes and an isolated caudal centrum of a juvenile individual of 57 Scelidosaurus cf. harrisonii (est. 1.3-1.5m long). NHMUK R5909: plaster 58 59 cast replicates of this partial skeleton. 60
24 Page 25 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 Referred material 6 7 8 1. Natural History Museum, UK 9 10 NHMUK R6704 (Rixon, 1968). A small (est. 1.5 metres long) juvenile, 11 specimen (closely comparable in size to the paratype associated skeletal 12 13 remains LYMPH 1998 6.1-6.7). Articulated mid-section of the body 14 15 comprising 13 dorsal vertebrae and ribs, 4 sacral vertebrae and 16 17 associated sacral ribs, 7 caudal vertebrae (including 11 ribs), 5 chevron 18 bones (haemal arches), ilia (r & l), pubes (r & l), ischia (r & l), femora (r & 19 20 l), 6 lateral osteoderms and scattered portions of ossified tendons remain 21 For Review Only 22 attached to neural spines along the dorsal and sacral series but most were 23 removed (preserved in small bundles) during the process of acid-mediated 24 25 removal from the marlstone dogger in which the original specimen was 26 27 embedded. (See also A.E. Rixon, ms. NHM preparation laboratory 28 archive). 29 30 NHMUK R12019. A posterior fragment of the occiput and an articulated 31 32 series of (at least) six cervical vertebrae, commencing with the axis, of a 33 34 large individual (est. 4 metres long) – remnants of the atlas seem to be 35 visible in cross-section on the fractured surface of the block. The 36 37 vertebrae are flanked by several sub-conical, horn-shaped osteoderms. 38 39 The specimen comprises a slab of marlstone that has been broken along 40 the midline; it has not yet been cleared of matrix. Provenance unrecorded 41 42 (?Charmouth). 43 44 NHMUK R10103. Indeterminate bone fragments 45 46 NHMUK OR28333. Two scapulae 47 48 49 NHMUK OR32396. Radius 50 51 NHMUK OR39516. Osteoderms. 52 53 NHMUK OR39517. Caudal or ?sacral vertebra, small distal caudal vertebra, 54 55 small distal caudal. 56 57 NHMUK OR39518. Haemal arch (chevron). 58 59 NHMUK OR39519. Two phalanges. 60
25 Zoological Journal of the Linnean Society Page 26 of 215
1 2 3 NHMUK OR39520. Ungual phalanx of ?pes. 4 5 NHMUK OR39521. Complete, uncrushed MT1. Several other bone 6 7 fragments. 8 9 NHMUK OR40503. Tibia (proximal end). 10 11 NHMUK OR41322. Femur, left, large (smaller than lectotype) near 12 13 complete, well-preserved 4th trochanter, which appears to be 14 15 thickened/reinforced by a layer of metaplastic bone. 16 17 NHMUK OR41323. Radius. 18 19 NHMUK OR41324. Scapular blade (central portion). 20 21 NHMUK OR41325.For Humerus Review (shaft of large Only specimen). 22 23 NHMUK OR41326. Fibula (proximal end). 24 25 26 NHMUK OR41327. Two caudal vertebrae. 27 28 NHMUK OR41328. Three metatarsals (associated II-IV) 29 30 NHMUK OR41329. Osteoderm. 31 32 NHMUK OR41330. Ischium (shaft only) 33 34 NHMUK OR42068. Six osteoderms (2 are articulated together via 35 36 remnants of a base-plate; these are anterior cervical osteoderms). 37 38 NHMUK OR42070. Phalanx. 39 40 NHMUK OR42072. Large bone fragment – indet. 41 42 43 NHMUK OR42074. Ulna (paired with NHM OR41323). 44 45 NHMUK OR43062. Shaft of a large humerus, crushed proximal femur, 46 fibula, fragmentary radius?, other large bone fragments. 47 48 49 NHMUK OR46011. Probable rib fragment – indet. 50 51 52 53 2. Bristol City Museum and Art Gallery, UK 54 55 BRSMG Ce12785. Partial skull and some disarticulated postcranial 56 57 elements including much of the hindlimb of an intermediate-sized 58 individual (est. 3m long), though heavily compressed. Associated 59 60
26 Page 27 of 215 Zoological Journal of the Linnean Society
1 2 3 marlstone slabs include patches of what appears to be the impressions of 4 5 a scaly skin texture. Collected on Black Ven (horizon undisclosed). 6 7 BRSMG Ce12787. Partly eroded femur and vertebra embedded in a 8 9 marlstone nodule. (Charmouth) 10 11 BRSMG Ce12788. Partly eroded and prepared marlstone nodule with 12 13 embedded cervical osteoderms. (Charmouth) 14 15 BRSMG Ce12789. Vertebra and bone fragments embedded in a marlstone 16 nodule. (Charmouth). 17 18 19 BRSMG Cf2781. Articulated, hemi-sected marlstone nodule (including nine 20 middle-distal caudal vertebrae and some haemal arches) with some 21 For Review Only 22 organic (kerogenised) remains of the skin preserved. Probably a large 23 24 individual (est. 3.5 metres long). No provenance information was recorded 25 with this specimen. (ex-Challinor Collection, Martill et al. 2000). 26 27 BRSMG LEGL 0004. Articulated skeleton of large size (est. 4 metres long) 28 29 including the skull, presacral vertebral column, the shoulder girdle and 30 31 proximal portions of the forelimb of the left side, most of the pelvic girdle 32 and most of the hindlimb (with the exception of the distal halves of the 33 34 metatarsals and toes). The proximal part of the tail (the first 13 caudal 35 36 vertebrae) are preserved, but the remainder of the tail is missing. In 37 38 addition, much of the dermal skeleton is preserved as an array of partly 39 articulated osteoderms attached to the skull and flanking both sides of the 40 41 neck, torso and proximal tail region. Collected from the Topstone Bed on 42 43 Black Ven (some parts had become separated and were collected 44 individually from the foreshore beneath Black Ven). 45 46 BRSMG LEGL 0005. Articulated partial skeleton (est. 3 metres long) 47 48 including part of the occiput and the presacral skeleton including the 49 50 forelimb, portions of the pelvis and sacrum and much of the hindlimb as 51 well as a few isolated caudal vertebrae. The articulated cervical osteoderm 52 53 arrays although incomplete are well preserved. One slab collected on the 54 55 foreshore below Black Ven by David Sole, and other portions were 56 retrieved from the Topstones Bed on Black Ven. 57 58 59 60
27 Zoological Journal of the Linnean Society Page 28 of 215
1 2 3 3. Dorset County Museum (Dorchester), UK 4 5 DORCM G.7542. Osteoderms (Charmouth) 6 7 8 DORCM G.7842. Three dorsal vertebrae in articulation (Locality: Seatown. 9 Horizon: Belemnite Marls). This occurrence is considerably higher in the 10 11 stratigraphic succession than all other known scelidosaur material. It is 12 13 considered more probable (by the author) that these specimens were 14 transported to Seatown by longshore drift, having previously eroded out 15 16 of the cliffs at Charmouth. 17 18 DORCM G.8590. Three dorsal vertebrae (‘Lias of Dorset’ - no other 19 20 documentation). 21 For Review Only 22 DORCM G.10817. Osteoderms (Charmouth). 23 24 25 26 4. Sedgwick Museum, Cambridge, UK 27 28 29 CAMSM X39256. Cranial: disarticulated skull and jaw elements. Axial: the 30 entire cervical, dorsal and sacral vertebral column; a large number of 31 32 cervical, dorsal and sacral ribs (many of which are complete). 33 34 Appendicular: includes the substantial parts of the pectoral and pelvic 35 girdles, as well as parts of the proximal fore and hind limb bones. Dermal 36 37 skeleton: an important collection of associated cervical osteoderms. 38 39 Estimated length approximately 2 metres. Collected from Black Ven, 40 Charmouth (horizon undisclosed). 41 42 43 44 45 5. Saint George Dinosaur Discovery Site, Utah, USA 46 47 SGDS 1311. Resin-based cast museum display mount. The display 48 specimen has been augmented in order to create the impression of a 49 50 more complete skeleton than the original upon which it is based. Much of 51 52 the cast is an accurate reproduction of the original Scelidosaurus skeleton 53 (BRSMG LEGL 0004) but there are some modifications and sculpted 54 55 additions. 56 57 58 59 Note concerning the usage of anatomical descriptors 60
28 Page 29 of 215 Zoological Journal of the Linnean Society
1 2 3 Terms used in this article to indicate direction and orientation of cranial 4 5 anatomical structures follow those used in the established literature 6 7 associated with reptilian anatomy (viz. Romer 1956). In the instance of 8 Scelidosaurus, whose skull was held horizontally, the following terms are 9 10 in essence self-explanatory: anterior (toward the front end) and posterior 11 12 (toward the neck and tail); medial (toward the midline) and lateral (away 13 from the midline); dorsal (above the skull) and ventral (toward the 14 15 bottom of the skull); proximal (near to) and distal (away from) the centre 16 17 of the skull. The terms median (on the midline) or the equivalent sagittal 18 are also used, but their use is dependent upon context and prior usage, 19 20 e.g. the sagittal crest on the parietals is a median structure. The 21 For Review Only 22 directional terms ‘rostral’, ‘cranial’ and ‘caudal’ are not used. 23 24 In the case of tooth descriptors: mesial (toward the mandibular 25 symphysis), distal (away from the mandibular symphysis), buccal (toward 26 27 the cheek or exterior), lingual (toward the tongue or interior of the 28 29 mouth), basal (toward the root or bottom of the alveolus) and apical 30 (toward the tip of the crown) will be used. This conforms, more or less, to 31 32 the usage found in most non-human dental literature (Edmund 1969). 33 34 These terms describe unambiguously the orientation and morphology of 35 36 individual teeth at any position in upper or lower dentitions and across a 37 dentition of any shape. This avoids the ambiguity implicit in terms such as 38 39 ‘anterior’, ‘posterior’, ‘medial’, ‘lateral’, ‘labial’, ‘ventral’, ‘dorsal’ and 40 41 ‘coronal’. 42 43 44 45 DESCRIPTIVE OSTEOLOGY: THE DERMAL SKULL ROOF 46 47 48 49 General comments 50 51 Up to the present, our knowledge of the skull of Scelidosaurus has been 52 53 wholly dependent upon the original description of the lectotype (NHMUK 54 55 R1111) provided by Owen (1861) – see Figure 4. Since the 1860s no 56 further descriptive work has been undertaken on this material; 57 58 nevertheless, the skull has been prepared chemically using immersion in 59 60 dilute acetic acid to dissolve the matrix and reveal the interior of the skull.
29 Zoological Journal of the Linnean Society Page 30 of 215
1 2 3 By the time that this preparatory phase had been completed (in the early- 4 5 mid 1970s) this specimen had become extremely fragile but was the best 6 7 three-dimensionally exposed early ornithischian skull then known. Since 8 that time this truly remarkable specimen has been progressively broken 9 10 and degraded. It is also evident, from simply comparing Owen’s original 11 12 lithographs of the skull (Fig. 4) to its present condition, that the natural 13 bony surface of the skull has deteriorated through a combination of 14 15 chemical action (leaching) and the need to apply layers of consolidant to 16 17 the bone surface. The lectotype skull, once three-dimensional and intact, 18 is now reduced to an assortment of bony elements: one major portion 19 20 comprises the right side of the facial region, the attached skull roof, 21 For Review Only 22 braincase and suspensorium. The left side of the skull has been broken 23 away from the rest of the skull and comprises broken palate bones, the 24 25 maxilla and lacrimal, the jugal arch and associated parts of the 26 27 suspensorium. There has also accumulated a considerable quantity of 28 small bony shards and powdered bone created by each fresh break and 29 30 this material has, from time to time, been collected up and disposed of. As 31 32 will be shown later, even isolated carefully stored bones have deteriorated 33 through mishandling or an underestimation of their extreme fragility. The 34 35 two mandibles are separate and seem robust (by comparison); they are 36 37 well preserved, but are thickly coated with consolidant. 38 39 By a combination of chance and persistence, further specimens of 40 41 Scelidosaurus have been collected across the years since the original 42 discoveries were made. Most important among these are the partial 43 44 skeleton and skull of an intermediate-sized individual collected by Simon 45 46 Barnsley in 1984 (now in the collections of the Bristol City Museum – 47 BRSMG Ce12785), and the articulated skull and skeleton of a larger 48 49 individual that was collected and assembled under the supervision of 50 51 David Sole from 2000 onwards (BRSMG LEGL 0004). Both of these latter 52 specimens were prepared (using vibropens and dilute acid). These 53 54 specimens show little evidence of chemical leaching and complement, to 55 56 an important degree, the anatomical information provided by the 57 lectotype. 58 59 60
30 Page 31 of 215 Zoological Journal of the Linnean Society
1 2 3 Cranial material, belonging to other taxa, that can be compared 4 5 realistically to the skull of Scelidosaurus is limited in range and variety. 6 7 Some comparisons are drawn with the skulls of the near-contemporary 8 (Hettangian-Sinemurian) basal ornithischian taxa Lesothosaurus and 9 10 Heterodontosaurus. However, the only truly comparable taxon is the 11 12 single specimen of the somewhat younger (Toarcian) thyreophoran 13 Emausaurus ernsti [EMAU SGWG 85] that is represented by a small 14 15 (140mm long) partial skull and lower jaw, as well as a few postcranial 16 17 fragments (Haubold 1990). A few other specimens have, in the past, been 18 regarded as comparable to Scelidosaurus: notably a lower jaw and skull 19 20 fragment of Tatisaurus oehleri (FMNH CUP 2088) from the (Early Jurassic) 21 For Review Only 22 Lufeng Formation of Yunnan, China. This latter specimen was referred to 23 the genus Scelidosaurus by Lucas (1996) but the material is poorly 24 25 preserved, non-diagnostic and regarded as a nomen dubium (Norman, 26 27 Butler & Maidment 2007). The cranial material associated with the 28 holotype of Scutellosaurus lawleri, from the Kayenta Formation in the 29 30 collections at Flagstaff, Arizona (MNA P1.175 – Colbert 1981) and a 31 32 considerable quantity of referred material in the collections at Austin, 33 Texas (TMM) and Berkeley, California (UCMP) – listed in Breeden (2016) is 34 35 unfortunately mostly fragmentary, crushed and distorted. 36 37 Other comparable taxa belong to the younger and more derived 38 39 thyreophoran clades Stegosauria and Ankylosauria (these together 40 41 constitute the Eurypoda sensu Sereno 1986) are referred to for 42 comparative purposes. Ankylosaurs are generally younger (Late Jurassic- 43 44 Cretaceous) and in many cases their skulls are of restricted comparative 45 46 value because they are highly modified. Much of the detailed cranial 47 anatomy of adult/mature specimens is obscured by a combination of 48 49 complete fusion between individual bones or obstructed by the presence 50 51 of a layer of thick, fused osteoderms (caputegulae [= ‘head tiles’] of 52 Blows 2001), see Coombs (1978) and Arbour & Evans (2017). There are 53 54 examples of partially eroded nodosaur skulls (e.g. Norman & Faiers 1996) 55 56 and several juvenile specimens of Pinacosaurus (Maryańska 1977, Hill et 57 al. 2011) that offer some osteological insights. Among stegosaurs the 58 59 most basal currently known taxon is Huayangosaurus taibaii from the 60
31 Zoological Journal of the Linnean Society Page 32 of 215
1 2 3 upper Middle Jurassic (Bathonian-Callovian) Lower Shaximiao Formation 4 5 of Sichuan Province, China (Sereno & Dong 1992). Unfortunately, the two 6 7 skulls currently known (ZDM 7001 and IVPP V6728) though preserved ‘in 8 the round’ are crushed and distorted, and their detailed anatomy is 9 10 obscured by adherent matrix. The only other anatomically well-known 11 12 stegosaur skull is that of the younger Upper Jurassic (Kimmeridgian) and 13 anatomically far more derived Stegosaurus (Gilmore 1914: pls 6-10). 14 15 Comments and comparisons are made below, where they are deemed to 16 17 be appropriate. 18 19 20 21
32 Page 33 of 215 Zoological Journal of the Linnean Society
1 2 3 sclerotic ring to be reconstructed with any confidence. The posterodorsal 4 5 margin of the orbit is slightly encroached upon by the overgrowth of a 6 7 crater-like osteoderm on the lateral face of the postorbital. 8 9 Beneath the rugose horizontal ledge that borders the antorbital 10 fossa and orbit the sidewall of the skull is smooth but deeply inset, 11 12 creating a pronounced horizontal buccal emargination (be) or cheek 13 14 recess, adjacent to the upper dentition and co-extensive with a similar 15 feature on the mandible. The infratemporal fenestra is large and ovoid 16 17 with its long axis is somewhat inclined posterodorsally; this fenestra is not 18 19 occluded by the overgrowth of any of the surrounding bones, or overlying 20 osteoderms. The quadrate is completely excluded from the margin of the 21 For Review Only 22 infratemporal fenestra by contact between finger-like processes of the 23 24 squamosal and quadratojugal. The jugal is elongate and deep posteriorly. 25 Its external surface is plastered by an extensive covering of periosteally 26 27 derived accreted bony tissue (exostoses). This latter tissue comprises 28 29 irregular bony strands in the area beneath the orbit; on the quadratojugal 30 wing of the jugal there are distinct patches of accreted tissue these have a 31 32 granular texture in places and strand-like in others. These discrete 33 34 patches are separated by smooth-surfaced grooves. The quadratojugal 35 36 bridges and seals off the posteriorly facing quadrate (paraquadrate) 37 foramen. The main body of the quadrate is pillar-like but has a shallowly 38 39 concave posterior margin, a short laterally flared jugal wing and a 40 41 considerably deeper, more expanded, pterygoid wing. 42 43 44 45
33 Zoological Journal of the Linnean Society Page 34 of 215
1 2 3 Radiating away from this central (granular) region is an array of elongate 4 5 bony strands that extend across the surface of adjacent skull roofing 6 7 bones. In the area adjacent to the supratemporal fenestrae the 8 postorbitals show small islets of periosteal bone that give the impression 9 10 that there is a process of encroachment around the adjacent margin of 11 12 the supratemporal fenestra. The internal surfaces of the supratemporal 13 fenestrae are smooth, providing areas that allowed the unhindered 14 15 movement of the adductor musculature that was attached to the margins 16 17 of these fenestrae. The margins of the supratemporal fenestra are 18 extremely rugose (notably along the pair of ridges that form the sagittal 19 20 crest) because they represent the anchoring sites for the jaw adductor 21 For Review Only 22 musculature. 23 24 The snout tapers gently anteriorly and the premaxillae are firmly 25 butt-sutured; there is a distinct transverse crease on its dorsal surface 26 27 (better seen in lateral view) that marks the base of the projecting portion 28 29 that supported a toothless ‘beak’ (rhamphotheca). The dorsolateral edge 30 of the skull, above the orbit and anterior part of the intertemporal arch is 31 32 dominated by a prominent, rugose, overhanging brow-ridge (Fig. 9) 33 34 formed by the arched and conjoined palpebral (anterior supraorbital) and 35 36 posterior supraorbital. The latter is sutured to a stubby tab-shaped 37 projection (and adjacent bowl-shaped slot) along the top edge of a 38 39 laterally positioned postorbital osteoderm. A narrow unossified space (see 40 41 Fig. 16, gap) remains between the posterior process of the palpebral and 42 the lateral margins of the prefrontal and middle supraorbital. This gap 43 44 (probably plugged with connective tissue and visible in the sub-adult 45 46 specimen (Fig. 16) might well become occluded by the continued spread 47 of exostotic tissue in fully adult individuals. 48 49 50 51 52
34 Page 35 of 215 Zoological Journal of the Linnean Society
1 2 3 (herein) as epivomers (epv). Farther posteriorly the pterygoids form 4 5 broadening, arched plates that separate the nasal cavity from the ethmoid 6 7 region/orbital cavities and shield the braincase; they also link laterally to 8 the upper jaw via the palatines and ectopterygoids. There is a narrow 9 10 interpterygoid vacuity, above and between which can be seen a narrow 11 12 parasphenoid rostrum (cultriform process). Farther posteriorly, the barrel- 13 like floor of the braincase is visible and flanked by two unusually large, 14 15 oblique, rugose tuberosities (bot) as well as prominent triangular pedicles 16 17 (op.ped) formed by ventral extensions of the opisthotics. 18 19 Perhaps most striking of all (in view of our previous dependence 20 upon the accuracy of Owen’s original description) is that the posterodorsal 21 For Review Only 22 margin of the occiput is ornamented by a pair of prominent, horn-shaped 23 24 osteoderms (Figs 16, oc.ost, 47). The reconstruction of the dorsal (Fig. 9) 25 and occipital aspects of the skull (Fig. 11) indicate the likely position of 26 27 the attachment sites (oc.f) for the pad-like base-plates that supported the 28 29 osteodermal horns. The occiput is reconstructed only tentatively because 30 no specimens show this surface particularly well. The lectotype provides 31 32 information about the general disposition of the occiput, but none of the 33 34 cranial sutures are clearly shown and seem to be obscured by a patina of 35 36 periosteally derived (exostotic) bone. The supraoccipital contributed to the 37 dorsal margin of the foramen magnum, and the pedicles of the 38 39 exoccipitals formed dorsolateral mounds on the occipital condyle. There is 40 41 an unusual spur-shaped process (Figs 11, 33, psp) on the dorsal margin 42 of the paroccipital process that surrounds most of the remnant of the 43 44 post-temporal fenestra (Fig. 11, ptf). A more medial position for any other 45 46 opening seems improbable in the skull of more mature individuals, given 47 the need for a sutural surface for anchorage of the occipital horns and the 48 49 extensive exostotic overlay. However, one juvenile specimen (CAMSM 50 51 X39256 – Fig. 32B,C, can) reveals the presence of a smooth oblique 52 groove representing what appears to be the lateral wall of a more 53 54 medially positioned oblique canal running from the occiput toward the 55 56 cranial cavity along the suture between the opisthotic and supraoccipital. 57 This canal presumably became plugged by accretionary bone growth 58 59 60
35 Zoological Journal of the Linnean Society Page 36 of 215
1 2 3 during ontogeny. The precise layout of the squamosals and parietals on 4 5 the occiput is not known. 6 7 8 9
36 Page 37 of 215 Zoological Journal of the Linnean Society
1 2 3 forms a small rostrum that is rugose, grooved (Fig. 13D, gr) and pitted 4 5 with tiny vascular openings. Viewed in profile this rostrum forms a 6 7 discrete, edentulous projection that is set off anteriorly relative to the 8 base of the median dorsal process by a transverse crease; the rostrum 9 10 has a slightly downturned tip (Figs 8, 13). It seems certain that the 11 12 rostrum was ensheathed in a keratinous rhamphotheca (beak). Posterior 13 to the small rostrum, the ventral edge of the external surface is 14 15 roughened (Fig. 13C, rug) and forms a curtain-like structure lateral to the 16 17 dentition. The surface morphology of this bone suggests that the 18 premaxillary dentition may have been at least partially sheathed (and 19 20 perhaps mechanically supported) by a posterior extension of the 21 For Review Only 22 rhamphotheca. 23 24 25 26
37 Zoological Journal of the Linnean Society Page 38 of 215
1 2 3 early Jurassic taxa such as Heterodontosaurus (Norman et al. 2011), and 4 5 as has been hinted at in Lesothosaurus (Porro et al. 2015). It is also worth 6 7 noting in passing that in general the premaxillary teeth are a little taller 8 and more conical, pointed and recurved compared to those on the maxilla 9 10 (Fig. 13B); however, the two posterior premaxillary teeth, though a little 11 12 larger, exhibit a transitional morphology and more closely resemble, in 13 shape and proportions, those of the maxillary series. 14 15 Mediodorsal to the alveolar parapet, the margins to the ventral roof 16 17 of the premaxilla display a row of foramina (sf – Edmund 1960; Horner et 18 19 al. 2004) connected by a shallow groove. Between the rows of special 20 foramina the premaxillae form a shallowly vaulted, triangular anterior 21 For Review Only 22 palatal roof, subdivided by an irregular median sutural line (Fig. 13B, 23 24 pms). Anterior to the tooth row and on the ventromedial edge of the 25 premaxillary rostrum, there is a pit (Fig. 13C, apo) marking the opening 26 27 for a canal that runs posterodorsally through the premaxilla and connects 28 29 to an anterior premaxillary foramen located in the anterior corner of the 30 external naris (Fig. 8, apf – Norman & Porro, in preparation). There is no 31 32 indication of an inter-premaxillary pocket or fossa in the palatal roof, as 33 34 described in Heterodontosaurus (Norman, et al. 2011: fig. 13) and 35 36 Hypsilophodon (Galton 1974: fig. 4B) and some other ornithischians. In 37 this latter respect the premaxillary palate of Scelidosaurus more closely 38 39 resembles that described in Lesothosaurus (Porro et al. 2015). The 40 41 premaxillae are butt-sutured along the midline and this suture was 42 evidently unfused in immature specimens (e.g. BRSMG Ce12785) because 43 44 the left and right premaxillae have separated and are offset from one 45 46 another (Fig. 13C, pms). Dorsal and posterior to the rostrum, each 47 premaxilla preserves the base of a short, curved, tapering dorsal process 48 49 (dp) that, when complete, evidently made sutural contact against a facet 50 51 preserved on the anterolateral margin of the anterior tip of the nasal in 52 BRSMG Ce12785 (Fig. 15, pms). This sutural arrangement is unusual 53 54 because in ornithischians the nasals usually diverge anteriorly and clamp 55 56 against either side of a tapering, median dorsal process formed by the 57 conjoined premaxilla. Clearly there was a complete bony rim to the 58 59 external naris. There is no obvious external evidence for an anterior 60
38 Page 39 of 215 Zoological Journal of the Linnean Society
1 2 3 premaxillary foramen in the anterior corner of the narial fossa, as 4 5 reported in Huayangosaurus (Sereno & Dong, 1992), Lesothosaurus 6 7 (Porro et al. 2015) and Heterodontosaurus (Norman et al. 2011). 8 However, MicroCT scan evidence confirms its presence (Norman & Porro, 9 10 in preparation). 11 12 Posterolaterally, each premaxilla produces an oblique, slightly 13 14 sinuous, tapering posterolateral process (Fig. 13D, pp) that wedges itself 15 between the maxilla and nasal and extends to a point just beyond the 16 17 medial anterior lamina of the lacrimal (although these bones do not make 18 19 sutural contact): premaxilla and nasal appear to be separated by the 20 dorsal lamina of the maxilla. The dorsomedial edge of the maxillary 21 For Review Only 22 lamina bears a narrow groove (Fig. 13B, gr) that represents a narrow 23 24 sutural contact area for the posterior premaxillary process. The nasal also 25 bears a step on its lower external surface and below this step is a medially 26 27 offset vertical ‘flange’ or, more accurately, a thin curtain-wall (Fig. 15, mf) 28 29 that offers a more substantial contact surface for the medial side of the 30 posterolateral process of the premaxilla. 31 32 33 The anterior sutural contact between the maxilla and premaxilla is 34 not clear. It is probable that this was strong and firm, and required an 35 36 anteromedially directed robust process of each maxilla to meet in the 37 38 midline, creating a wedge-like structure that slotted into a recess in the 39 posterior margin of the body of the premaxilla dorsal to the premaxillary 40 41 palatal roof (Fig. 10). The disarticulated maxilla of Emausaurus exhibits 42 43 just such a structure (Haubold 1990: taf. III). 44 45 Apart from the extreme rostral tip and adjacent surfaces, which are 46 47 all somewhat rugose and pitted (reflecting the active growth and 48 mechanical support for a keratinous rhamphotheca) there is little 49 50 indication of a superficial coating of periosteally derived bone, prevalent 51 52 on the exterior surfaces of many of the other skull bones. However, this 53 may well be a juvenile feature in the instance of BRSMG Ce12785 because 54 55 the distal tip of the posterolateral premaxillary process of the lectotype 56 57 (Fig. 12A) has a rugose surface texture. 58 59 60
39 Zoological Journal of the Linnean Society Page 40 of 215
1 2 3 Comparisons. In the only other reasonably well preserved and 4 5 morphologically comparable, stratigraphically early (Lower Toarcian) and 6 7 systematically basal thyreophoran Emausaurus ernsti (Haubold 1990: figs 8 2, 4) the left premaxilla is well preserved. Five premaxillary tooth 9 10 positions have been reported, as described in Scelidosaurus. The 11 12 premaxilla is not laterally flared and tapers anteriorly. Unfortunately, the 13 anterior tip and much of the dorsal portion of this bone is missing. 14 15 However, the posterolateral process is preserved; it is transversely 16 17 flattened and becomes broad and spatulate posteriorly, unlike the 18 morphology seen in Scelidosaurus. Among eurypodans the basal 19 20 stegosaur Huayangosaurus is reported to possess seven premaxillary 21 For Review Only 22 teeth (Sereno & Dong 1992: fig. 6A); however, in the more derived 23 Stegosaurus the premaxilla is edentulous. Ankylosaurid premaxillae are 24 25 edentulous and some have large subsidiary foramina positioned dorsally 26 27 and laterally (Maryańska 1977: fig. 3). In contrast several nodosaurids: 28 Pawpawsaurus (Lee 1996), Silvisaurus Eaton 1960), Gargoyleosaurus 29 30 (Carpenter, Miles & Cloward 2000) and Hungarosaurus (Ösi 2005) are 31 32 reported to possess premaxillary teeth (as many as seven in the case of 33 Gargoyleosaurus). 34 35 36 Maxilla (Figs 8, 12-14). In the lectotype, NHMUK R1111 (Fig. 12) the left 37 maxilla (Mx) is disarticulated, but truncated by an oblique abrasion 38 39 surface anteriorly; this latter surface reveals cortical and trabecular bone 40 41 exposed by erosion. The partially prepared but articulated skull BRSMG 42 LEGL 0004 (Fig. 14), reveals most of the superficial features of the 43 44 maxilla, and these are echoed in the smaller, partly disarticulated 45 46 specimen (BRSMG Ce12785 – Fig. 13A). 47 48 49 50
40 Page 41 of 215 Zoological Journal of the Linnean Society
1 2 3 recess in the posterodorsal midline between the premaxillae (cf. Haubold 4 5 1990). Viewed externally, the anterior end of the maxilla has a smoothly 6 7 rounded edge that contacts the premaxilla (Fig. 8); there is no obvious 8 offset between the alveolar margins of these two bones as is also the case 9 10 in Emausaurus (Haubold 1990) and Huayangosaurus (Sereno & Dong 11 12 1992). On the lateral surface of the maxilla, close to the anteroventral tip 13 of the maxilla, there is a prominent, slit-like foramen (Figs 8 & 14, amf). 14 15 The dorsolateral surface of the maxilla is laterally compressed, forming a 16 17 thin lamina (lacrimal lamina) that forms the lateral wall of the nasal 18 cavity; it displays exostotic bone on its external surface that indicates the 19 20 probable existence of overlying keratinous scales. The dorsal edge of the 21 For Review Only 22 maxilla is obliquely inclined and has a slightly sinuous contact with the 23 premaxilla. In BRSMG Ce 12785 (Fig. 13B, gr) the edge of the lamina is 24 25 grooved to provide a very narrow, trough-like suture with the 26 27 posterolateral premaxillary process. Beyond its premaxillary contact the 28 dorsal edge of the maxilla continues posterodorsally along a suture with 29 30 the nasal and then meets the lacrimal at its posterior edge. Beneath and 31 32 anterior to the suture line with the lacrimal there is an extensive scarf 33 suture with the lateral surface of lacrimal (Fig. 12, ls). The apex of the 34 35 maxilla is rounded and the suture line between maxilla and lacrimal 36 37 descends in a slightly sinuous fashion ventrally; this suture line terminates 38 at a notch that forms the smooth anterior margin of the antorbital 39 40 fenestra (Fig. 12A, af) located at the rear of the antorbital fossa (aof). At 41 42 this point, the antorbital fenestra forms a smooth-surfaced channel that 43 curves anteromedially, beneath the base of the lacrimal lamina of the 44 45 maxilla, and opens out on to the lateral floor of the nasal cavity (Fig. 12B, 46 47 af). In lateral view, this foramen widens on to the external surface and 48 merges with the smooth medial wall of the shallow antorbital fossa. 49 50 51 The antorbital fossa (Figs 8, 12-14, aof) has an ovoid outline in 52 lateral view; it is marked off from the surrounding area of the maxilla 53 54 because it has a smooth, concave medial wall that is delineated ventrally 55 56 by an everted and rugose ledge. The latter also forms the dorsal margin 57 of the buccal emargination (be) and, farther posteriorly, a laterally 58 59 everted shelf that forms the jugal suture. Anteriorly, along the ventral 60
41 Zoological Journal of the Linnean Society Page 42 of 215
1 2 3 margin of the antorbital fossa, this rough edge develops into a smooth, 4 5 sharp-edged rim that curves tightly, but smoothly, posterodorsally in 6 7 order to define the upper margin of the antorbital fossa. The anteromedial 8 corner of the antorbital fossa forms a pocket (sediment-filled) that 9 10 indicates the position of the anterior antorbital foramen (aaf). A clear, 11 12 narrow ridge marks the dorsal margin of the fossa anteriorly but this 13 gradually fades away as the ridge subsides and merges with the base of 14 15 the dorsal (lacrimal) lamina of the maxilla. The posterior border of the 16 17 antorbital fossa is closed by a bridge formed by the jugal process of the 18 lacrimal; the latter forms an oblique overlapping suture against the 19 20 anterior tip of the jugal (Fig. 14). The entire lacrimal (sub-orbital) process 21 For Review Only 22 of the jugal is sutured (ventromedially) against a laterally flaring ledge 23 formed along the entire posterodorsal surface of the maxilla (Fig. 12A, js). 24 25 Immediately beneath the area occupied by the antorbital fossa, the 26 27 body of the maxilla reaches its maximum transverse thickness. Beneath 28 29 the rugose ledge that forms the upper boundary of the buccal 30 emargination, the external surface of the maxilla is punctured by a 31 32 number of larger foramina that lead into a large sinus within the body of 33 34 the maxilla. 35 36 The portion of the maxilla immediately posterior to the antorbital 37 38 fossa forms an oblique, shelf-like structure that underlies the anterior 39 process of the jugal. The dorsal surface of this shelf is rugose and 40 41 transversely concave, forming an elongate trough into which the ventral 42 43 surface of the maxillary (sub-orbital) process of the jugal appears to fit 44 snugly (NHMUK R1111). The posterior extremity of the maxilla is acutely 45 46 pointed, lies beneath the main body of the jugal and is wedged between 47 48 the jugal dorsally and the ectopterygoid medially (Fig. 12B, ecs). The 49 posterior ventral margin of the maxilla is sharp-edged beneath its contact 50 51 with the jugal and is angled anteroventrally in a slightly concave curve. 52 53 This edge is interrupted by a small pit, which appears to be an incipient 54 alveolus (plugged with calcite in the lectotype); this curved edge is finally 55 56 interrupted by the concave scalloped edge and alveolus representing the 57 58 last (preserved) maxillary tooth position (18th). 59 60
42 Page 43 of 215 Zoological Journal of the Linnean Society
1 2 3 In lateral view, the entire alveolar border of the maxilla is 4 5 scalloped, where the bone of each alveolus is molded to form a collar 6 7 around the root of each tooth. The alveolar margin is mildly sinuous along 8 its length – slightly arched anteriorly and bowed ventrally posteriorly – 9 10 mirroring the profile of the mandibular dentition; this configuration is seen 11 12 in all three examples (Figs 12-14) and is not an artefact of preservation. 13 Medially, the alveolar wall of the maxilla shows evidence of tooth 14 15 replacement along the entire length of the tooth row. A row of ‘special’ 16 17 foramina (Edmund 1960, Horner et al. 2004) is located parallel to, and 18 just a few millimetres above, the alveolar margin of each functional tooth 19 20 (Fig. 12B, sf – see Fig. 44). Similar foramina were reported in a variety of 21 For Review Only 22 ankylosaurs and stegosaurs (Edmund 1960: fig. 48). These foramina do 23 not form an arched, array as seen in more derived forms such as 24 25 ornithopods and ceratopians. In almost all instances, the foramina show 26 27 the presence of a replacement crown (see Figure 44 for more detail); 28 these range from small emergent crown tips through various stages to 29 30 emerged replacements. Unusually, among ornithischians, the widest part 31 32 of the emergent crowns is adjacent to the regions of destruction of the 33 medial wall of the alveolus. This creates a visual equivalent of incised 34 35 interdental plates at intervals along the dentition – see the Dentition 36 37 section later in this article. Interdental plates are more commonly found 38 on the alveolar walls in theropod dentitions (Holtz, Molnar & Currie 2004). 39 40 41 The inner surface of the main body of the maxilla is smooth and 42 bulges slightly medially, above the dental margin, before curving laterally 43 44 to form a rounded shelf (Fig. 12B, mxs) that partially floors the nasal 45 46 passage before contacting the lacrimal, nasal and premaxilla. In 47 combination, all these bones form the floor, lateral and dorsal walls of the 48 49 snout/nasal passages. The lacrimal lamina is triangular and marked 50 51 medially by a concave area (Fig. 12, ls – shaped like a thumbprint) that is 52 covered with low ridges indicating that connective tissue secured this 53 54 surface against the lacrimal in an extensive scarf suture. This area is 55 56 marked off from the main body of the maxilla by the smooth notch (Fig. 57 12B, af) that forms the anterior edge of the antorbital fenestra. Posterior 58 59 to the lacrimal lamina, the maxilla attains its maximum width. Its medial 60
43 Zoological Journal of the Linnean Society Page 44 of 215
1 2 3 and dorsal surfaces bear closely packed ridges running sub-parallel to one 4 5 another. This large area represents the sutural attachment for the 6 7 palatine. The palatine has a dorsolateral process that caps the maxilla and 8 contacts the jugal, as well as an extensive ventral lamina (Fig. 12B, mwp) 9 10 that covers the medial wall of the maxilla. Posteriorly, the border of the 11 12 maxilla is reflected sharply laterally and displays more sutural scarring, 13 this time for the attachment of the anterolateral edge of the ectopterygoid 14 15 (ecs); the latter bone connects the body of the pterygoid to the maxilla 16 17 and jugal. 18 19 The dorsal surface to the maxilla, immediately posterior to the 20 lacrimal lamina, is marked by a large elliptical trough/fossa (Fig. 27F, 21 For Review Only 22 mxf) that communicates with a row of large foramina that emerge on the 23 24 lateral surface of the maxilla. This fossa is sandwiched between the 25 elongate sutural surface for attachment of the jugal, and the broad mound 26 27 that forms the sutural surface for the palatine. The lateral edge of the 28 29 base of the palatine forms a narrow suture against the medial edge of the 30 jugal and jugal process of the lacrimal, forming a bridge across the central 31 32 part of the maxillary fossa. 33 34 Three areas on the external surface of the maxilla are uncoated 35 36 with either granular or fibrous exostoses: a posterior portion of the 37 38 lacrimal lamina dorsal to the antorbital fossa (Figs 8, 14), the medial wall 39 of the antorbital fossa itself and the buccal emargination; these areas 40 41 would have been lined with soft epidermal tissue. The small unadorned 42 43 patch of bone on the lacrimal lamina is difficult to explain, but may have 44 some relationship with the antorbital fossa that is positioned directly 45 46 beneath. 47 48 The dorsal surface of the maxilla, between the boundary to the 49 50 buccal emargination and the premaxilla, is coated with low-relief bony 51 52 strands; these are mostly arranged longitudinally, but farther posteriorly, 53 near the lacrimal, a pattern of curved ridges develops. A similar pattern of 54 55 ‘swirly’ ridges is also seen just dorsal of this area on the nasal 56 57 (particularly well displayed on the nasals of BRSMG Ce12785 – Fig. 15). 58 The prominent ledge beneath the antorbital fossa is rugose; this rugosity 59 60 may indicate the anchorage of the ornithischian equivalent of a
44 Page 45 of 215 Zoological Journal of the Linnean Society
1 2 3 mammalian buccinator muscle. The rugosities extend posteriorly as 4 5 irregular strands of exostotic bone on the lateral surface of the lacrimal 6 7 process of the jugal beneath the orbit (Fig. 14). 8 9 Comparisons. In Emausaurus ernsti (Haubold 1990: fig. 5), the 10 maxilla has, overall, a similar morphology to that seen in Scelidosaurus. A 11 12 buccal emargination is well developed, the sinuous dental profile is the 13 14 same and the dentition appears to be bowed medially (better seen in the 15 dentary than the maxilla); an oblique, flared sutural shelf for attachment 16 17 of the jugal is similar to that described in Scelidosaurus. An extensive 18 19 rugose area on the posteromedial wall of the maxilla (Haubold 1990: taf. 20 III: 4) indicates the attachment of a medial lamina of the palatine, as 21 For Review Only 22 seen clearly in Scelidosaurus. Additionally, the anteromedial 23 24 (premaxillary) process is well preserved; this region is not well exposed in 25 all known specimens of Scelidosaurus. However, in striking contrast, the 26 27 preserved margin of antorbital fossa (Haubold 1990: taf. III, 2) suggests 28 29 that this is considerably larger than that seen in Scelidosaurus and more 30 reminiscent of that seen in basal ornithischians such as Lesothosaurus and 31 32 heterodontosaurids (Porro et al. 2015; Norman et al. 2011). Among 33 34 eurypodans the maxilla is excavated laterally and backs a deep buccal 35 36 emargination (covered by prominent cheek osteoderms in ankylosaurs); 37 however, no antorbital fossa has been recorded. In the basal stegosaur 38 39 Huayangosaurus (Sereno & Dong 1992: fig. 6) the maxilla resembles, in 40 41 its general proportions those of Scelidosaurus and may well have a large 42 elliptical foramen close to the premaxilla-maxilla suture; it also displays a 43 44 triangular antorbital fossa. In the more derived Stegosaurus the maxilla is 45 46 lower and more elongate compared to that of Scelidosaurus (Gilmore 47 1914: pl. 5); it also has a long anterior process that underlies the 48 49 premaxilla; a small antorbital fossa is present and resembles that seen in 50 51 Scelidosaurus. 52 53 54 55
45 Zoological Journal of the Linnean Society Page 46 of 215
1 2 3 Nasal (Figs 8,9,14-16). The nasals are visible in ventral view in the 4 5 lectotype skull and are smooth and shallowly arched; their dorsal surfaces 6 7 are roughly textured but the skull as preserved is far too fragile to 8 examine closely (see Fig. 4D). Fortunately the nasals are visible externally 9 10 on the referred skull (BRSMG LEGL 0004: Figs 14, 16), whereas in the 11 12 intermediate-sized specimen (BRSMG Ce12785) both nasals are preserved 13 separated from the remainder of the skull and folded together about the 14 15 dorsal mid-line and are nearly complete (Fig. 15). 16 17 The nasals are elongate, transversely arched plates, broad 18 19 posteriorly and tapering gradually to a bluntly rounded anterior tip. Just 20 posterior to the anterior tip, each nasal bears a small facet on its lateral 21 For Review Only 22 surface that must represent the sutural contact for the median dorsal 23 24 processes of the premaxilla (Fig. 15A,B, pms). Unusually among 25 ornithischians the conjoined tips of the nasals interpose themselves 26 27 between the posterior ends of the premaxillary processes. The nasals 28 29 meet in the midline along an elongate, irregular, tongue-in-groove suture 30 (Fig. 15B, ns). The ventral edge of the anterior tip of the nasal is 31 32 transversely rounded and gentle arched and meets the posterolateral 33 34 process of the premaxilla, pinching off the border of the external naris. 35 36 The suture with the premaxilla bows ventrally along its length following 37 the dorsal profile of the premaxilla and then meets the lacrimal lamina of 38 39 the maxilla. In addition, this suture is stepped medially to produce a 40 41 curtain-like flange (Fig. 15, mf) that lies against the medial surface of the 42 posterolateral process of the premaxilla and probably backs the lacrimal. 43 44 Posteriorly, the external sutural margin of the nasal curves dorsomedially 45 46 forming an embayment where it contacts the prefrontal (Figs 15A, pfs, 47 16). However, the ventral aspect of the nasals of the lectotype reveals 48 49 that this latter suture was scarf-like, forming a continuation of the medial 50 51 flange that underplates the prefrontal. The well-preserved isolated nasals 52 (Fig. 15) as well as those of the lectotype are smooth internally; this 53 54 contrasts with the external surface, which is coated with an array of 55 56 linear, irregular and curved ridges, and some strand-like exostoses (Figs 57 8, 9, 15, 16). The posterior suture with the frontals is scarf-like, but here 58 59 60
46 Page 47 of 215 Zoological Journal of the Linnean Society
1 2 3 the nasals overlap the frontals, reversing the pattern seen on the lateral 4 5 side of the snout. 6 7 Comparisons. In Emausaurus the nasals are not preserved. Among 8 9 eurypodans more generally, ankylosaurs have nasals that are obscured by 10 large caputegulae (with the exception of very juvenile examples e.g. 11 12 Burns et al. 2011) whereas in the basal stegosaur Huayangosaurus, as 13 14 well as the more derived Stegosaurus, nasals resemble more closely in 15 general shape those seen in Scelidosaurus. However, they are not so 16 17 obviously encrusted with exostoses and their anterior tips are illustrated 18 19 lapping laterally against the sides of a common median dorsal 20 premaxillary process (as seen more typically in ornithischians). 21 For Review Only 22 23 24 25
47 Zoological Journal of the Linnean Society Page 48 of 215
1 2 3 parietal and skirting the anterior margin of the supratemporal fenestrae, 4 5 before butting up against the blunt medial processes of the postorbitals. 6 7 Anterior to the postorbital suture each frontal contacts the middle 8 supraorbital (mso) forming a curved and crenulate scarf suture that cuts 9 10 back anteromedially across the roof of the skull, before extending 11 12 anteriorly to meet the edge of the prefrontal. The margin then turns 13 medially to contact the posterior edge of the nasals on either side of the 14 15 mid-line. The nasals and frontals meet along an irregular scarf-suture. The 16 17 nasals overlap the dorsal surface of the frontals. Ventrally, the outline of 18 the frontals and relationship to most of the surrounding skull bones can be 19 20 discerned. The most noteworthy features are the smoothness of the bone 21 For Review Only 22 surface (in comparison with the dorsal surface texture) and the ‘hourglass’ 23 ridges that mark the upper internal edges of the ethmoid cartilages that 24 25 walled the orbital cavities, leaving a median channel that represents the 26 27 roof of the passage for the olfactory lobes. 28 29 The bony textures preserved on the dorsal surface of the frontals 30 (see Figs 8, 16) are striking. Adjacent to the parietal suture the posterior 31 32 portion of the frontal plate bears a patch of longitudinal strands of bony 33 34 tissue. Anterior to this patch a median trough subdivides the frontals and 35 36 the surrounding area is covered by densely granular-textured bone. The 37 trough flattens out anteriorly and the granular texture becomes one of 38 39 strands of radiating bony fibres that extend beyond the nasal suture. 40 41 These bony fibres extend laterally as a sort of ‘zone’ surrounding the 42 central (granular) area of the frontal plate. These radiating fibres extend 43 44 toward all the peripheral skull roofing bones. Fibrous looking strands, as 45 46 well as irregular arrays of bony fibres or grains and some bony pustules 47 coat the external surfaces of all adjacent skull-roofing bones, except in 48 49 the areas of bone in which muscles were either attached directly or where 50 51 bone forms the smooth inner walls of the adductor chambers. The texture 52 on the external surface of the skull bones of Scelidosaurus most nearly 53 54 resembles that seen on the skulls of chelonians that are covered in life by 55 56 a mosaic of large, keratinous scales (pers. colln. Chelonia mydas). 57 58 Comparisons. In Emausaurus the structure of the frontals is not 59 well preserved. In outline, its proportions resemble those of 60
48 Page 49 of 215 Zoological Journal of the Linnean Society
1 2 3 Scelidosaurus, but the same is true of many ornithischians. There is no 4 5 description of surface textures equivalent to those described in 6 7 Scelidosaurus, but this may also be an artefact of its preservational 8 condition (e.g. the palpebral is one of the few skull bones to exhibit 9 10 surface texturing equivalent to that seen more extensively on the skull of 11 12 Scelidosaurus). It appears, based upon the reconstruction (Haubold 1990: 13 fig. 2), that the frontal forms the dorsal orbital margin, even though it is 14 15 flanked laterally by the palpebral. The presence of a middle supraorbital 16 17 (or its cartilaginous equivalent) between the frontal margin and the 18 palpebral is suspected, and this suspicion is reinforced by the structure of 19 20 the dorsal orbital margin of the postorbital. The latter displays an oblique 21 For Review Only 22 rugose facet for attachment of a supraorbital. The frontals of the 23 eurypodan Stegosaurus retain a midline suture but otherwise have similar 24 25 relationships to the surrounding skull bones (Gilmore 1914: pl. 6), despite 26 27 the skull being narrower and much more anteroposteriorly stretched. In 28 contrast, the basal stegosaur Huayangosaurus has been illustrated with a 29 30 frontal that is excluded from contact with the middle supraorbital by the 31 32 prefrontal and postorbital (Sereno & Dong 1992: fig. 6B). 33 34 Parietals (Figs 9, 16). The parietals (P) are completely fused together to 35 36 create a saddle-shaped structure (Figs 4D, 9, 16). The lectotype has the 37 parietal plate exposed dorsally (Fig. 4D), but it was evidently substantially 38 39 eroded prior to collection and has also been somewhat degraded after 40 41 long periods of acid immersion. The referred skeleton (BRSMG LEGL 0004: 42 Fig. 16) has a well-exposed parietal plate, although it is visible only in 43 44 dorsal aspect. 45 46 On either side of the midline, the sides of the parietal plate are 47 48 concave and form the dorsal portion of the sloping internal walls of the 49 adductor chamber. These surfaces are smooth to allow movement of the 50 51 adductor musculature that was attached along the margins of the 52 53 supratemporal fenestra. The ventral edges of the parietal plate contact 54 the proötic and opisthotic along a horizontal suture line. The internal 55 56 structure of these sutural surfaces is visible as counterparts on the 57 58 corresponding surfaces of the laterosphenoid, opisthotic and supaoccipital 59 60
49 Zoological Journal of the Linnean Society Page 50 of 215
1 2 3 in the lectotype (Figs 31A, 35B); these include mortice-and-tenon-like 4 5 structures that would have reinforced the suture with the skull roof. 6 7 The sagittal crest(s) is (are) rugose and prominent but, slightly 8 9 unusually, comprise a pair of crests separated by a narrow median gully 10 (Figs 9, 16). The midline gully has a smooth surface, whereas the raised 11 12 parasagittal crests are rough-textured – no doubt reflecting the 13 14 ligamentous attachment of medial portions of the adductor musculature. 15 Anteriorly these raised crests diverge to form the anteromedial margins of 16 17 the supratemporal fenestrae. The dorsal area between these divergent 18 19 ridges flattens out (becoming mildly concave transversely) and terminates 20 anteriorly along an irregular (coarsely crenulate and interdigitate) open 21 For Review Only 22 suture line that is concave anteriorly where it meets the central portion of 23 24 the frontal plate (Fig. 16). Anteriorly, the dorsal surface between the 25 divergent parasagittal crests is roughened by the presence of 26 27 anteroposteriorly oriented superficial bony strands. Posteriorly, the 28 29 sagittal crests diverge above the occiput and produce a pair of diverging 30 gullies that cap the mediodorsal edge of the occiput and overlap the 31 32 squamosals. Immediately anteroventral to these crested edges (and lining 33 34 the posterior face of the adductor cavity) lie thin curved laminae of the 35 36 squamosals. The lateral extent of the divergent parietal wings is 37 uncertain, but they clearly capped the supraoccipital medially, the 38 39 paroccipital farther laterally and backed the medial part of the ‘shoulder 40 41 region’ of each squamosal. The robustness of these occipital parietal wings 42 makes it seem probable that one of their roles was to offer support and/or 43 44 anchorage for the occipital osteoderm horns (Fig. 16, oc.ost). 45 46 Comparisons. Parietals are not preserved in Emausaurus. Among 47 48 eurypodans the parietal is poorly known but forms a broad saddle-shaped 49 cap to the braincase in stegosaurs and nothing is known of the form of the 50 51 sagittal crest (or if it even existed). The parietal plate of mature 52 53 ankylosaurs is covered by dermal ossifications (although parietals are 54 unfused in a juvenile Pinacosaurus – Maryańska 1977) and cannot be 55 56 usefully compared to Scelidosaurus. 57 58 Lacrimal (Figs 8, 12-14). The external surfaces of both left and right 59 60 lacrimals (La) of the lectotype (NHMUK R1111: Fig. 12A) appear coated
50 Page 51 of 215 Zoological Journal of the Linnean Society
1 2 3 by rough-textured exostotic bone that partly obscures sutural contacts 4 5 with adjacent bones. The internal surface of the disarticulated left lacrimal 6 7 (Fig. 12C) helps to clarify some its internal structure. A left lacrimal is 8 visible in external view on the articulated skull (BRSMG LEGL 0004: Fig. 9 10 14) and its sutural relationships are a little clearer. The left and right 11 12 lacrimals of the intermediate-sized specimen (BRSMG Ce12785: Fig. 13A) 13 are also preserved in partial articulation with their respective maxillae. 14 15 The lacrimal has comparatively little external exposure on the skull. 16 17 It is overlapped anteriorly by the lacrimal lamina of the maxilla and is 18 19 overlapped by the prefrontal posterodorsally; it is also partly overlain by 20 the base of the large palpebral (Fig. 14, Pp). The laterally exposed surface 21 For Review Only 22 of the disarticulated lacrimal of the lectotype (Fig. 12) and referred skull 23 24 (Fig. 14) is partly coated with bony strands. In the lectotype (Fig. 12A) 25 the osseous coating that covers the posterior portion of the lacrimal ends 26 27 abruptly at an oblique line and step. Anterior to the step, the external 28 29 surface of the lacrimal is smooth and seems to represent an exposed part 30 of the scarf suture with the lacrimal lamina of the maxilla – the lamina of 31 32 the maxilla appears to have partly broken away. The equivalent area in 33 34 the articulated referred skull (Fig. 14) also shows the posterior half of the 35 36 lateral wall of the lacrimal covered by strand-like superficial bony fibres. 37 The strands terminate abruptly (in the area of the suture between maxilla 38 39 and lacrimal) and the external surface of the lamina of the maxilla 40 41 becomes perfectly smooth and unadorned for a short distance before 42 becoming once again coated in rough-textured curved ridges. 43 44 Posteroventrally, the lacrimal produces a tapering, curved jugal 45 46 process (Fig. 12C, jp) that forms the anteroventral margin of the orbit. 47 48 The posterior tip of the jugal process is arched ventrally and twisted 49 axially, and wraps around the mediodorsal surface of the finger-like 50 51 anterior end of the jugal (Figs 12A, 21). The medial edge of the jugal 52 53 process also has a thin contact with the lateral edge of the base-plate of 54 the palatine. 55 56 57 The foramen for the nasolacrimal duct (Fig. 12, ld) is transversely 58 compressed and positioned quite low down on the posterior (orbital) 59 60 surface of the lacrimal. The nasolacrimal duct is enclosed within a narrow
51 Zoological Journal of the Linnean Society Page 52 of 215
1 2 3 sinus-like chamber in the body of the lacrimal. Medial to the nasolacrimal 4 5 foramen there is a thin, transverse orbital wall; farther dorsally this 6 7 partition wall becomes thicker as it approaches the area of the suture with 8 the prefrontal but this area is badly broken in the lectotype lacrimal. Just 9 10 anterior to the orbital partition there is a recess floored by a short, blunt, 11 12 anteromedial extension of the lacrimal; this would have been roofed by 13 the prefrontal/nasal in the articulated skull. Anterior to the orbital 14 15 partition wall, the lacrimal comprises two laminae: the lateral and more 16 17 substantial one fits snugly against the medial surface (ls) of the dorsal 18 lamina of the maxilla; the medial lamina (Fig. 12C, mw) of the lacrimal is 19 20 a thin partition wall that lined the nasal passage. The central area of this 21 For Review Only 22 medial lamina appears to be pierced by a fenestra, but this is in fact a 23 fractured area in the thin bony wall (through which can be seen the cavity 24 25 for the nasolacrimal duct). Farther anteriorly, the nasolacrimal duct opens 26 27 into the nasal cavity via a slot-like anteriorly directed opening (Fig. 13C, 28 ld). 29 30 Externally the lacrimal seems not to have contacted the nasal; 31 32 however, the nasal has a long, recessed, posteroventral curtain-like 33 34 sutural surface that supported (backed on to) the premaxilla-maxilla 35 36 contact and this may well have overlapped the anterodorsal portion of the 37 lacrimal. Posterodorsally, the lacrimal is overlapped/capped by the ventral 38 39 ‘foot’ of the prefrontal. Posteriorly, the dorsolateral portion of the lacrimal 40 41 bears a small shoulder (Fig. 13) for attachment of the prefrontal; the 42 latter bone wraps around on to the medial edge of the lacrimal so that the 43 44 prefrontal is locked in place on the anterior orbital margin. 45 46 Comparisons. The lacrimal of Emausaurus is incomplete, but 47 48 includes a long, curved jugal process that evidently wrapped itself around 49 the anterior tip of the jugal in a similar fashion to that seen in 50 51 Scelidosaurus. The lacrimal also formed the posterodorsal margin of the 52 53 antorbital fossa and, unlike Scelidosaurus, backed the dorsal portion of 54 the antorbital fossa and bordered the dorsal half of the antorbital 55 56 fenestra; and its sutural relationship with the maxilla and prefrontal are 57 58 not clear. Among more derived eurypodans such as Huayangosaurus and 59 Stegosaurus the lacrimal resembles that of Scelidosaurus in its superficial 60
52 Page 53 of 215 Zoological Journal of the Linnean Society
1 2 3 characters, though it does have a short pillar-like jugal process that forms 4 5 the posterior boundary to the antorbital fossa. No particularly fruitful 6 7 comparisons can be drawn with currently known ankylosaurs; the juvenile 8 Pinacosaurus skull (Maryańska 1977: fig. 3) reveals a superficially oblong 9 10 bone with a long jugal process, but there is no antorbital fossa for it to 11 12 bridge. 13 14 Prefrontal (Figs 8, 9, 14, 16, 17). The left prefrontal (Pf) of NHMUK 15 R1111 is crushed and broken, and was mostly destroyed by mishandling 16 17 of the acid-prepared skull. Its external features are obscured by a 18 19 superficial layer of bony tissue that may also be a portion of the fused 20 ‘footplate’ (basal body) of the overlying palpebral bone (Fig. 12, pp?). The 21 For Review Only 22 right prefrontal of the lectotype is visible internally, so its general shape 23 24 and sutural relationships can be determined. The external (dorsal) surface 25 of the prefrontal is well-displayed in the referred skull (BRSMG LEGL 26 27 0004: Fig. 16) where it is seen to be covered by an irregular array of bony 28 29 strands that extend to the lateral surface; however, its lateral surface is 30 mostly obscured by the overlying palpebral (Pp). An isolated, partial, right 31 32 prefrontal is preserved in the intermediate-sized specimen (BRSMG 33 34 Ce12785: Fig. 17A,E). 35 36 37 38
53 Zoological Journal of the Linnean Society Page 54 of 215
1 2 3 distinctly rugose (Fig. 17A); following this margin as it curves ventrally 4 5 (toward the lacrimal) the orbital margin becomes more smoothly rounded 6 7 and then thickens to form a mound that culminates in a relatively 8 prominent lateral facing, roughly triangular facet against which the 9 10 footplate of the palpebral was anchored (Fig. 17A, ppf). Medially (Fig. 11 12 17E), the prefrontal develops a transverse partition wall that extends from 13 the frontal suture on the mediodorsal surface, above the orbital cavity and 14 15 descends across the front of the orbital cavity as a sharp edge that widens 16 17 as it approaches the equivalent partition wall on the medial side of the 18 lacrimal. 19 20 Comparisons. In Emausaurus, the prefrontal has a broadly similar 21 For Review Only 22 shape to that seen in Scelidosaurus (Haubold 1990: fig. 2); its 23 24 morphology is partly obscured by a large shield-like palpebral that is 25 clearly sutured to the lateral edge of the prefrontal alone (Haubold 1990: 26 27 figs 2, 6). Eurypodans (as well as basal ornithischians more generally) 28 29 exhibit rather similarly proportioned prefrontals. 30 31 32 33
54 Page 55 of 215 Zoological Journal of the Linnean Society
1 2 3 between the edge of the middle supraorbital and palpebral. The palpebral 4 5 has a generally rugose, irregular bony texture. 6 7 Comparisons. In Emausaurus (Haubold 1990: figs 2, 6) the 8 9 palpebral is well preserved. It exhibits an extensive sutural surface 10 medially showing that it was very firmly attached to the lateral margin of 11 12 the prefrontal and that its base did not encroach upon the lacrimal. This 13 14 bone is roughly triangular in outline, with a slightly concave orbital 15 margin. Its external surface is markedly sculptured and, when articulated 16 17 with the prefrontal, there is a tapering gap between the orbital margin of 18 19 the frontal and the inner margin of the palpebral. The orbital margin of 20 the frontal is rugose and bevelled; this perhaps indicates an area for the 21 For Review Only 22 attachment of the middle supraorbital (whether ossified or unossified) that 23 24 spanned the gap between the frontal and palpebral. Among more derived 25 eurypodans the stegosaurs Huayangosaurus and Stegosaurus possess 26 27 similarly robust palpebrals attached to the prefrontal and curving 28 29 backwards parallel to the orbital margin. Gilmore (1914: pl. 6) also 30 illustrates an unossified gap between the palpebral and adjacent middle 31 32 supraorbital. The juvenile ankylosaur Pinacosaurus (Maryańska 1977: figs 33 34 2, 3) displays a broadly similar pattern. 35 36 Middle Supraorbital. In NHMUK R1111, an angular, slightly domed, 37 38 bone is visible on both sides of the dorsal aspect of the skull roof (Fig. 39 4D). The bone is positioned on the dorsal orbital margin, lateral to the 40 41 frontal and forms coarse interdigitate sutures with the prefrontal, frontal 42 43 and postorbital. This bone is seen clearly in the dorsal view of the slightly 44 smaller, referred specimen (BRSMG LEGL 0004: Fig. 16). Its dorsal 45 46 surface is roughened by a coating of bony strands, whereas its shallowly 47 48 domed ventral (orbit) wall is quite smooth. 49 50 The intermediate-sized specimen (BRSMG Ce12785: Fig. 17) 51 52 preserves an isolated right middle supraorbital that articulates loosely 53 against the postorbital and lies anteromedial to the tent-shaped posterior 54 55 supraorbital. This small, inverted dished bone has an approximately 56 57 pentagonal outline; its sutural edges are rough and slightly thickened and 58 the entire bone is vaulted because it roofs a portion of the orbital cavity. 59 60 The dorsal surface of this bone does not bear any irregular, superficial,
55 Zoological Journal of the Linnean Society Page 56 of 215
1 2 3 bony fibres and this is probably a reflection of the comparative 4 5 ontogenetic immaturity of the skeleton. Nevertheless, its dorsal surface is 6 7 pock-marked by a number of foramina – perhaps hinting at elevated 8 levels of biological activity associated with the growth of overlying tissue. 9 10 In contrast, the ventral (orbital) surface appears comparatively smooth 11 12 and unmarked. The orbital margin, which lies adjacent to the medial edge 13 of the distal process of the palpebral, is round-edged but slightly rugose. 14 15 Comparisons. In Emausaurus a middle supraorbital is not reported 16 17 as being present. Eurypodans generally exhibit a similar pattern of 18 19 circumorbital bones with both stegosaurs and ankylosaurs possessing 20 middle supraorbitals that widen the roof of the orbital cavity. 21 For Review Only 22 23 Postorbital (Figs 8, 9, 16-18). In the lectotype (NHMUK R1111), the 24 postorbitals (Po) are fragmentary because they appear to have been 25 26 damaged and almost completely removed by erosion. The now 27 28 disarticulated portion of the skull comprising the complete jugal, 29 quadratojugal and quadrate displays just the jugal process of the 30 31 postorbital and its extensive oblique suture to the jugal (Fig. 21A). The 32 33 dorsal part of the jugal-postorbital suture is open; nevertheless, the distal 34 tip of the jugal process of the postorbital appears to be fused to the jugal 35 36 (apparently bound by strands of exostotic tissue). The orbital rim of this 37 38 process is concave and, viewed anteriorly, there is a minor degree of 39 interdigitation between the jugal and postorbital (Fig. 21C). There is also 40 41 a line of foramina (vas) lying medial and parallel to the lateral edge of the 42 43 orbital margin of the postorbital; these foramina seemingly mark the 44 existence of an interface between the lateral wall of the postorbital and a 45 46 crater-shaped area of osteodermal tissue plastered against the lateral 47 48 surface of the postorbital (see also Figs 14, 17A,C). 49 50 The orbital surface of the postorbital is smooth and slightly 51 52 concave. The medial margin of the internal orbital wall of the postorbital 53 forms a raised ridge that marks the incomplete bony partition (probably 54 55 completed by connective tissue) between the orbital and adductor 56 57 chambers. Posteromedially, the jugal process is smooth and flat and its 58 ventral edge forms a slightly curved diagonal line that represents the 59 60 upper extent of the overlap by the jugal internally. Similar anatomical
56 Page 57 of 215 Zoological Journal of the Linnean Society
1 2 3 features can be seen on the slightly smaller, better-preserved, referred 4 5 specimen (Figs 14, 18, 19). In the latter, the postorbital is triradiate and 6 7 displays a diffuse, but large, depressed patch of rugose bony tissue on its 8 dorsolateral surface; the postorbital of Emausaurus has a similarly 9 10 depressed area on the lateral surface of the postorbital (Haubold 1990: 11 12 Taf. I, 2), but the surface does not appear to be so strongly textured. 13 14 Above the laterally positioned osteodermal crater on the postorbital 15 is a discrete pup-tent-shaped posterior supraorbital that was bound in 16 17 position to the distal end of the palpebral by connective tissue (Fig. 18, 18 19 pso). This suture becomes ossified later in development and may be 20 overgrown by strands of periosteal bone. 21 For Review Only 22 23 The dorsal part of the postorbital-jugal suture is open, as it is in the 24 lectotype. Another curious feature, relating to the superficial layering of 25 26 bone on the skull surface, is visible in the dorsal aspect of the skull (Fig. 27 28 16). It seems that this surface of the postorbital is encrusted with strands 29 of exostotic tissue and an area of little ‘islets’ of superficial bone in the 30 31 area of the skull roof medial to the cap-shaped posterior supraorbital. 32 33 However, in the area surrounding the supratemporal fenestra, where the 34 adductor mandibulae musculature is anchored, the squamosal and parietal 35 36 processes are smooth-surfaced and free of exostotic bone. The small islets 37 38 of superficial bone encroach on the periphery of these areas as part of 39 what has the appearance of an on-going process of competitive spreading. 40 41 42 43 44
57 Zoological Journal of the Linnean Society Page 58 of 215
1 2 3 supratemporal fenestra. The ventral process (jp) curves anteriorly and 4 5 forms a spatulate distal tip; the whole process twists about its long axis 6 7 and displays a large, posteromedially-facing suture (Fig. 19C, j.s). The 8 dorsal edge of this sutural surface is marked by a curved contour line (see 9 10 Fig. 21B). In anterior view, the orbital margin is smooth and shows a line 11 12 of vascular foramina positioned laterally, beyond which there is an area of 13 ornamented bone: an osteoderm. The dorsal portion of this osteoderm 14 15 forms a hook-shaped projection (hk) that protrudes slightly into the 16 17 orbital margin. Immediately above this bony projection, the surface of the 18 postorbital is excavated into a crude, bowl-shaped structure (Fig.19A, 19 20 bwl) that merges into the base of the anterior process of the postorbital 21 For Review Only 22 above the orbital margin. The latter process projects anteriorly in lateral 23 view (Fig. 19A), but when viewed in dorsal aspect (Fig. 19D) this process 24 25 can be seen to bend medially, before ending in a truncated, rough- 26 27 textured sutural surface against the frontal (fs). 28 29 In isolation, this curious superficial structure on the postorbital 30 would be puzzling; fortunately, other material preserved with BRSMG 31 32 Ce12785 provides an explanation. The recess above the postorbital 33 34 osteoderm proves to be a surface for attachment of a pup-tent-shaped 35 36 posterior supraorbital. 37 38 Comparisons. Among eurypodans the fine details of the structure of 39 the postorbital are not well preserved. The basal stegosaur 40 41 Huayangosaurus has been described as possessing a distinctive horn-like 42 43 mound positioned immediately anterior to the supratemporal fenestra 44 (Sereno & Dong 1992: fig. 6A-C, poh); its external surface was also 45 46 covered by a large posterior supraorbital, which looks very similar to that 47 48 seen in Stegosaurus (Fig. 20). The structure of the suture between the 49 postorbital and jugal of Stegosaurus (Gilmore 1914: fig. 5) is very similar 50 51 to that seen in Scelidosaurus; this may also be true for Huayangosaurus, 52 53 but cannot be established with certainty on the basis of the published 54 illustrations. Among ankylosaurs, the postorbital cannot be described 55 56 because it is obscured by osteoderms (caputegulae). 57 58 Posterior supraorbital (Figs 8, 9, 16,17). This small osteoderm (Fig. 59 60 17A,B, pso) fits snugly into a bowl-shaped recess above, and medial to,
58 Page 59 of 215 Zoological Journal of the Linnean Society
1 2 3 the hook-like projection formed by the osteoderm on the lateral surface of 4 5 the postorbital. It projects laterally as a horizontal mound (Fig. 17A,B) 6 7 that is a continuation of the brow ridge formed by the palpebral bone. The 8 anteromedial edge of this supraorbital is truncated and rugose and forms 9 10 a sutural attachment site for the middle supraorbital (mso: Fig. 17B). It 11 12 also bears a short facet anteriorly that is a contact point for the distal end 13 of the palpebral process. The combination of palpebral and posterior 14 15 supraorbital, create a brow ridge (see Figs 8, 18) that would have been 16 17 sheathed by keratin and doubtless served as a shield for the eye, but may 18 also have served as a signaling device. 19 20 Comparisons. The postorbital of Emausaurus bears an oblique ridge 21 For Review Only 22 and partial facet on its anterodorsal corner suggestive of an attachment 23 24 site for a posterior supraorbital (Haubold 1990: taf I, 2); and furthermore 25 an isolated postorbital of Stegosaurus displays an equivalently positioned 26 27 posterior supraorbital (USNM 6645: Fig. 20, pso); a very similarly 28 29 positioned bone is also illustrated on the skull of Huayangosaurus (Sereno 30 & Dong 1992). In ankylosaurids describe a posterior supraorbital 31 32 caputegulum (Arbour & Currie 2016: figs 6, 7, 9, psca) that occupies a 33 34 similar topographic location to the posterior supraorbital osteoderm of 35 36 Scelidosaurus. 37 38 39 40
59 Zoological Journal of the Linnean Society Page 60 of 215
1 2 3 4 5
60 Page 61 of 215 Zoological Journal of the Linnean Society
1 2 3 anterolateral face of the quadratojugal. This feature is not visible on the 4 5 lectotype because of damage in this area (Fig. 21, br), but is well 6 7 preserved in the intermediate-sized specimen (Fig. 13A). The notch on the 8 posterior edge of the jugal is reflected in a complementary wedge-shaped 9 10 ridge on the lateral surface of the quadratojugal. The ventral edge of the 11 12 jugal is comparatively thick and describes a shallow arch 13 anteroposteriorly. 14 15 The external surface of the jugal has superficial patches of granular 16 17 textured and more strand-like arrays of exostotic bone (Figs 13, 14, 21A, 18 19 p.os). 20 21 In the intermediate-sizedFor Review specimen Only(BRSMG Ce12785: Fig. 13) the 22 23 external surface of the lacrimal process is smooth near its tip; however, 24 beneath the orbit longitudinal strands of exostotic tissue lead posteriorly 25 26 into a mound-shaped area with a more granular texture on the central 27 28 body of the jugal (directly beneath the postorbital process). There is then 29 an oblique gap before another area of dense bony strands is found on the 30 31 posteroventral portion of the jugal (adjacent to the quadratojugal suture). 32 33 The larger specimen (BRSMG LEGL 0004: Fig. 14) displays external 34 textures on the left jugal that have a very similar pattern to those seen on 35 36 the intermediate-sized individual; however the larger more granular area 37 38 is found on the deeper posteroventral blade-like part of the jugal. This 39 latter area is raised and displays an array of foramina and what appear to 40 41 be associated vascular grooves (or natural channels between bony 42 43 strands) that all share a generally posteroventral orientation. In contrast 44 other areas, such as the postorbital process and the margin of the lateral 45 46 temporal fenestra have a smooth, cortical bone surface. 47 48 Comparisons. In Emausaurus the left jugal is well preserved 49 50 (Haubold 1990: fig. 7) and has a similar morphology to that seen in 51 52 Scelidosaurus. The lacrimal process is finger-like and wedges between the 53 lacrimal and maxilla, and there is a ectopterygoid facet on its medial 54 55 surface ventral to the orbital margin. The postorbital suture is large, 56 57 pocket-like and faces the orbit; and the quadratojugal process is blade- 58 like and notched along its posterior edge to accommodate a ridge on the 59 60 external surface of the quadratojugal. Little is known of the detailed
61 Zoological Journal of the Linnean Society Page 62 of 215
1 2 3 structure of the jugal in eurypodans. Superficially, the jugal of 4 5 Huayangosaurus and Stegosaurus has an elongate, narrow lacrimal 6 7 process and the body of the jugal is undercut and appears shelf-like. The 8 relationship between jugal and quadratojugal is vague and somewhat 9 10 stylized (Sereno & Dong 1992: fig. 6) but has been given a structure that 11 12 resembles that seen in Scelidosaurus, whereas in Stegosaurus the 13 quadratojugal process of the jugal is somewhat abbreviated. The jugal of 14 15 ankylosaurs also displays and elongate lacrimal process and the body of 16 17 the jugal forms a distinct undercut shelf beneath the orbit (Maryańska 18 1977: fig. A2), but the remainder of its relationships are obscured by 19 20 dermal bone. 21 For Review Only 22 Quadratojugal (Figs 8, 14, 21, 22). The left quadratojugal (Qj) of the 23 24 lectotype is partially obscured because it articulated between the jugal 25 and quadrate and this area is also fractured (Fig. 21). This is also the case 26 27 with the large referred specimen (BRSMG LEGL 0004: Fig. 14). However 28 29 the right quadratojugal of the lectotype is disarticulated and complete 30 (Fig. 22A,B), as also is that of the intermediate-sized specimen (BRSMG 31 32 Ce12785: Fig. 22C,D). 33 34 35 36 37
62 Page 63 of 215 Zoological Journal of the Linnean Society
1 2 3 and, in so doing, spans (and closes off) a nearly circular notch in the 4 5 quadrate that forms a quadrate (paraquadratic) foramen (qf). 6 7 Immediately above the quadrate foramen the posterior margin of the 8 quadratojugal is notched (more obviously so in the lectotype) and twists 9 10 itself around the edge of the jugal wing of the quadrate to help secure its 11 12 position. 13 14 The quadratojugal thus becomes reduced to a tapering splint of 15 bone that is bound to the leading edge of the jugal wing of the quadrate 16 17 (Fig. 22A, qs) until it contacts an equivalent descending process of the 18 19 squamosal (sqc). The apical portion of the left and right quadratojugals 20 exhibit a short, concave sutural surface medially; in the articulated portion 21 For Review Only 22 of the left jugal arch in the lectotype (Fig. 21) the distal tip of the 23 24 squamosal can be seen contacting the quadratojugal. Vascular foramina 25 (Fig. 22C, for) are present on the medial side of the quadratojugal of the 26 27 smaller referred specimen; these are not apparent in the larger specimen 28 29 and may well represent a juvenile feature linked to active bone growth. 30 31 Comparisons. The quadratojugal of Emausaurus is similar in 32 33 morphology to that described above and, though not completely 34 preserved, probably had a similar dorsal contact with the squamosal (as 35 36 suggested by Haubold 1990: fig. 2). Eurypodans, particularly stegosaurs, 37 38 display part of the structure of this bone. In Huayangosaurus this bone is 39 overlain by the jugal anteriorly and has been reconstructed with a 40 41 tapering dorsal (squamosal) process that fails to meet the squamosal; it 42 43 also appears that the quadratojugal extends ventrally to lie adjacent to 44 the condylar surface of the quadrate, as in Scelidosaurus. In Stegosaurus 45 46 the quadratojugal is more restricted: having a narrow jugal process and 47 48 posteriorly ascends the quadrate for a short distance; it may not extend 49 ventrally to reach the condylar region, although this apparent abbreviation 50 51 may also be a quirk of preservation (see Gilmore 1914: fig. 6). In 52 53 ankylosaurs the quadratojugal is completely obscured because of fusion 54 between adjacent bones and its being is overlain by large osteoderms. 55 56 57 Squamosal (Figs 8, 9, 16, 18). The squamosals (Sq) are broken and 58 eroded on both sides of the skull of the lectotype (NHMUK R1111: Fig. 59 60 21A) but are well exposed on the surface of the articulated referred skull
63 Zoological Journal of the Linnean Society Page 64 of 215
1 2 3 (BRSMG LEGL 0004: Fig. 18). None are preserved in the intermediate- 4 5 sized individual (BRSMG Ce12785). 6 7 On the left side of the lectotype (Fig. 21A), the only part remaining 8 9 is the anterior descending process that forms a tapering structure 10 attached along the anterior edge of the jugal wing of the quadrate. Its 11 12 upper end is broad and clearly expanded as it approaches the central body 13 14 of the squamosal. The anterior surface is smoothly rounded transversely, 15 while the posterior surface is grooved along its length so that it can wrap 16 17 itself tightly around the edge of the jugal wing. Distally, the squamosal 18 19 process tapers to a point that overlaps the dorsal tip of the quadratojugal 20 (excluding the quadrate from the margin of the infratemporal fenestra). 21 For Review Only 22 The right squamosal is poorly preserved. 23 24 As can be seen in the articulated skull of the referred specimen 25 26 (BRSMG LEGL 0004: Figs 16, 18), the squamosal has a typically 27 28 dinosaurian morphology. The bone has a robust central portion that 29 houses the vaulted recess that is the cotylus for the quadrate head. From 30 31 this central region radiate a number of processes: a long, slightly-arched, 32 33 anterior (postorbital) process, a shorter anteromedially-curved (parietal) 34 process, and a posteroventral wing-like (paroccipital) process. The 35 36 postorbital process tapers anteriorly and contacts a complementary 37 38 process of the postorbital (creating the intertemporal bar). The suture 39 between these bones appears to be a diagonal overlap in dorsal (Fig. 16) 40 41 and lateral (Fig. 18) aspects; this was, however, evidently more complex: 42 43 the process of the squamosal appearing to wrap itself around the dorsal 44 and medial sides of the postorbital as it develops (ontogenetically). The 45 46 lateral surface of the postorbital process of the squamosal has an arched 47 48 ledge that curves ventrally beyond the quadrate cotylus (Fig. 18); this 49 ledge and the area beneath would have anchored a superficial portion 50 51 (MAMES) of the adductor musculature. The lower edge of the postorbital 52 53 process is transversely rounded and, along with the postorbital, forms the 54 dorsal margin of the infratemporal fenestra; its medial surface is smooth 55 56 and slightly concave (facing medioventrally) and rises to an acute internal 57 58 edge. This latter edge is sculpted with ropey-textured bony fibres 59 indicating the attachment site for deeper portions (MAMEP) of the 60
64 Page 65 of 215 Zoological Journal of the Linnean Society
1 2 3 adductor musculature (Fig. 16). This dorsal margin traces the lateral rim 4 5 of the supratemporal fenestra. Posteromedially the dorsal edge of 6 7 squamosal becomes noticeably thinner and sharper, curving medially to 8 create a deep lamina that overlaps the posterolateral process of the 9 10 parietal within the adductor chamber. The ventral edge of this squamosal 11 12 lamina overlies the opisthotic. 13 14 Posterolaterally, the squamosal forms a thick shoulder that curves 15 ventrally toward the paroccipital process and forms a wing-like structure 16 17 that overlies its anterior surface. The occipital portion of the squamosal is 18 19 mostly hidden from view in the referred specimen (BRSMG LEGL 0004: 20 Fig. 16) by the presence of two horn-shaped occipital osteoderms (Fig. 21 For Review Only 22 16). The two occipital osteoderms are attached to the sloping dorsal 23 24 portion of the occiput, judged by the position that they now occupy – as 25 suggested in Figure 11 (see also Fig. 47). The left occipital horn comprises 26 27 a superficially positioned sub-conical osteoderm that is attached to the left 28 29 occiput via a wedge-like base-plate (Fig. 16, ba). The latter is sutured to 30 the posterodorsal occipital wall and would have been anchored, in part at 31 32 least, to the exposed occipital surfaces of the adjacent squamosal, parietal 33 34 and supraoccipital. The right horn is a little displaced, as preserved (Fig. 35 36 16). 37 38 Comparisons. The right squamosal of Emausaurus appears to be 39 damaged (Haubold 1990: fig. 8) but exhibits a long, tapering 40 41 quadratojugal process, a well-developed cotylus and a lateral ledge on the 42 43 postorbital process, all of these features resemble those seen in 44 Scelidosaurus. Among eurypodan ankylosaurs, the squamosal tends to be 45 46 obscured by dermal bone. The basal stegosaur Huayangosaurus appears 47 48 to show a squamosal with a curved, external ledge and a tapering 49 quadratojugal process; however, as reconstructed (Sereno & Dong 1992: 50 51 fig. 6) the squamosal and quadratojugal did not make contact on the 52 53 margin of the infratemporal fenestra. The squamosal of Stegosaurus is 54 highly modified to accommodate the oblique, laterally compressed head of 55 56 the quadrate; there is no quadratojugal process, but the dorsolateral 57 58 margin of the postorbital process is ledge-like. 59 60
65 Zoological Journal of the Linnean Society Page 66 of 215
1 2 3 DESCRIPTIVE OSTEOLOGY: THE PALATAL COMPLEX 4 5 6 7 8 Very little is known of the palate of eurypodans, so few comparisons are 9 possible. 10 11 Vomer (Figs 23-25, 28). The vomers (V) are preserved almost in their 12 13 entirety in the lectotype (NHMUK R1111). Figure 23 presents sketches 14 15 that indicate the deterioration of these bones that has occurred in a 16 decade. Figure 23A represents an accurate sketch that I made in 1999; it 17 18 can be compared directly with a similar sketch made in 2009 (Fig. 23B,C). 19 20 The posterodorsal edges of both vomers have been broken and lost. This 21 For Review Only is unfortunate because the loss of these parts of the vomers means that 22 23 the evidence regarding the positioning of these bones with respect to the 24 25 pterygoid has also been lost. A small square fragment of vomer preserved 26 attached to the pterygoid (Fig. 23E, vfr) fitted perfectly in the ‘notch’ (Fig. 27 28 23A) present on the posterodorsal corner of the intact vomers and 29 30 permitted the sketched reconstruction seen in Figure 23E. The anterior 31 portion of the vomers is visible in articulation in the partially prepared 32 33 detached snout nodule of the smaller referred specimen (BRSMG LEGL 34 35 0004: Fig. 24). 36 37 38 39
66 Page 67 of 215 Zoological Journal of the Linnean Society
1 2 3 in position by flattened, tapering splints that project from the posterior 4 5 margin of the premaxilla (Fig. 23, Pm). The lateroventral margin of the 6 7 sutural surface is notched and has a small ventral projection (lip) that 8 secures this suture. 9 10 11 12 13
67 Zoological Journal of the Linnean Society Page 68 of 215
1 2 3 skull was severely damaged by a palaeoichthyologist. Photographic 4 5 images of the skull made at intervals during the phases of the skull’s acid 6 7 preparation do not (unfortunately) reveal the original position of these two 8 bones. I suspect that they were positioned deep within the inverted skull: 9 10 that is to say, close to the roof of the snout (and effectively ‘invisible’ to 11 12 the lighting and amateur photographic techniques used at the time), 13 otherwise there would have been some indication of their whereabouts on 14 15 these photographs. Checking back through my notes I found that in the 16 17 late 1970s (prior to the event that so badly damaged the skull) I had 18 made crude pencil sketches (Fig. 25) of the articulated, incompletely acid- 19 20 prepared, skull of the lectotype (NHMUK R1111). I was, at this time, 21 For Review Only 22 attempting to decipher the layout of the palate of Iguanodon, in 23 preparation for two monographic studies (Norman 1980, 1986). The 24 25 sketches indicate the presence of two thin sheets of bone lateral to the 26 27 vomers. Judged from my annotations at the time I clearly assumed that 28 these were probably anterior extensions of the palatine bones. However, 29 30 they are more likely to be the two bones in question (Fig. 25, ?epv) 31 32 because the palatines of Scelidosaurus do not extend either anteriorly or, 33 more importantly, dorsally toward the roof of the snout. 34 35 36 These bones are thin and mildly dished. Their edges are damaged 37 in places but the two bones (allowing for damage) appear to be mirrored 38 39 structures (Fig. 26C,D) that can be positioned on either side of the sagittal 40 41 plane. As preserved (Figs 26A,B) these bones are roughly triangular and 42 each has a somewhat thickened and bevelled edge (bev) that has been 43 44 found, by cautious manipulation, to be complementary in shape to the 45 46 bevelled dorsolateral edges on each vomer. One of these bones (Fig. 26B) 47 has a few irregular perforations that may have been created by excessive 48 49 acid leaching of these rather thin bony plates. The other bone is 50 51 imperforate. When an attempt is made to articulate these bones with the 52 vomers they form vaulted ‘wings’ (Fig. 26E) that project laterally, 53 54 perpendicular to the principal planes of the vomers. These bone, if they 55 56 are articulated correctly, formed a substantial part of the roof to the nasal 57 cavity (nc). The epivomers extend laterally toward the internal surface of 58 59 the nasals; whether they contacted (for example) the deep curtain-like 60
68 Page 69 of 215 Zoological Journal of the Linnean Society
1 2 3 premaxillary flange on each nasal cannot be ascertained. The incomplete 4 5 free anterior and posterior edges of these bones may indicate that these 6 7 structures were ossified portions of more extensive sheets of tissue that 8 roofed the nasal cavity. 9 10 These bones appear to be unique to Scelidosaurus. 11 12 13 14 15
69 Zoological Journal of the Linnean Society Page 70 of 215
1 2 3
6 7 8 The left palatine is more complete and preserved in articulation with 9 the maxilla (Fig. 12B, 27F,G). As shown in medial view (Fig. 12B) the 10 11 maxillary suture is extensive: as previously described, it caps the dorsal 12 13 surface of the maxilla and furthermore develops a thin, deep, ventral 14 flange forming a sheet that is sutured to the medial maxillary wall (mwp). 15 16 This palatine sheet extends ventrally to just above the row of special 17 18 foramina that lie parallel to the alveolar margin. In ventral view (Fig. 27G) 19 where it is articulates with the ectopterygoid and maxilla, the 20 21 posterolateral cornerFor of the Review left palatine spans Only a small gap in the 22 23 anteromedial corner of the palatine (suborbital) fenestra (Fig. 27G, pf). 24 25 Ectopterygoid (Fig. 27). The left ectopterygoid (Ec) is disarticulated in 26 27 the lectotype (NHMUK R1111: Fig. 27C,D,E,G). It is a robust, U-shaped 28 bone forming a smooth, concave, anteroventral border to the adductor 29 30 chamber/sub-temporal fenestra. The ectopterygoid forms an important 31 32 structural link, binding together the jugal, maxilla, palatine and pterygoid, 33 as well as bordering the palatine/suborbital fenestra and reinforcing the 34 35 lateral edge of the pterygoid flange. Laterally, the ectopterygoid produces 36 37 a posteriorly pointed, transversely-flattened process that bears a facet on 38 its lateral surface (js) that contacts an equivalent facet on the medial wall 39 40 of the jugal (Fig. 21B, ecs). Anterior to this facet its anterior edge curves 41 42 medially and becomes immediately suturally bound to the medial surface 43 of the maxilla (Fig. 27F,G); this edge then becomes smooth and concave 44 45 anterolaterally because it forms the posterior margin of the 46 47 palatine/suborbital fenestra (pf). Farther medially, the edge of the 48 ectopterygoid becomes rugose again, where it forms a sutural surface for 49 50 the palatine (pal.s) before the edge turns posteriorly as a scarf-style 51 52 sutural flange (on its ventral surface – Fig. 27G, pt.s) for contact with the 53 pterygoid and backing the pterygoid flange. The ventral surface (Fig. 54 55 27C,D & G) shows the pterygoid suture clearly as well as the oblique 56 57 channel that is directed medioventrally from the palatine/suborbital 58 fenestra (pf). 59 60
70 Page 71 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5
71 Zoological Journal of the Linnean Society Page 72 of 215
1 2 3 medial wall of the basal articulation (Fig. 28B). This medial wall curves 4 5 laterally and then swings dorsally as it creates the dorsal edge of the 6 7 quadrate wing (qw) of the pterygoid. The lateral surface of the basal 8 articulation is at the base of the quadrate wing of the pterygoid. Quite 9 10 near the origin of the dorsal edge of the quadrate wing sits a discrete 11 12 epipterygoid (Ep). The dorsal edge of the quadrate wing is very thin and 13 curves dorsally, following the margin of the pterygoid wing of the 14 15 quadrate; in contrast the ventral edge of the wing runs horizontally 16 17 posteriorly and is considerably thicker, it also supports on its medial side 18 an unusual, narrow, pocket-like slot (Fig. 28B, sl). The posterior margin of 19 20 the quadrate wing is broadly emarginated at mid-height so that the wing 21 For Review Only 22 appears to be broadly forked. 23 24 Lateral to the base of the quadrate wing of the pterygoid there is a 25 posteroventrally-oriented shallow trough. The lateral border of this trough 26 27 is the deflected dorsal edge of the vomerine process. This angled trough 28 29 provides room for a scarf-style suture with the overlying medial wing of 30 the palatine. The trough continues posteroventrally and forms a rather 31 32 thin and insubstantial pterygoid flange (fl); the latter is evidently 33 34 reinforced (structurally) by the attachment of the robust medial arm of 35 36 the ectopterygoid, to which it is sutured (ecs). 37 38 Epipterygoid (Fig. 28). Epipterygoids (Ep) are rarely recorded as being 39 present in archosaurs generally (Romer 1956; see also Holliday & Witmer 40 41 2009 for its elimination in crocodile evolution) and are rarely reported 42 43 among dinosaurs (e.g. some theropods – see Weishampel, Dodson & 44 Osmólska 2004). Among ornithischians, Maryańska & Osmólska (1974) 45 46 reported an epipterygoid in the well-preserved skull of the 47 48 pachycephalosaurid Prenocephale. Maryańska (1977: 116) was also able 49 to report the presence of epipterygoids in some ankylosaurids: Saichania, 50 51 Pinacosaurus and Euoplocephalus. Epipterygoids have not so far been 52 53 reported in stegosaurs or in any other ornithischian sub-clades. 54 55 A left epipterygoid is attached to the quadrate wing of the pterygoid 56 57 of the lectotype (Fig. 28, Ep). It has an anteroposteriorly expanded and 58 laterally-flattened base that is firmly bound/sutured against the lateral 59 60 surface of the pterygoid. The dorsal edge of the pterygoid is sculpted
72 Page 73 of 215 Zoological Journal of the Linnean Society
1 2 3 adjacent to the area of attachment of the epipterygoid: producing a 4 5 flattened, pillar-shaped structure (pil) that backs the medial side of the 6 7 base of the epipterygoid; equally, the anteroventral edge of the 8 epipterygoid is warped (w) so that it wraps around the edge of the 9 10 pterygoid creating an evidently snug fit. Dorsally, the epipterygoid is 11 12 recurved and tapers to a narrow point that would have been positioned 13 close to the proötic though there is no clear evidence of sutural or 14 15 articular contact between the epipterygoid and any of the adjacent 16 17 braincase bones. 18 19 Comparative comments. The palatal structure of the basal 20 thyreophoran Emausaurus is unknown and the palatal elements of 21 For Review Only 22 eurypodans such as ankylosaurs are highly derived. The palate of 23 24 Huayangosaurus is largely unknown; however the palate of Stegosaurus 25 was illustrated by Gilmore (1914: pl. 7) and bears some general similarity 26 27 to the features seen in Scelidosaurus, but there is not enough information 28 29 to allow for detailed comparisons to be made. 30 31 32 33
73 Zoological Journal of the Linnean Society Page 74 of 215
1 2 3 quadrate and its ventral (mandibular) condyles. Although smooth and 4 5 articular in appearance, the quadrate head is located in a deep socket in 6 7 the body of the squamosal. The presence of anterior and posterior 8 processes that, in effect, clamp the proximal portion of the quadrate shaft 9 10 would have limited palinal (fore-aft) rotation between the head of the 11 12 quadrate and the squamosal cotylus; however, in contrast, mediolateral 13 rotation and consequent flexure of the quadrate shaft cannot completely 14 15 be ruled out. The posterior edge of the shaft, directly beneath the 16 17 quadrate condyle, descends vertically as a transversely compressed 18 buttress (cb). This condylar buttress extends for about 20% of the length 19 20 of the shaft, beyond which the shaft is bowed evenly anteriorly, creating 21 For Review Only 22 an overhang. The anterior edge of the quadrate head descends obliquely 23 before dividing into widely divergent laminae, the jugal and pterygoid 24 25 wings. 26 27 28 29
74 Page 75 of 215 Zoological Journal of the Linnean Society
1 2 3 in the quadrate of Heterodontosaurus (Norman et al. 2011: fig. 14, ?pn). 4 5 The ventral portion of the jugal wing of the quadrate is overlain by the 6 7 posterior edge of the quadratojugal (Fig. 29, 30C) and there is a well- 8 marked elongate sutural surface (Fig. 30, qjs). Beneath this scarred 9 10 surface the jugal wing merges with the main body of the quadrate above 11 12 the lateral expansion that supports the oblique surangular portion (sac) of 13 the mandibular condyle. 14 15 The pterygoid wing originates as a ridge on the anteromedial corner 16 17 of the quadrate condyle; this ridge becomes progressively narrower, and 18 19 the wing very deep, as it expands medioventrally in a nearly straight line 20 to overlap the quadrate wing of the pterygoid (Figs 28B, 29, 30). The 21 For Review Only 22 pterygoid wing reaches its greatest expansion at about mid-shaft where it 23 24 forms a convex bulge before contracting back, sharply, to merge with the 25 lower part of the shaft well above the expansion for the mandibular 26 27 condyle. The medial surface of the pterygoid wing bears an unusual deep, 28 29 sutural area for the thin, bifurcated quadrate wing of the pterygoid. Just 30 posterior to this sutural surface the pterygoid wing bears a broad 31 32 depression (Fig. 30, dep), although the presence of this feature varies 33 34 between individual specimens: well developed in CAMSM X39256 and 35 36 BRSMG Ce12785, yet poorly developed in the lectotype. The quadrate 37 wing of the pterygoid comprises a thin tapering process that is pressed 38 39 against the mediodorsal portion of the pterygoid wing; and a slightly more 40 41 robust process that is pressed against the medioventral surface of the 42 pterygoid wing of the quadrate (Fig. 28). The degree of development of 43 44 the sutural surfaces for the pterygoid varies between the left and right 45 46 quadrates of NHMUK R1111, but this may well be an artefact of 47 preparation. 48 49 The mandibular condyle of the quadrate is expanded transversely 50 51 and has the rough outline of an elongate rounded triangle in distal end 52 53 view (Fig. 10), the apex curving laterally and the base directed medially. 54 The posterior edge of the articular surface lies roughly perpendicular to 55 56 the long axis of the skull. The lateral corner is relatively narrow 57 58 anteroposteriorly, but roller-like anteroposteriorly with an oblique axis 59 (sac). More medially the condylar surface expands steadily to form an 60
75 Zoological Journal of the Linnean Society Page 76 of 215
1 2 3 enlarged and more anteroposteriorly expanded, near-spherical, medial 4 5 condylar surface (ac). A shallow, oblique groove seems to separate what 6 7 are distinguished as two condylar areas. The articular surfaces are 8 comparatively smooth and would have been coated with articular 9 10 cartilage. A framework of thin curved ridges that surround the margins of 11 12 the condylar area mark the limits of the articular cartilage. The 13 asymmetry evident in the structure of the lateral and medial portions of 14 15 this quadrate articular surface reflects the structure of the mandibular 16 17 cotylus and this may well have constrained the motion of the lower jaw 18 when in articulation. 19 20 Comparisons. The quadrate of Emausaurus is not sufficiently well- 21 For Review Only 22 preserved to allow comparison to be made. In eurypodans the quadrate is 23 24 inconsistently described. Among stegosaurs the basal taxon 25 Huayangosaurus exhibits a quadrate that is similar in shape to that of 26 27 Scelidosaurus (Sereno & Dong 1992: fig. 6C): it is illustrated with a larger 28 29 medial condyle and an oblique lateral condyle, as seen in Scelidosaurus, 30 and there is even an indication of a pit, notch or recess dorsomedial to the 31 32 quadrate foramen. The quadrate of Stegosaurus is highly specialised in 33 34 having an obliquely compressed proximal head (Gilmore 1914: fig.5), but 35 36 it does have the asymmetrical distal articular condyles as seen in 37 Scelidosaurus, and as reconstructed in Huayangosaurus. Among 38 39 ankylosaurs the quadrate tends to be cryptically positioned beneath a 40 41 shield of dermal osteoderms (caputegulae) in ankylosaurids, although it is 42 visible in posteroventral view. In nodosaurids the quadrate is a little better 43 44 known and illustrated. The ventral portion of the quadrate, where visible, 45 46 conforms in its sutural relationships with those seen in Scelidosaurus and 47 the mandibular condyle displays a similar asymmetry. 48 49 50 51 52 DESCRIPTIVE OSTEOLOGY: THE BRAINCASE 53 54 55 56 General comments 57 58 The temporal region of the skull of Scelidosaurus is completely open, 59 60 having large fenestrae bounded by the intertemporal bar and jugal arch.
76 Page 77 of 215 Zoological Journal of the Linnean Society
1 2 3 In terms of the general anatomical setting of the braincase with respect to 4 5 the surrounding cranial framework, it is comparable to the structure 6 7 described in Lesothosaurus (Porro et al. 2015) and Heterodontosaurus 8 (Norman et al. 2011). It is particularly unfortunate that the braincase of 9 10 the basal thyreophorans Emausaurus and Scutellosaurus are not 11 12 preserved. This contrasts markedly with the largely roofed over or fully 13 closed structure created by the presence of caputegulae (cranial 14 15 osteoderms) in the more derived ankylosaurs. As a consequence, 16 17 complete braincase morphology though generally understood (see 18 Vickaryous, Maryańska & Weishampel 2004) is less well known in fine 19 20 detail. Parts of the braincase are, more often than not, visible in ventral, 21 For Review Only 22 posterior and ventrolateral aspect (Coombs 1978, Vickaryous & Russell 23 2003, Vickaryous, Maryańska & Russell 2004). There are of course a few 24 25 exceptions such as the isolated partly eroded nodosaurid braincase that 26 27 was attributed to Polacanthus (Norman & Faiers 1996). Equally Tumanova 28 (1987) described partial braincases of Talarurus, Pinacosaurus and 29 30 ‘Amtosaurus’. The cranial material of Saichania (Maryańska 1977), 31 32 Pawpawsaurus (Lee 1996), Cedarpelta (Carpenter et al. 2001) also 33 includes portions of their braincases. Several juvenile specimens of 34 35 Pinacosaurus have also been recovered (Maryańska 1977, Hill et al. 2003) 36 37 but, as is often the case, crushing and distortion obscure details and 38 confound comparative assessment. Stegosaurs, notably Stegosaurus, 39 40 have a more open temporal framework (see Gilmore 1914: fig. 10). Apart 41 42 from the original work of Gilmore and the paper by Galton (2001 – which 43 expands on the work of Gilmore but is solely focused on Stegosaurus) 44 45 surprisingly little is known of the range and variation in structural 46 47 anatomy of stegosaur braincases. 48 49 In most regards it is close to a truism to say that the braincases of 50 51 basal ornithischians (and I include Scelidosaurus in this regard) seem to 52 be structurally conservative, with forms such as Lesothosaurus and 53 54 Heterodontosaurus differing relatively little from that seen in 55 56 Scelidosaurus with regard to the general layout of their component parts. 57 58 59 60
77 Zoological Journal of the Linnean Society Page 78 of 215
1 2 3 Parasphenoid (Figs 10, 31). The parasphenoid forms a dermal shield 4 5 beneath the floor of the braincase; it is often fused indistinguishably with 6 7 the basisphenoid (Romer 1956). It extends posteriorly toward the basal 8 tubera (which provide an attachment site for the sub-vertebral 9 10 musculature). Anteriorly, the parasphenoid often forms a narrow rostrum 11 12 (pr) that projects anteriorly on the midline and bisects the pterygoid 13 vacuity; this process forms a physical keel beneath the dorsally located 14 15 ethmoid cartilages. 16 17 An isolated rostrum (cultriform process) (Fig. 31A, pr) is preserved 18 19 among the many skull fragments and comprises a ventrally bowed, 20 tapering rod with a V-shaped cross-section. The posterior end of this 21 For Review Only 22 process is truncated, with a smooth-walled recess (beyond which this part 23 24 of the braincase was presumed to have been cartilaginous), so it was 25 evidently disconnected (osteologically) from the adjacent pituitary fossa 26 27 (the anterior walls of which are themselves unossified) and the ossified 28 29 floor of the braincase. 30 31 32 33
78 Page 79 of 215 Zoological Journal of the Linnean Society
1 2 3 enter the pituitary fossa ventrolaterally through the dorsum sellae. These 4 5 canals emerge on the external surface of the basisphenoid just posterior 6 7 to the base of each basipterygoid process (Fig. 31, vid); they represent 8 the Vidian canals for the arterial (carotid) supply to the pituitary and 9 10 endocranial cavity. 11 12 Projecting ventrolaterally and slightly posteriorly from the body of 13 14 the basisphenoid below the pituitary fossa is a pair of large basipterygoid 15 processes (bpt); these project ventrolaterally, taper slightly toward their 16 17 distal ends and then terminate as discrete, ovoid, convex articular 18 19 surfaces for the basal articulation between the braincase and palate (Fig. 20 28B, b.ar). Between the basipterygoid processes the ventral surface of the 21 For Review Only 22 basisphenoid is arched transversely and anteriorly, and starts to form a 23 24 ventral keel that lies close to the unossified proximal end of the 25 parasphenoid rostrum (cultriform process). Posteriorly, the ventral surface 26 27 of the basisphenoid becomes shallowly vaulted in the midline. The 28 29 vaulting deepens (Fig. 31B) and eventually, like an inverted adit, tunnels 30 into the basisphenoid-basioccipital suture above the everted, rugose 31 32 ridges (svm) for the insertion of the cervical sub-vertebral musculature. 33 34 Posterior to the bases of the basipterygoid processes, the lateral sides of 35 36 the basisphenoid contract strongly before re-expanding to form enlarged, 37 oblique and pedicle-like structures that form a slightly open suture against 38 39 the anterolateral surface of the basioccipital and the enlarged, oblique, 40 41 basioccipital tuberosities (bot). 42 43 The anterior lateral wall of the basisphenoid dorsal to the 44 basipterygoid process is comparatively smooth and forms, at its dorsal 45 46 edge, a concave ventral floor to the large fossa that forms at the junction 47 48 of the three major branches of the Trigeminal nerve (cn.V). The anterior 49 edge of the basisphenoid medially contacts a foot-like ventral projection of 50 51 the laterosphenoid. Posteriorly, the basisphenoid is overlain by a 52 53 triangular flange formed by the proötic. Below this area, the surface of the 54 basisphenoid is obliquely grooved for the palatine branch of the Facialis 55 56 nerve (cn.VII); the posterior edge of this groove merges with the pedicle- 57 58 like portion of the basisphenoid. The pedicle extends dorsally into the 59 auditory fossa and partially separates the fenestra ovalis from the more 60
79 Zoological Journal of the Linnean Society Page 80 of 215
1 2 3 ventrally-positioned fenestra rotunda. The suture between basisphenoid 4 5 and basioccipital appears to lie in the floor of the auditory fossa. 6 7 Basioccipital (Figs 10, 11, 31, 32). The basioccipital (Bo) is reasonably 8 9 well exposed in the lectotype skull (Fig. 31) and the referred specimen 10 (CAMSM X39256: Fig. 32). In ventral view, its anterior margin forms a V- 11 12 shaped, rugose-edged lip that underlies the body of the basisphenoid. The 13 14 ventral midline behind this everted lip forms a narrow keel that merges 15 with the body of the basioccipital prior to the posterior expansion of the 16 17 occipital condyle (Fig. 31B). On either side of the ventral lip the 18 19 ventrolateral walls of the basioccipital are expanded to form two rugose- 20 faced obliquely-oriented tuberosities (bot); these latter extend up the 21 For Review Only 22 lateral sides of the basioccipital and meet a prominent foot-like process 23 24 (op.ped) at the most ventral portion of the opisthotic. Apart from this 25 sutural junction, preservation does not permit an accurate delineation of 26 27 the suture between the basioccipital and opisthotic/exoccipital. 28 29 The expansion forming the occipital condyle is reniform when 30 31 viewed posteriorly (Figs 32, 33); there is a broad indent dorsally, marking 32 33 the floor of the foramen magnum (fm). The condyle appears to face 34 directly posteriorly and its principal articular axis does not seem to be 35 36 deflected ventrally. The dorsolateral corners of the occipital condyle form 37 38 small raised promontories where they meet the pedicles of the 39 exoccipitals; these promontories contribute to the formation of the 40 41 occipital condyle (CAMSM X39256: Fig. 32). There is no evidence to 42 43 suggest that the exoccipitals met medially to form the floor of the 44 foramen magnum (CAMSM X39256 and BRSMG LEGL 0005). There is also 45 46 evidence (CAMSM X39256 – Fig. 32) for a shallow notochordal ‘dimple’ in 47 48 the upper centre of the occipital condyle, but this might represent a 49 juvenile ontogenetic feature. 50 51 52 53 54
80 Page 81 of 215 Zoological Journal of the Linnean Society
1 2 3 lectotype (Fig. 33), but its sutural boundaries are not clear. In the juvenile 4 5 specimen (CAMSM X39256) although the basioccipital-exoccipital suture is 6 7 completely open, there is no indication of a suture between the left 8 exoccipital and opisthotic (Fig. 32A), suggesting that these bones must 9 10 have fused together early in ontogeny. As preserved in the juvenile 11 12 specimen just mentioned the exoccipitals meet the supraoccipital on 13 either side of the dorsal margin of the foramen magnum, rather than 14 15 meeting the opposing exoccipital – thereby excluding the supraoccipital 16 17 from this boundary. The juvenile specimen (CAMSM X39256: Fig. 32B) 18 demonstrates that the exoccipital pedicles have articular surfaces and 19 20 contribute to the formation of the occipital condyle. 21 For Review Only 22 23 24
81 Zoological Journal of the Linnean Society Page 82 of 215
1 2 3 The area occupied by the supraoccipital slopes anterodorsally and 4 5 runs toward and comes to lie beneath the overlying parietal plate. In the 6 7 juvenile specimen (Fig. 32B,C) the structure of the supraoccipital suture 8 and shape of the paroccipital wing (adjacent to the suture) indicates the 9 10 presence of an oblique canal (can) that runs forward between these bones 11 12 in the general direction of the otic capsule. In the lectotype there is no 13 indication of an equivalent canal (unless this canal has been filled with 14 15 consolidant). The midline of the supraoccipital in the lectotype bears a low 16 17 midline ridge (Fig. 33, sr); on either side of this ridge the dorsal surface 18 slopes ventrolaterally and appears to be mildly concave. On either side of 19 20 the midline ridge the concave flanks are marked by an array of transverse 21 For Review Only 22 rugae; these latter may well indicate the attachment areas for bands of 23 ligaments associated with a nuchal osteoderm. The latter has been 24 25 identified in CAMSM X39256 and BRSMG LEGL 0004 (Norman [Part 3]). 26 27 28 29 Opisthotic (Figs 11, 31-33). The opisthotic (Op) cannot be differentiated 30 31 from the exoccipital, and this may be linked, in part, to the reinforcement 32 33 of the occipital plate and proximate braincase bones for the anchorage of 34 the occipital horns. The opisthotic-exoccipital forms the majority of the 35 36 lower posterolateral wall of the braincase and appears to be co-extensive 37 38 (ventrally) with the basioccipital upon which it sits. A number of discrete 39 foramina (or larger fossae containing subsidiary foramina) penetrate its 40 41 ventrolateral wall. The interpretation of these fossae and foramina is 42 43 partly conjectural and based upon several comparative accounts, most of 44 which originate in the classic work by Romer (1956). 45 46 47 Posteriorly there is a large dorsoventrally elliptical foramen (Fig. 48 31A) that probably contained the Hypoglossal nerve (cn.XII). Anterior to 49 50 this there is a large, apparently double, opening that may have served to 51 52 convey a vein as well as the Accessory (cn.XI), Vagus (cn.X) and 53 Glossopharyngeal (cn.IX) nerves; slightly anterior and dorsal to this is 54 55 another large foramen that most likely contained the jugular vein. 56 57 Immediately anterior to the jugular foramen is a stout pillar of bone that 58 terminates ventrally in a prominent pedicle (op.ped) that lies dorsal, and 59 60 adjacent, to the large basioccipital tuberosity (bot).
82 Page 83 of 215 Zoological Journal of the Linnean Society
1 2 3 Anterior to this pillar-like structure is the auditory fossa, which is 4 5 partially subdivided into anterodorsal and anteroventral canals by an 6 7 (apparently) incomplete septum. The dorsal foramen is the fenestra ovalis 8 for the footplate of the stapes (Fig. 34). A.J. Charig (pers. comm. circa 9 10 1980) found and removed the ossified portion of the stapes in position as 11 12 the acid preparation of the skull progressed during the 1970s, but its 13 exact position and orientation were not recorded at the time. Immediately 14 15 beneath the fenestra ovalis is an equally large foramen that may have 16 17 conveyed the Glossopharyngeal nerve (cn.IX) as well as housing the 18 fenestra rotunda. Anterior to the auditory fossa is another elliptical 19 20 foramen that is tucked under the ventral margin of the proötic; this was 21 For Review Only 22 undoubtedly the opening for the Facialis nerve (cn.VII) and its two major 23 branches (palatine and hyomandibular). In the example provided by the 24 25 juvenile referred specimen (CAMSM X39256: Fig. 32) the lateral wall of 26 27 the opisthotic is greatly simplified by comparison to the lectotype: two 28 large, elliptical foramina (for) exit through its sidewall; anterior to these 29 30 foramina is the pillar-like opisthotic pedicle. The latter pillar is notched 31 32 anteriorly (indistinctly) by the eroded remnant of the auditory fossa. It is 33 therefore highly likely that the appearance of a larger number of discrete 34 35 cranial foramina in the lateral wall of the opisthotic in larger individuals is 36 37 a progressive developmental feature associated with the continued 38 differentiation of passages for the nerves and blood vessels that exit the 39 40 braincase. 41 42 43 44
83 Zoological Journal of the Linnean Society Page 84 of 215
1 2 3 opisthotic that is mostly consistent with the structure seen in the 4 5 lectotype, with regard to shape and orientation of this process. The dorsal 6 7 edge of the paroccipital process forms a narrow, saddle-shaped structure 8 that supports the parietal medially and squamosal laterally. The lectotype 9 10 (Fig. 33) displays a curious additional structure on its dorsal edge. There 11 12 is a notch representing a channel whose path traverses this upper edge. 13 This notch/channel appears as a shallow gulley on the occipital face of the 14 15 paroccipital process. The notch is almost entirely enclosed by a spur-like 16 17 structure (Fig. 33, psp). The dorsal edge of the paroccipital wing (medial 18 to the spur-like process) is rugose, reflecting the sutural attachment of 19 20 the parietal and squamosal. The spur-like process encloses a channel for 21 For Review Only 22 occipital vasculature that drains into the adductor chamber and thence 23 into the transverse sinus high on the lateral wall of the endocranial cavity. 24 25 This opening probably represents a remnant of the post-temporal 26 27 fenestra. 28 29 In the posterior view of the occiput of the lectotype (Fig. 33) a 30 bulge on the internal wall of the braincase to accommodate the otic 31 32 capsule is clearly visible (otc), as are the paired foramina for the 33 34 Abducens nerve (cn.VI) in the floor of the braincase. 35 36 Proötic (Fig. 31). The proötic (Pr) is clearly demarcated on the sidewall of 37 38 the braincase. It has a roughly triangular flange, with a clear sutural 39 boundary, that extends posteriorly and overlaps the otic capsule contained 40 41 largely within the body of the opisthotic. The ventral portion of this suture 42 43 continues anteroventrally and then curves directly ventrally to contact the 44 basisphenoid behind the Trigeminal fossa (cn.V) and then curves dorsally 45 46 to form the upper posterior margin of that fossa where it forms a rugose 47 48 butt-style suture with the laterosphenoid. The dorsal margin appears to be 49 a more or less horizontal where it meets the edge of the parietal plate. At 50 51 the junction between the proötic and opisthotic there is a shallow, roughly 52 53 triangular, fenestra that marks the entry of the vena capitis dorsalis via 54 the braincase wall into the transverse sinus that lines the wall of the 55 56 endocranial cavity. This fenestra is also seen in the articulated skull of the 57 58 referred specimen (BRSMG LEGL 0004: Fig. 16, vcd) 59 60
84 Page 85 of 215 Zoological Journal of the Linnean Society
1 2 3
85 Zoological Journal of the Linnean Society Page 86 of 215
1 2 3 the edges of the laterosphenoids. Dorsally the ‘blades’ form a partial, 4 5 inclined floor to the olfactory lobes, whereas ventrally the handles and 6 7 blade margins create a heart-shaped window opening on to the ethmoid 8 region. 9 10 The distal ends of the ‘handles’ are hooked anteriorly on the midline 11 12 forming a symphysis (sym). The latter, given its median position, 13 14 probably contacted the anterodorsal edge of the dorsum sellae and lies 15 immediately above the pituitary fossa. This implies that the ventral part of 16 17 the brain and hypothalamus passed through the fenestra so that the 18 19 ventral wall of the hypothalamus and its underlying pituitary body were 20 draped over the orbitosphenoid symphysis and down into the pituitary 21 For Review Only 22 fossa. 23 24 Dorsally, the ‘handles’ merge with the expanded ‘blades’. The lower 25 26 edge of each ‘blade’ curves abruptly anteromedially for a short distance 27 28 forming a pendent structure (forming the notch at the top of the heart- 29 shaped fenestra), and opposing medial edges meet at a, slightly open, 30 31 narrow tongue-in-groove suture. The symphysis between the ‘blades’ 32 33 extends for a short distance anterodorsally and forms what appears to be 34 a modest, oblique keel. The osseus medial edges of the blades then 35 36 diverge, leaving an irregular V-shaped notch between that may have been 37 38 spanned by connective tissue. The dorsal margin of each ‘blade’ then 39 swings laterally along a thickened and somewhat irregular edge that ends 40 41 in a dorsolateral spur (c) that seems to represent a contact point against 42 43 an equivalent structure (c*) on the anterior margin of each 44 laterosphenoid. The lateral edge of the ‘blade’ is embayed beneath the 45 46 spur and ends in a second spur; beyond the latter the lateral edge merges 47 48 into the ‘handle’ region. The ‘handles’ are circular in cross-section, 49 compared to the thin ‘blades’. Each blade has a small foramen on either 50 51 side of the midline adjacent to the symphysis. The orbitosphenoids lie 52 53 obliquely against the front edges of the laterosphenoids (Fig.31), 54 extending from the basisphenoid-presphenoid region on the floor of the 55 56 braincase, up across the laterosphenoids and approaching the skull roof; 57 58 they frame a median braincase fenestra and leave narrow slots around its 59 edges. The keeled portion of the orbitosphenoids probably linked ventrally 60
86 Page 87 of 215 Zoological Journal of the Linnean Society
1 2 3 to ethmoid cartilages that supported the interorbital septum and, 4 5 ultimately, connected to the cultriform process of the parasphenoid 6 7 ventrally. 8 9 The V-shaped notch between the orbitosphenoids in all probability 10 marks the passage for the olfactory lobes, which run forward toward the 11 12 snout region, directly under the skull roof. Ventral to this area the discrete 13 14 foramen in each blade may represent passages for the Trochlear nerve 15 (cn.IV). The large heart-shaped fenestra framed the presumably 16 17 cartilaginous casing for the anteroventral portion of the brain as well as 18 19 the vasculature supplying the floor of the brain, the orbital, nasal and 20 palatal regions and the Optic nerve (cn.II) and (perhaps) the ophthalmic 21 For Review Only 22 branch of the Trigeminal nerve (cn.V). 23 24 Ethmoid region. Anterior to the orbitosphenoids, no further trace of 25 26 ossifications has been discovered that might derive from the ethmoid 27 28 region of the skull. The presence of ethmoid cartilages surrounding the 29 brain, olfactory lobes and partitioning the orbital cavities can be inferred 30 31 from the V-shaped cross-section of the parasphenoid rostrum (cultriform 32 33 process) and, more arguably, the modest keel formed by the 34 orbitosphenoid blades and curved ridges on the roof of the frontals. 35 36 37 38 39 DESCRIPTIVE OSTEOLOGY: THE MANDIBLE 40 41 General comments 42 43 The mandible is stout and comparatively straight, even though the 44 45 dentition itself is bowed medially and slightly sinuous along its length. 46 There is no external mandibular fenestra and a coronoid eminence rises 47 48 just above the occlusal plane of the dentition; there is, however, no 49 50 prominent coronoid process. To date a bony predentary has not been 51 recovered; however, there is circumstantial evidence for the presence of 52 53 this bone (or a membranous/cartilaginous precursor). The dentition of the 54 55 mandible is markedly inset from the lateral surface and framed by a well- 56 defined buccal emargination. Much of external surface of the dentary is 57 58 coated with a layer of fibrous exostotic bone, except in the areas 59 60 associated with muscle attachment (surrounding the coronoid eminence),
87 Zoological Journal of the Linnean Society Page 88 of 215
1 2 3 and the fleshy structures associated with the buccal emargination. The 4 5 mandible has one particularly large mound-shaped exostosis (p.os); the 6 7 growth of the latter is centred upon the external surface of the angular 8 bone alone (BRSMG LEGL 0004: Figs 14, 39). In the largest and most 9 10 mature specimen currently known (the lectotype – NHMUK R1111), 11 12 strands of this superficial tissue encroach upon the adjoining surangular 13 and dentary bones. 14 15 Both mandibles of the lectotype (Figs 36, 37) have been prepared 16 17 chemically and are reasonably well preserved and nearly complete. 18 19 Unfortunately, erosion of the original skull nodule prior to discovery 20 means that both mandibles lack their anterior ends and the symphysis. An 21 For Review Only 22 isolated, well-preserved anterior portion of the dentary ramus that 23 24 includes the symphyseal suture is present among the disarticulated skull 25 fragments of the intermediate-sized referred specimen (BRSMG Ce12785: 26 27 Fig. 40). 28 29 Comparisons. Portions of the mandible of Emausaurus are known 30 31 and were used to reconstruct its entire structure (Haubold 1990: figs 2, 9; 32 33 taf. II, 2-4). The mandible, as reconstructed, is similar in its layout and 34 proportions to that seen in Scelidosaurus, except that the anterior portion 35 36 of the ramus of Emausaurus is more slender and sinuous. The most 37 38 obvious differences in anatomy, based on Haubold’s reconstruction, 39 include the presence of a large external mandibular fenestra at the 40 41 junction of the dentary, angular and surangular, and the absence of an 42 43 osteoderm centred upon the angular. The only skull of Emausaurus known 44 at present is comparatively small (~140mm long); this specimen is 45 46 smaller than that of the intermediate-sized specimen of Scelidosaurus 47 48 (BRSMG Ce12785: ~180mm). The relative paucity of dermal 49 ornamentation on the skull bones of Emausaurus might be attributable to 50 51 its potentially immature ontogenetic status. 52 53 Among eurypodans more generally the mandible is at least as 54 55 robust as that seen in Scelidosaurus. In the basal stegosaur 56 57 Huayangosaurus (Sereno & Dong 1992: fig.6A) only the external surface 58 is visible, nevertheless there is a general morphological similarity to that 59 60 of Scelidosaurus, notably in the close packing of the teeth and sinuosity of
88 Page 89 of 215 Zoological Journal of the Linnean Society
1 2 3 the mandibular ramus. However, in contrast there is a well-developed 4 5 external mandibular fenestra, a much higher coronoid eminence (probably 6 7 better described as a ‘process’) and there is no indication of a exostosis or 8 dermal ossification adhering to the posterior external surface of the 9 10 mandible. The more derived Stegosaurus (Gilmore 1914: pl. 5; Sereno & 11 12 Dong 1992: fig. 10A) has a mandible that is proportionally more elongate, 13 tapering and has a lower, more smooth, coronoid eminence. Gilmore 14 15 (1914: pl. 5) did not illustrate an external mandibular fenestra, and this 16 17 general area of the mandible in USNM 4934 (lateral to the junction of the 18 dentary, surangular and angular) gives the appearance of being covered 19 20 by a raised pad of tissue that resembles an osteoderm. However, a 21 For Review Only 22 substantial external mandibular fenestra was illustrated in a 23 reconstruction of the same skull by Galton (1990) by reference to earlier 24 25 reports by Berman & McIntosh (1986); this reconstruction was repeated 26 27 by Sereno & Dong (1992). Re-examination of Gilmore’s specimen by 28 Matthew Carrano (pers. comm. 2018) confirms the presence of a 29 30 mandibular fenestra, despite the presence of what he considers might be 31 32 exostotic bone in this area of the mandible. Among ankylosaurs the 33 mandible is extremely robust, very deep posteriorly and with a narrower 34 35 anterior tip. In general its osteology is less well known (for comparative 36 37 purposes) because it is covered externally by a massive mandibular 38 osteoderm (Coombs 1978). This plate of bone obscures sutural 39 40 relationships within the mandible and covers the area that would be the 41 42 location of the external mandibular fenestra; it is nevertheless clear that 43 the exostosis that is centred upon the angular of Scelidosaurus is similar 44 45 and potentially a structural predecessor of the much larger osteodermal 46 47 plate that characterises the external surface of the mandible in 48 ankylosaurs. 49 50 51 52 53
89 Zoological Journal of the Linnean Society Page 90 of 215
1 2 3 (Fig. 36, Ar); it is supported ventrally by a posterior extension of the 4 5 angular (Fig. 38A). The articular has a thickened medial ledge (Fig. 36C) 6 7 that overhangs the prearticular so that the whole jaw articulation, in 8 dorsal view, is hooked medially. The inturning of the articular region 9 10 compensates for the lateral bowing of the posterior half of the mandibular 11 12 ramus and ensures that the jaw joint and dentition are aligned 13 anteroposteriorly (see Figs 36C, 37C). 14 15 In the lectotype, the left articular is well exposed and shows less 16 17 damage than the right (Fig. 37, br), so this description is based primarily 18 19 upon the left. The smaller specimen (CAMSM X39256: Fig. 38) includes an 20 articulated portion of the posterior left mandible on which the articular is 21 For Review Only 22 exposed medioventrally because the posteromedial prearticular splint has 23 24 been broken away. The dorsal surface of the articular is subdivided by a 25 prominent transverse ridge into anterior and posterior areas. The former 26 27 is a smooth articular facet (gl) for the larger medial condyle (Fig. 29, ac) 28 29 of the quadrate. The glenoid, while cup-shaped medially, is bounded 30 laterally by a low oblique ridge that marks the lateral edge of the articular 31 32 bone. The lateral ridge is close, rather than being sutured, to the medial 33 34 wall of the surangular; the latter bone bears a less clearly defined lateral 35 36 extension of the glenoid (gl). The portion of the glenoid on the surangular 37 articulated with the oblique, lateral portion of the quadrate condylar 38 39 surface (Fig. 29, sac). In CAMSM X39256 the left quadrate and articular 40 41 can be physically articulated to confirm this articular relationship. 42 43 The posterior portion of the dorsal surface of the articular slopes 44 posteroventrally from the ridge that separates it from the glenoid area. 45 46 This surface tapers posteriorly and forms a major part of the retroarticular 47 48 process (rp); it comprises a rounded midline ridge on either side of which 49 the surface slopes away steeply and is roughened; this surface would 50 51 have been an area of insertion of m. depressor mandibulae (MDM). There 52 53 is a small foramen tucked just behind the dividing ridge on the medial 54 edge of the retroarticular process. The lateral margin of the articular 55 56 portion of the retroarticular process is closely bound to the medial surface 57 58 of the (splint-like) flattened, posterior extremity of the surangular. The 59 60
90 Page 91 of 215 Zoological Journal of the Linnean Society
1 2 3 posterior end of the surangular curves, medially, around the posterior end 4 5 of the articular. 6 7 The medial surface of the articular has a prominent dorsal ledge; 8 9 the dorsal surface of this ledge is pierced by a foramen. Beneath this 10 ledge, the articular is undercut so that the ledge overhangs the posterior 11 12 extension of the prearticular. The posterior end of the prearticular is 13 14 transversely compressed and its ventral edge rests upon the underlying 15 angular; both these bones appear to support, and contain, the ventral and 16 17 medial surfaces of the articular. 18 19 In both lectotype mandibles there is an unossified gap between the 20 21 anterior end of theFor articular Review and the adjacent Only edges of the surangular and 22 23 prearticular; this space was probably plugged by articular cartilage in life. 24 25 26 27
91 Zoological Journal of the Linnean Society Page 92 of 215
1 2 3 posterodorsal tip of the splenial is rounded and lies against the coronoid 4 5 on the medial surface near the base of the coronoid eminence (just 6 7 posterior to the last tooth position). Its medial surface is visibly striated 8 and this probably indicates a part of the area of insertion of the adductor 9 10 mandibulae complex. From this apex the dorsal edge descends in a 11 12 smooth curve that runs diagonally across the coronoid (Co) to a point 13 adjacent to tooth 14; it then runs anteriorly parallel to the ventral edge of 14 15 the latter bone, following the line of alveolar foramina on the medial side 16 17 of the dentary (compare Figs 36 & 37). Farther anteriorly, the dorsal 18 margin of the splenial is obscure on the left mandible – there seems to be 19 20 a mosaic of splenial fragments on the medial face of the dentary ramus. 21 For Review Only 22 The right mandible (Fig. 37B) has a slightly displaced splenial that 23 continues anteriorly for about two-thirds of the length of the dentition 24 25 before curving downward toward the ventral edge. It may well taper to a 26 27 point close to, if not at, the symphysis. 28 29 The ventral margin of the splenial follows the contour of the lower 30 edge of the mandible and further posteriorly wraps around the ventral 31 32 edge of the angular so that it may just be visible as a sliver in lateral 33 34 view. The splenial terminates as a pointed, flattened process lying in a 35 36 shallow facet (partly exposed – Fig. 37B) on the medial surface of the 37 angular at a point midway beneath the adductor fossa. From this area the 38 39 splenial curves forward and upward, crossing the prearticular; its edge 40 41 runs roughly parallel to the anterior margin of the adductor fossa before 42 meeting the dorsal tip. At the point at which the splenial transitions from 43 44 the suture with the angular to the prearticular the suture opens to form a 45 46 narrow, vertical slot between these two bones (imf); this feature is seen 47 on both mandibles and may indicate the presence of a remnant of an 48 49 infra-Meckelian fenestra/foramen (Romer 1956). 50 51 Angular (Figs 36-38). The angular bone (An) is partly obscured by a 52 53 raised mound of exostotic tissue that gives the impression of having 54 expanded radially from the centre of the angular. This outwardly 55 56 osteoderm-like feature is eroded and irregular on the lectotype mandibles 57 58 (compare Figs 36 & 37). The other large specimen (BRSMG LEGL 0004: 59 60
92 Page 93 of 215 Zoological Journal of the Linnean Society
1 2 3 Figs 14, 39) reveals a less damaged exostosis on the surface of the 4 5 angular. 6 7 The posterior portion of the angular is thick, curved and tapers to a 8 9 rounded point that terminates just short of the overlying surangular. This 10 part of the angular underplates and supports several postdentary bones: 11 12 the prearticular medially, the articular dorsally and the surangular 13 14 laterally. The dorsolateral edge of the angular displays a well-defined 15 suture with the surangular. This is best shown by the surangular of 16 17 BRSMG Ce12785 (Fig. 38B); the suture curves anteriorly (running parallel 18 19 to the upper margin of the surangular) and then slopes upward in a 20 gradual but more or less sigmoid path, which curves steadily 21 For Review Only 22 anteroventrally before meeting the dentary. The angular becomes a thin 23 24 sheet of bone that is applied to the lateral surface of surangular. On the 25 dentary ramus the angular develops a narrow, laterally compressed, 26 27 finger of bone that tapers out along the ventrolateral margin of the 28 29 dentary (Fig. 37A, ans). It also has a thin, tongue-in-groove ventral 30 contact with the edge of the splenial. Medially the angular underlies the 31 32 ventral edge of the prearticular and, a little farther anteriorly, is 33 34 overlapped by a thin posterior extension of the splenial. 35 36 The contacts between the posterior portion of the angular and the 37 38 surangular, articular and prearticular are well exposed in CAMSM X39256 39 (Fig. 38A). Although the lateral surface of the angular (where it is 40 41 preserved) is thick and its lateral surface smooth but irregularly mounded 42 43 it does not show clear indications of substantial exostotic growth. The 44 right and left sides of the articulated sub-adult skull (BRSMG LEGL 0004: 45 46 Figs 14, 39) reveal further details of the angular and its superficial bony 47 48 coating: the left angular (Fig. 14) is preserved in articulation with the 49 surangular, although the dentary is badly damaged. The right surangular 50 51 (Fig. 39) is partly disarticulated from the surangular and can be seen to 52 53 overlap the dentary ramus anteriorly. In both instances the exterior of the 54 angular is heavily textured. The centre of the angular is covered by thick, 55 56 rugose granular textured bone that forms a mound. Around this mound, 57 58 creating a sort of halo-effect, is a radiating pattern of bony strands that 59 cover the anterolateral exposure of the angular; however, posteriorly, 60
93 Zoological Journal of the Linnean Society Page 94 of 215
1 2 3 where the angular develops the thickened finger-like structure that 4 5 underlies the jaw joint, its external surface becomes smooth. The 6 7 uncoated surfaces correspond to areas where ligaments and muscles 8 controlling motion at the jaw joint would have been attached. Some 9 10 strands of superficial bone can also be seen on the body of the surangular, 11 12 just beneath the surangular foramen (on both sides of the skull). In the 13 lectotype, which is larger than the specimen just described, the superficial 14 15 bony texture on the angular appears to have spread so that it partly 16 17 encroaches on to adjacent surfaces of the surangular and dentary 18 19 20 21
94 Page 95 of 215 Zoological Journal of the Linnean Society
1 2 3 Posterior to this apex the dorsal surface forms a thick rounded bar that 4 5 arches posteroventrally; this arch is interrupted posteriorly by a boss-like 6 7 projection that flares laterally as a ledge and medially as a mound. The 8 lateral ledge overhangs a clearly marked surangular foramen (saf) that 9 10 runs as a canal anteroventrally into the rear corner of the adductor fossa. 11 12 An additional elliptical foramen is also present anteriorly, directly 13 underneath the arched bar, in the roof of the adductor fossa (Figs 36,37, 14 15 asf); this foramen exits from the surangular anteriorly, as a narrow slot 16 17 along the suture between the surangular and coronoid process of the 18 dentary. The anterior edge of the lateral ledge extends horizontally 19 20 forward as a narrow ridge along the outer wall of the surangular before 21 For Review Only 22 swinging dorsally toward the base of the coronoid process of the dentary. 23 The narrow ridge encloses a smooth dorsolateral surface on the 24 25 surangular that may have anchored a substantial bodenaponeurosis for 26 27 several of the mandibular adductor muscles. 28 29 In medial view the coronoid eminence forms a thick, cylindrical arch 30 of bone roofing the adductor fossa. The anterior end of this arch is 31 32 clamped between the coronoid and two processes of the dentary. A 33 34 considerable degree of complexity of the dentary-surangular suture is 35 36 expected since this is the main strain-accommodating junction during 37 biting: between the loading imposed upon the tooth-bearing dentary and 38 39 stress exerted upon the muscle-bearing postdentary region of the 40 41 mandible. 42 43 44 45
95 Zoological Journal of the Linnean Society Page 96 of 215
1 2 3 Farther anteriorly, the coronoid thins and narrows, turning into a 4 5 long strip of bone that lies medial to and screens the dentary alveolar 6 7 parapet. Farther anteriorly, on the left mandible, the top of the alveolar 8 parapet and emerging replacement crowns are just visible (Fig. 36). In 9 10 contrast, the right mandible shows the coronoid following the dorsal 11 12 margin of the alveolar parapet and tapering to a point near the base of 13 the most anterior tooth that is preserved; the parapet and alveolar 14 15 foramina are more clearly visible anteriorly (Fig. 37). 16 17 18 19 Dentary (Figs 36, 37, 40). The dentary (D) is the largest and stoutest of 20 21 the bones in the mandible.For Review In lateral view, Only its highest point is where it 22 23 develops a modest, spur-like dorsal projection (coronoid process) on the 24 anterior margin of the coronoid eminence. From this point, its posterior 25 26 edge descends obliquely forward against the surangular and then meets 27 28 the angular along a scarf suture. The ventral edge of the dentary is 29 slightly bowed (convex ventrally – Fig. 37A). The dorsal margin has a 30 31 slightly sinuous profile convex anteriorly, concave posteriorly, 32 33 complementing the sinuosity of the maxilla. The dorsal margin is scalloped 34 because the alveolar bone shapes itself to form partial collars around the 35 36 roots of the teeth. The external surface of the dentary is marked by some 37 38 traces of superficial bone that seem to have spread from a centre on the 39 angular. There are traces of superficial bony strands (Fig. 36, rug) on the 40 41 external surface of the dentary ramus beneath the prominent curved ridge 42 43 (be) that marks the ventral border of the buccal emargination; above this 44 ridge there are many large neurovascular foramina. 45 46 47 Much of the internal surface of the dentary is obscured by 48 postdentary bones in the lectotype. The alveolar parapet and alveolar 49 50 foramina are exposed on the medial surface of the dentary where the 51 52 coronoid has been displaced. Because of the anterior truncation of the 53 mandibles, they provide no evidence concerning the presence of a 54 55 predentary bone, or the nature of the dentary symphysis. By chance the 56 57 intermediate-sized specimen (BRSMG Ce12785: Fig. 40) includes a small 58 (25mm long) fragment from the anterior tip of the left dentary, which 59 60
96 Page 97 of 215 Zoological Journal of the Linnean Society
1 2 3 preserves the symphyseal suture, the alveoli and the roots or alveoli for 4 5 six (6) anterior teeth. 6 7 8 9
97 Zoological Journal of the Linnean Society Page 98 of 215
1 2 3 anterior end of the Meckelian groove (Mg). Beneath this slot, the surface, 4 5 down to its ventral margin, is marked by a series of prominent curved 6 7 ridges and rugose facets that represent ligament attachment areas that 8 bound together opposing dentaries at the symphysis (sym). 9 10 Dorsally (Fig. 40C), the positions for five teeth (three broken roots 11 12 (rt), two empty alveoli (alv)) can been seen in a closely packed row, these 13 14 follow a slight (laterally concave) curve. A sixth tooth root is preserved as 15 shards on the broken posterior edge of the dentary ramus. Because of the 16 17 torsion in the dentary, the alveoli/teeth run parallel to the plane of the 18 19 symphysis, rather than converging on the midline. This arrangement 20 complements that seen in the positioning of the teeth in the premaxilla. 21 For Review Only 22 Each tooth is surrounded by alveolar bone and there is a horizontal slot on 23 24 the medial side of the alveoli marking the break between the cortical bone 25 of the dentary and the alveolar trough packed with alveolar bone. The 26 27 internal alveolar wall is scalloped and narrow alveolar foramina (sf) are 28 29 visible. The alveolar scallops aligned along the dorsal edge of the dentary 30 lie slightly proud of the inner dentary wall (Fig. 40B, ap). Anteriorly, the 31 32 dorsal part of the dentary splays laterally so that the plane of its medial 33 34 wall faces dorsomedially. The symphysis forms a slightly irregular vertical 35 36 wall beneath the first three alveoli. The extreme anterior tip of the 37 dentary (sh) is marked by a short, tongue-like projection. 38 39 40 41 42 THE DENTITION 43 44 Premaxillary teeth: morphology and wear 45 46 Premaxilary teeth are not preserved with the lectotype because the 47 48 premaxillary portion of the skull is missing. The other large referred 49 specimen (BRSMG LEGL0004: Figs 14, 39) has eroded premaxillae and 50 51 has lost most of its premaxillary dentition. Fortunately the premaxillae are 52 53 nearly complete in the intermediate-sized individual (BRSMG Ce12785: 54 Fig. 41) and most of its teeth are also preserved. 55 56 Five premaxillary teeth are present (contra Barrett 2001) and 57 58 comprise a slightly heteromorphic array (Fig. 41). All these teeth have 59 60 stout roots that are elliptical rather than circular in cross-section; the
98 Page 99 of 215 Zoological Journal of the Linnean Society
1 2 3 roots contract apically, as they approach the enamel-coated crown, 4 5 creating a distinct ‘neck’ before re-expanding into the base of the crown. 6 7 The crown base is swollen mesiodistally as well as buccolingually and 8 there are narrow curved ledges on the mesial and distal corners of the 9 10 lingual side of the crown. Apical to this bulge, the crown contracts 11 12 gradually to form a roughly conical structure whose buccal and lingual 13 flanks are compressed, to form mesial and distal carinae; these carinae 14 15 can be denticulate, but not consistently so. The crown terminates apically 16 17 at a slightly recurved, narrow, round-ended tip. 18 19 20 21
99 Zoological Journal of the Linnean Society Page 100 of 215
1 2 3 than recurved-conical profile. However, most striking is the presence of a 4 5 straight line of eleven (11) small rounded denticles on its mesial carina 6 7 and in contrast six (6) larger and more prominent denticles on the distal 8 carina; these latter are more similar (though slightly smaller) to the 9 10 denticles on maxillary crowns. The base of each carina merges with the 11 12 swollen portion of the crown base. 13 14 There is no clear evidence of wear or abrasion facets on any of the 15 premaxillary crowns; however, the apical third of their enamelled surfaces 16 17 is smooth and polished, and more basally the surface texture of the 18 19 enamel is coarse. The exposed lingual surface of the apical portion of the 20 replacement crown (Fig. 41B, re) that is not yet in occlusion displays the 21 For Review Only 22 same coarse enamel texture seen on the basal portions of the crowns of 23 24 other premaxillary teeth. The polished appearance of the apical portions of 25 fully emerged premaxillary crowns is most likely a consequence of 26 27 repeated abrasion by foliage bitten by these anterior teeth prior to 28 29 pulping/chewing further back in the mouth. The tips of most premaxillary 30 crowns are smoothly rounded, whereas a few appear to be obliquely 31 32 fractured; these fractures may reflect excavation, preparation damage or 33 34 tooth-tooth impacts (perhaps even abrasion) during feeding. The detailed 35 36 structure of these broken surfaces is now being examined using high- 37 power microscopy (Norman & Porro, in preparation). 38 39 40 41 42 Maxillary teeth: morphology and wear 43 44 The maxillary dentition is exposed in several specimens: the lectotype 45 (Figs 42-45), the large sub-adult articulated skull and skeleton that has 46 47 been referred to this taxon (BRSMG LEGL 0004: Figs 14, 39) and the 48 49 intermediate-sized partial skull and skeleton (BRSMG Ce12785: Fig. 13). 50 The lectotype has 18 tooth positions preserved in its left maxilla (Fig. 42), 51 52 although the maxilla is incomplete anteriorly. There are 18 tooth positions 53 54 in the more complete maxillae of the slightly smaller referred specimen 55 (BRSMG LEGL 0004: Figs 14, 39) and there are 17 positions in the 56 57 complete right maxilla of the intermediate-sized specimen (BRSMG 58 59 60
100 Page 101 of 215 Zoological Journal of the Linnean Society
1 2 3 Ce12785: Fig. 13). Fully-grown individuals are unlikely to have had fewer 4 5 than 22 maxillary tooth positions. 6 7 8 9 BRSMG LEGL 0004 (Figs 14, 24, 39). 10 11 Because of its current state of preparation only the buccal aspect of the 12 13 maxillary and dentary dentitions can be examined. Teeth in the maxilla 14 15 show remarkably little variation in morphology along the length of the 16 17 dentition, except at the extreme posterior end where the last two teeth 18 seem to be more overtly recurved and the last is considerably smaller 19 20 than the rest. As with the premaxillary teeth, the root is elongate, robust, 21 For Review Only 22 elliptical in cross-section and merges into the base of the crown via a 23 modest waist. The crown, evenly coated with enamel, swells mesiodistally 24 25 as well as buccolingually; above this swollen region the crown has a 26 27 roughly trapezoidal outline in profile (the mesial denticulate margin being 28 slightly convex whereas the distal edge is straight). In their natural 29 30 positions within the dentition, crowns are arranged en echelon with the 31 32 mesial basal portion of each crown being overlapped, on their buccal side, 33 by the distal basal portion of the preceding crown. Viewed mesially or 34 35 distally each crown appears more asymmetrical because the apicobasal 36 37 axis is inclined lingually; this creates a mound (an almost knee-like 38 structure) at the base of the buccal surface of the crown that is well- 39 40 shown in BRSMG LEGL 0004 (Figs 14, 24, 39). 41 42 In buccal view the mesial and distal edges at the base of the crown 43 44 diverge at an acute angle and are comparatively prominent because each 45 is formed from a thickened roll of enamel. Apical to these structures the 46 47 mesial and distal edges of the crown are ornamented by prominent rows 48 49 of denticles ranging between 5-8 on the mesial edge and 6-8 on the distal 50 edge. Parallel buttressing ridges are prominent and extend from the 51 52 denticles down the surface of the crown before merging into the crown 53 54 surface. The buccal surface of the crowns (seen on both sides of the skull 55 of BRSMG LEGL 0004: Figs 14, 39), reveal no evidence of abrasion, and 56 57 collectively they form a remarkably even, palisade-like row of apparently 58 59 fully emerged crowns (with a minor exception near the anterior end of the 60
101 Zoological Journal of the Linnean Society Page 102 of 215
1 2 3 left maxillary dentition – Fig. 14). In striking contrast the dentition of the 4 5 left maxilla of the slightly larger lectotype specimen (Fig. 42A,B) exhibits 6 7 unevenness in the height and degree of eruption of crowns as well as 8 multiple examples of abrasion on the crown apices. Apart from minor 9 10 fracturing of the apical margin of a few crowns, there are no abrasion 11 12 facets on the buccal surface of maxillary crowns because of the overbite. 13 14 15 16
102 Page 103 of 215 Zoological Journal of the Linnean Society
1 2 3 Dentary teeth: morphology and wear 4 5 Seventeen tooth positions are preserved on the left mandible of the 6 7 lectotype (NHMUK R1111: Fig. 42C); the most anterior tooth is 8 9 represented by just an oblique cross-section through its root and 10 corresponding alveolus. It is obvious that the tooth count must have been 11 12 higher in the complete skull. In the slightly smaller skull of BRSMG LEGL 13 14 0004: Figs 14, 39) at least 21 tooth positions can be counted (with a fair 15 degree of confidence) but this also is incomplete. At least 26 dentary 16 17 tooth positions are likely to have been present in mature individuals. 18 19 20 21 For Review Only 22 BRSMG LEGL 0004 (Figs 14, 24). 23 24 The right dentary dentition is effectively missing; however, the posterior 25 portion of the left dentary dentition of this specimen is well preserved, 26 27 and a few of the anterior teeth are similarly well preserved (Fig. 14). What 28 29 (again) is striking about the dentition of this individual is its apparent 30 evenness and minimal evidence of tooth abrasion: most of the crowns 31 32 seem to be fully emerged and packed en echelon (just one replacement 33 34 crown can be seen in Figure 14). The dentition seems, in effect, to be 35 pristine. There are signs of abrasion (af) on a few crowns and these 36 37 isolated abrasion facets are generally spindle-shaped and restricted to the 38 39 most prominent part of the buccal surface: the ‘knee-like bulge at the 40 base of the crown (Figs 14, 39, af). 41 42 43 The general morphology of the teeth resembles that seen in the 44 maxillary dentition with few subtle differences. The roots are stout and 45 46 elliptical in cross-section and merge into the expanded base of the crown 47 48 without a clearly defined neck or waist (the root simply expands smoothly 49 into the base of the crown). The overall outline of each crown is 50 51 trapezoidal (inverted by comparison to the maxillary crowns): with a pair 52 53 of slightly thicker, divergent ledges mesially and distally that seem to 54 support the somewhat fan-like apical margin with its coarsely denticulate 55 56 carinae. The carinae are slightly asymmetrical: the mesial carina slightly 57 58 convex and the distal one straight or slightly concave. The marginal 59 denticles are prominent and tongue-shaped, and there are generally 6-7 60
103 Zoological Journal of the Linnean Society Page 104 of 215
1 2 3 of these denticles on the mesial carina and 5-6 on the distal carina. In this 4 5 specimen the denticles are ‘supported’ by buttressing ridges that extend 6 7 down the buccal surface of the crown, looking a little like close-packed 8 ribs. The denticle ridges are, again, not nearly so prominent in the 9 10 lectotype, but the crowns of the latter are much more heavily abraded so 11 12 a complete comparison is not possible. As was the case with the maxillary 13 teeth, the dentary crowns are offset (tilted) lingually and this creates a 14 15 similar mound or ‘knee’-like zone of flexure at the centre of the base of 16 17 the crown, which is precisely where vertical abrasion facets seem to be 18 located on just a few of the dentary teeth. 19 20 21 For Review Only 22 23
104 Page 105 of 215 Zoological Journal of the Linnean Society
1 2 3 complete, but displays a small triangular wear facet on the distal side of 4 5 its apical margin. Tooth 8 has a small break on its mesial ledge that may 6 7 be linked to contact between this ledge and the adjacent surface of crown 8 7. An extensive wear facet covers the distal half of the crown, and this has 9 10 extended mesially across the crown surface; basally the edge of this wear 11 12 facet is marked by an oblique lip. Tooth 9 has a crown that retains the 13 rounded mesial ledge, but its buccal surface displays a very large wear 14 15 facet that ends basally at an oblique lip. Tooth 10 is little worn, although 16 17 the distal denticulate carina is worn away, possibly indicating the first 18 stage in the development of a distal wear facet. Teeth 11 and 12 are 19 20 similarly heavily worn, both bearing large facets that cover major portions 21 For Review Only 22 of their buccal surfaces. Teeth 13 and 14 show damaged carinae, but are 23 unworn; this is also the case with the final two teeth (15 & 16), which are 24 25 less fully emerged than previous examples, and show unworn denticulate 26 27 carinae, and no obvious traces of abrasion on their buccal faces. 28 29 The influence of cranial and jaw morphology, tooth structure and 30 wear facets on the jaw action in this dinosaur are considered in the 31 32 Commentary and Discussion section below. 33 34 35 36 37 Tooth growth and replacement (maxillary dentition) 38 39 The lectotype (NHMUK R1111) provides some general information 40 concerning the pattern of tooth emergence and replacement in the left 41 42 maxilla only (Fig. 44). The dentary dentition yields less information about 43 44 growth and replacement because the alveolar region is masked by the 45 coronoid bone. The alveolar parapet of the right mandible is entirely 46 47 covered by the coronoid and splenial bones. The referred large specimen 48 49 (BRSMG LEGL 0004: Figs 14, 39) though well preserved can only be 50 examined externally, so successional growth patterns cannot be 51 52 determined. MicroCT scanning has been undertaken and processing of 53 54 these image files will permit a more thorough description of tooth growth 55 and the replacement pattern in this species (Norman & Porro, in 56 57 preparation). 58 59 60
105 Zoological Journal of the Linnean Society Page 106 of 215
1 2 3 Maxillary dentition. The internal surface of the left maxilla (Fig. 44) 4 5 reveals a row of elliptical alveolar foramina (sf) in a row parallel to the 6 7 dentition and interconnected by a shallow groove (g) that indicates the 8 channel for blood vessels and nerves supplying the dental lamina. The 9 10 alveolar parapet is more or less continuous, but is breached in one or two 11 12 places. The breaches represent the rapid eruption of replacement crowns 13 on the medial side of an alveolus and subsequent lateral migration of this 14 15 tooth into a vacated alveolus. The edges of the breaches in the alveolar 16 17 wall appear to be reforming prior to coalescing so that the alveolar 18 foramen is reformed prior to the growth and emergence of the next 19 20 replacement tooth. Successional pulses of emergence of teeth along the 21 For Review Only 22 length of the dentitions are present, judged by the pattern of eruption 23 that is visible along the maxillary dentition. 24 25 26 27 28
106 Page 107 of 215 Zoological Journal of the Linnean Society
1 2 3 within both dentitions will be clarified after the segmentation of MicroCT 4 5 scanned images of both jaws (Norman & Porro, in preparation). 6 7 8 9 EXTRACRANIAL OSTEOLOGY 10 11 12 13 14 Stapes (Fig. 34). A single example of the bony portion of the columella 15 auris (stapes) was recovered during extraction of the braincase of the 16 17 lectotype. It was reportedly found in position in the auditory recess (A.J. 18 19 Charig pers. comm. circa late 1970s), but I have been unable to trace a 20 photographic record of this. It appears that I sketched this bone in 21 For Review Only 22 position in the late 1970s (Fig. 25, st) but I cannot remember whether I 23 24 did this because I knew of its approximate location (it having been 25 removed). Because the skull had been almost completely freed of matrix 26 27 at the time I sketched it I strongly suspect that the stapes had already 28 29 been removed. 30 31 This specimen is well-preserved and figured in two views (Fig. 32 33 34A,B). The stapedial footplate (ftp) is expanded and has a slightly 34 oblique almost condylar appearance. Distal to the footplate, the shaft 35 36 contracts sharply before lengthening and thickening and then becoming 37 38 abruptly truncated. As preserved the stapes was quite clearly not long 39 enough to reach the position of the tympanic membrane (assumed to 40 41 have been located just posterior to the dorsal end of the quadrate), so the 42 43 truncation probably reflects the attachment site for a cartilaginous 44 extrastapedial extension (est) of the columellar shaft. 45 46 Stapes are delicate and rarely discovered; they have been reported 47 48 among ankylosaurs in Pinacosaurus (Maryańska 1977) and 49 50 Gargoyleosaurus (Carpenter et al. 1998). 51 52 53 54
58 nd 59 2 Ceratobranchial (Figs 14, 45). Several examples of an ossified, 60 curved rod-shaped portion of the hyoid skeleton are known. In two cases:
107 Zoological Journal of the Linnean Society Page 108 of 215
1 2 3 the lectotype (NHMUK R1111: Fig. 45B) and the referred specimen 4 5 (CAMSM X39256: Fig. 45A) these were recovered as distinctive bowed, 6 7 slender, isolated elements. In the other, referred, specimen (BRSMG LEGL 8 0004: Fig. 14, cb2) this bone is preserved in what is presumed to be close 9 10 to its natural position near the posteromedial corner of the mandible. In 11 12 conformity with earlier observations in ornithischians Norman (1980 and 13 references therein) referred to this bone as an ossified first 14 15 ceratobranchial. 16 17 Recent study of a well-preserved juvenile skull of the ankyosaurid 18 19 Pinacosaurus (Hill et al. 2015) exposed an articulated hyoid apparatus 20 lying between the mandibles. This new discovery indicates that the curved 21 For Review Only 22 elongate bones with ovoid articular surfaces at either end, and found 23 24 occasionally alongside the mandibles of dinosaurs, is more likely to be the 25 homologue of the second ceratobranchial. 26 27 28 In the lectotype the anterior end of this ceratobranchial is sheared 29 off obliquely (Fig. 45B, br). The remainder of this bone is bowed ventrally 30 31 and although comparatively slender, terminates posteriorly in an 32 33 expanded, slightly convex, articular pad (art). In the referred specimen 34 (Fig. 45A) this ceratobranchial is preserved (apparently complete) but it is 35 36 poorly preserved. Its anterior end is narrow, bluntly truncated and slightly 37 38 concave and was, in all probability, capped by an articular cartilage. This 39 would have allowed articulation with the hyoid body in the gular area 40 41 between the mandibles. These structures supported the tongue and 42 43 helped to anchor its associated musculature (see Hill et al. 2015). 44 45 46 47 Epistyloid bones (Figs 46, 47). A pair of well-preserved, elongate, blade- 48 49 ended bones (styl) can be seen projecting posteroventrally from behind 50 the shaft of the quadrate (Q) on the left side of the skull/anterior neck 51 52 block of the referred specimen BRSMG LEGL 0004 (Fig. 46). There is no 53 54 trace of equivalent bones on the other side of the block and it is concluded 55 that these bones are a pair that were originally positioned on either side 56 57 of the midline adjacent to the braincase and projecting diagonally 58 59 posteroventrally, as reconstructed in Figure 47. 60
108 Page 109 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5
109 Zoological Journal of the Linnean Society Page 110 of 215
1 2 3 Occipital osteoderms (Figs 16, 47). The remarkable preservation of the 4 5 dermal skeleton in the referred specimen (BRSMG LEGL 0004) has 6 7 revealed, for the first time, that the skull was ornamented by two large, 8 curved, occipital horns. These latter are each composed of two fused 9 10 osteoderms that are attached to the dorsal portion of the occiput on either 11 12 side of the midline (as shown in Figure 11). The deeper part of the horn 13 structure comprises a wedge-shaped base-plate pad (Fig. 16, ba), 14 15 whereas the more superficial and externally visible horn is referred to as 16 17 an osteoderm – both bones are of dermal origin. The occipital horns 18 project posterodorsally from the occiput (Figs 16, 47), diverge from the 19 20 midline, curve outward a little along their length and bear a rounded 21 For Review Only 22 carina laterally. There is little doubt that these osteoderms were capped 23 by epidermal keratinous horns (see Commentary and Discussion below 24 25 and Norman [Part 3]). 26 27 28 29 COMMENTARY AND DISCUSSION 30 31 32 Generalities 33 34 The skull of Scelidosaurus conforms, in its general anatomical 35 configuration, to that of almost any hypothetical model of that of a basal 36 37 dinosaur. Posteriorly, it resembles a rectangular box framed by the skull 38 39 table, temporal and suspensorial elements that combine to form a scaffold 40 that surrounds and anchors the braincase. This bony scaffold encloses 41 42 large apertures (the temporal fenestrae) for attachment and movement of 43 44 jaw the muscles and the housing of other soft tissues and organs. In front 45 of the orbits, the snout tapers gradually to a small rugose, toothless beak. 46 47 The mandible is robust and bar-like, supporting a sinuous palisade of 48 49 overlapping, leaf-shaped teeth and it can reasonably be inferred that the 50 mandible is similarly tipped by a small toothless beak that acts as the 51 52 counterpart of that seen in the upper jaw. 53 54 Unlike the generalised dinosaur skull, there is widespread evidence 55 56 of superficial bony tissue (exostoses) coating most of its external 57 58 surfaces; this suggests the existence in life of an external mosaic-like 59 covering of shield-like keratinous scales reminiscent of the keratinous 60
110 Page 111 of 215 Zoological Journal of the Linnean Society
1 2 3 plates seen covering the skulls of many extant chelonians. A prominent 4 5 brow-ridge overhangs each orbit, and a prominent pair of horns adorns 6 7 the occiput. The lateral walls of the snout house comparatively small, 8 shallow antorbital fossae and there is no evidence of an external 9 10 mandibular fenestra. The rostrum is rugose, and there is a step, in profile 11 12 view, that demarcates the small rostrum from the remainder of the 13 premaxillae; this creates the impression that the rostrum forms a discrete 14 15 ‘cap’ on the premaxillae. The premaxilla bears five teeth that exhibit mild 16 17 heterodonty: the anterior ones are pointed, recurved and have narrow 18 carinae, some of which bear the remnants of marginal serrations, whereas 19 20 the posterior ones transition toward the more leaf-shaped morphology 21 For Review Only 22 seen in the maxillary dentition. The lower sides of the snout are marked 23 by pronounced buccal emarginations that extend from the jugal flange 24 25 and coronoid eminence anteriorly to the premaxilla-maxilla suture. The 26 27 palate is deep but (with the exception of the vomers) shows no sign of 28 fusion of any of its component parts. The roof of the nasal cavity may 29 30 have been formed by epivomers (unique to Scelidosaurus); and, a 31 32 comparative rarity among dinosaurs, a conical epipterygoid is present 33 (sitting vertically and sutured to the dorsolateral edge of the quadrate 34 35 wing of the pterygoid). In ontogenetically mature individuals the occiput 36 37 becomes fused to form an obliquely inclined plate that anchors the 38 occipital horns and a median, nuchal osteoderm. The quadrate exhibits a 39 40 well-defined pit medial to the quadrate (paraquadrate) foramen in one 41 42 immature individual; there is a fossa associated with this pit on the 43 posterior surface of the shaft of the quadrate that resembles a remnant 44 45 pneumatic foramen (as previously reported in Heterodontosaurus). This 46 47 latter feature is lost in larger (more mature) individuals. Unusually, for 48 any ornithischian dinosaur skull, a slender, elongate, blade-ended bone 49 50 (the epistyloid) projects posteroventrally from the interior of the skull into 51 52 the proximal neck region on either side. These bones are not comparable 53 (structurally or positionally) to any known dinosaur ceratobranchials and 54 55 may also be unique to Scelidosaurus. 56 57 58 59 The premaxillary beak 60
111 Zoological Journal of the Linnean Society Page 112 of 215
1 2 3 The median edentulous rostral portion of the premaxillae has the outward 4 5 appearance and texture of an osteoderm (highly rugose, grooved and 6 7 foraminate). This structure lies against the premaxilla and is slightly 8 separated from it by the crease that runs transversely across the 9 10 premaxilla at the base of the medial dorsal processes (see Fig. 13). This 11 12 rostrum would have been covered by a keratinous rhamphotheca in life; 13 the growth of the latter was facilitated by the rich vascular supply to the 14 15 rostrum. The beak-supporting rostrum could have been derived either 16 17 from an osteoderm, or as an exostosis on the surface of the premaxillae. 18 Against the idea of the origin of the ornithischian premaxillary beak being 19 20 a median osteodermal ‘cap’ to the premaxilla, the rostrum has a median 21 For Review Only 22 suture line (Fig. 13B, pms); this suggests that either there was a pair of 23 capping osteoderms or that each premaxilla produced a rostral bony 24 25 exostosis that supported the growth of the keratinous beak. Whatever the 26 27 explanation, it is curious that the rostrum of Scelidosaurus (as opposed to 28 other basal ornithischians such as Lesothosaurus and Heterodontosaurus) 29 30 seems to form a more discrete projection anterior to the base of the 31 32 dorsomedial processes of the premaxillae. 33 34 35 36 Ornithischian supraorbital bones 37 38 Positionally, the endochondral bone that contributes to the dorsal margin 39 40 of the orbit, and is positioned between the prefrontal, frontal and 41 42 postorbital, in a wide range of reptiles, is identified as the postfrontal 43 (Romer 1956). Postfrontals are present in a range of non-dinosaurian 44 45 archosaurs (Romer 1956, Schoch 2007, Nesbitt 2011). Gilmore (1914), on 46 47 the advice of Robert Broom, identified a postfrontal in the skull roof of 48 Stegosaurus. However, it has become accepted (following the 49 50 comprehensive review by Romer) that the endochondral postfrontal is lost 51 52 in dinosaurs and a few other more derived archosaur clades. Romer 53 (1956) re-identified Gilmore’s ‘postfrontal’ in Stegosaurus as a neomorph 54 55 osteoderm, a supraorbital. 56 57 The identity and implied homology of palpebral/supraorbital bones 58 59 in ornithischians was reviewed by Maidment & Porro (2010). They posited 60
112 Page 113 of 215 Zoological Journal of the Linnean Society
1 2 3 that the dorsal orbit margin was flanked primitively (using Lesothosaurus 4 5 and Heterodontosaurus as basal representatives), by a slender arched 6 7 palpebral bone that projects posterodorsally (chord-like) across the orbit. 8 This bone has a flattened basal region that articulates against the 9 10 prefrontal. Rugosities along the medial edge of the tapering posterior 11 12 palpebral process and the adjacent orbit margin indicate the existence of 13 a connective tissue sheet that spanned this gap. Additionally, they noted 14 15 that a few ornithischians are known to have single ossifications positioned 16 17 immediately distal (posterior) to the palpebral process and lying against 18 the postorbital (e.g. Norman 1980: figs 2,3,11). Ossification of the 19 20 connective tissue-filled gap between the palpebral process and the lateral 21 For Review Only 22 (orbital) margin of the frontal would account for the formation of a 23 neomorph middle supraorbital; this structural conformation agrees with 24 25 the pattern observed around the orbital margin of Scelidosaurus, as well 26 27 as with the textural changes noted in the middle supraorbital during its 28 apparent ontogeny: BRSMG Ce12785>BRSMG LEGL 0004>NHMUK R1111. 29 30 The pattern of three supraorbital bones: palpebral, middle 31 32 supraorbital and posterior supraorbital is seen consistently within 33 34 thyreophorans (or at least insofar as those in which their sutures are 35 36 visible) and with varying modifications in other ornithischian sub-clades 37 (Maidment & Porro 2010). What is curious, and of potential relevance in a 38 39 general ‘pre-dinosaur’ phylogenetic context, is the description by Schoch 40 41 (2007: fig. 8) of three bones identified as supraorbitals positioned in a 42 linear array near the dorsal orbital margin in the skull of the osteoderm- 43 44 covered Norian (Late Triassic) pseudosuchian Aetosaurus ferratus. 45 46 In small, highly cursorial ornithischians such as Lesothosaurus (see 47 48 Porro et al. 2015) and Heterodontosaurus (Norman et al. 2011) the skull 49 is small and compact and the orbits are extremely large open pockets on 50 51 the side of the face. The elongate, slender palpebral bone projects across 52 53 the dorsal part of the orbit toward the postorbital; the latter is 54 correspondingly thickened adjacent to the distal end of the palpebral 55 56 process and forms a modest almost tab-like structure suggestive of an 57 58 attachment site for ligaments connecting the tip of the palpebral process 59 to the adjacent surface of the postorbital. The chord-like arrangement of 60
113 Zoological Journal of the Linnean Society Page 114 of 215
1 2 3 the palpebral creates what is structurally, a strap (harness-like structure) 4 5 that might have been capable of stabilising the eyeball within such wide 6 7 and very open orbital cavities. 8 9 In Scelidosaurus – although the structural layout of the 10 supraorbitals was not correctly recognised by Maidment & Porro (2010: 11 12 fig. 4C,D) they correctly deduced that three supraorbitals were likely to 13 14 have been present – the palpebral has an enlarged base that is anchored 15 to the lateral surface of the prefrontal and has an arched posteriorly 16 17 directed process that is clearly anchored to the orbital margin of the 18 19 prefrontal. The palpebral process traces the margin of the orbit, rather 20 than being oriented chord-like across the orbital cavity. The palpebral 21 For Review Only 22 process extends toward the postorbital where it meets, and sutures 23 24 against (and ultimately fuses with) the posterior supraorbital. The 25 combined palpebral-posterior supraorbital form a prominent shelf or ‘brow 26 27 ridge’ that can be appreciated to have had an overtly protective function 28 29 and, more conjecturally, a behavioural role. 30 31 32 33 Postorbital osteoderms and ornament 34 35 Perched on the dorsolateral corner of the postorbital is a pup-tent-shaped 36 37 posterior supraorbital. This element appears to be osteodermal and is 38 39 sutured to the palpebral process. The lateral surface of the postorbital is 40 seemingly concave (as appears to be the condition in basal dinosaurs and 41 42 dinosauriforms). In Scelidosaurus the lateral postorbital concavity is filled 43 44 by a crater-shaped osteoderm, judged by the interface between 45 endochondral and dermal bone revealed as a line of vascular foramina on 46 47 the orbital margin. The posterior and ventral edges of this osteoderm 48 49 seem to merge with the external surface of the jugal process of the 50 postorbital creating the appearance of an exostosis giving the impression 51 52 that these two forms of superficial bone formation can become intimately 53 54 linked in particular regions of the skull. Nevertheless, the dorsal portion of 55 the postorbital osteoderm thickens and creates a horizontal ledge and, 56 57 adjacent to the orbit margin, a stubby prong or ‘hook’ (Figs 17, 19, hk); 58 59 60
114 Page 115 of 215 Zoological Journal of the Linnean Society
1 2 3 the ledge and hook combine to form a sutural facet for attachment of the 4 5 posterior supraorbital. 6 7 8 9 Epivomers 10 11 Referring to these bones as epivomers implies the existence of a unique 12 13 (or previously unrecognized) pair of bones that articulate against the 14 15 dorsal edge of the vomers and (partly) roof the nasal cavity. It is 16 17 unfortunate that no detailed record of the precise position of these 18 symmetrical, thin and fragile bones was made at the time of their 19 20 exposure. It is clear that these bones have a degree of symmetry as well 21 For Review Only 22 as finished edges indicating that they are not random broken fragments 23 derived from some other bones of the snout or palate. They can be 24 25 manipulated so that they articulate snugly against what appear to be 26 27 complementary sutural surfaces running along the anterodorsal edges of 28 the vomers. If correctly interpreted (from my original rough sketches) 29 30 these bones are unlikely to be septomaxillae because such bones are 31 32 normally positioned either anterolaterally: very close to the external 33 nares, and in some instances contacting the lacrimal posteriorly, or they 34 35 form part of the floor of the nasal cavity – often associated with the 36 37 roofing for a vomeronasal organ (Romer 1956);. 38 39 In two remarkably well-preserved Cretaceous ankylosaurids: 40 Saichania and Pinacosaurus Maryańska (1977) was able to describe nasal 41 42 conchae and maxilloturbinal structures in their nasal cavities. However, 43 44 the structures she described bear no topographic (or structural) similarity 45 to the bones described and named herein as epivomers. 46 47 48 49 50 Sclerotic ring 51 52 There was a sclerotic ring in the orbit of Scelidosaurus. A few sclerotic 53 54 bones have been recovered from the acid-prepared lectotype skull, 55 adhering to the internal surface of the left jugal (Fig. 21B) and a few 56 57 others have been found among the debris collected during preparation of 58 59 the skull (one was found attached to the parasphenoid rostrum - 60
115 Zoological Journal of the Linnean Society Page 116 of 215
1 2 3 cultriform process). These few scattered remnants are insufficent to be 4 5 able to model or reconstruct the sclerotic ring of this dinosaur. 6 7 8 9 Predentary (Fig. 8) 10 11 A predentary bone has never been reported in Scelidosaurus and it seems 12 13 paradoxical that the two best-preserved skulls (NHMUK R1111 and 14 15 BRSMG LEGL 0004) are missing the anterior ends of their mandibles. 16 17 Nevertheless, circumstantial evidence implies that a small keratinous beak 18 was present and capped the mandibular symphysis. Based upon the 19 20 structure of the fragment described in the intermediate-sized individual 21 For Review Only 22 (BRSMG Ce12785 – Fig. 40) the dentary rami meet at a symphysis 23 ventrally. However, the tooth rows above the symphyseal area do not 24 25 converge on the midline (Fig. 40C). The anterior portions of the left and 26 27 right tooth rows remain parallel to one another, separated by a gap 28 created by the lateral flaring of the dorsal edges of each dentary ramus. 29 30 The gap between the anterior ends of the dentary rami is framed by 31 32 rugosities on the exposed surfaces of each dentary, the presence of a 33 34 ledge backed by vascular foramina (indicating a significant blood supply to 35 this area), and a pair of stout (tenon-like) processes project from the 36 37 anterior end of each dentary. There is clearly space, as well as the 38 39 provision of mechanical support, for a comparatively small crescentic 40 predentary to cap the end of the dentaries. The numerous foramina 41 42 surrounding the anterior end of the dentary are suggestive of an adequate 43 44 blood supply for the growth and active renewal (in response to attrition) 45 of a keratinous beak (rhamphotheca). The absence of any substance to 46 47 the inferred predentary suggests that this element was either lost prior to 48 49 collection or was still unossified in this ontogenetically immature specimen 50 51 These observations and inferences accord with the structure of the 52 53 upper jaw: the anterior tip of the premaxillae forms a small downturned 54 edentulous lobe that evidently supported a keratinous beak 55 56 (rhamphotheca) that opposed an equivalent structure on the lower jaw. 57 58 59 60
116 Page 117 of 215 Zoological Journal of the Linnean Society
1 2 3 Was there a mandibular fenestra? 4 5 There is no evidence, in any of the specimens of Scelidosaurus recovered 6 7 to date, of an archosaurian-dinosaurian external mandibular fenestra. This 8 9 view is reinforced by the structure of the medial wall of the mandible that 10 can be seen through the adductor fossa. This report appears to contrast 11 12 with the anatomy of the mandible of the closely similar (in anatomy and 13 14 geological age) thyreophoran Emausaurus. Haubold (1990: fig. 2) 15 illustrated a skull reconstruction with a large external mandibular fenestra. 16 17 Judging from the condition of the preserved remnants of the lower jaw of 18 19 the latter genus, which lacks much of the lower half of the mandibular 20 ramus, the presence of an external mandibular fenestra is not 21 For Review Only 22 unequivocally established. 23 24 External mandibular fenestrae are, nevertheless, present in the 25 26 well-preserved skulls of the basal ornithischians Lesothosaurus (Porro et 27 28 al. 2015) and Heterodontosaurus (Norman et al. 2011). It is also the case 29 that some stegosaurs (of which Huayangosaurus and Stegosaurus are 30 31 well-known examples) retain an external mandibular fenestra. However, 32 33 ankylosaurs resemble Scelidosaurus in show no evidence of a mandibular 34 fenestra. However, it is the case that the lateral surface of the mandible is 35 36 invariably obscured by the presence of a large osteoderm. 37 38 39 40 Epipterygoid 41 42 43 Epipterygoids are rarely preserved in dinosaurs (Weishampel, Dodson & 44 Osmólska 2004). In diapsids more generally (particularly squamates) the 45 46 bone is a slender rod that forms the equivalent of a strut or ‘spacer’. This 47 48 bone has articular condyles at both ends and spans the gap between the 49 pterygoid and the overlying braincase. The epipterygoid acts as a ‘spacer’ 50 51 during palinal movements of the pterygoids, which act as push-rod 52 53 linkages connecting the snout and quadrate. Streptostylic rotation of the 54 quadrate drives the concomitant elevation-depression of the snout and 55 56 the epipterygoids prevent the palate-pterygoid complex from being drawn 57 58 dorsally against the floor of the braincase. The retention of epipterygoids 59 that are evidently fixed securely to (rather than jointed to) the dorsal 60
117 Zoological Journal of the Linnean Society Page 118 of 215
1 2 3 edge of the pterygoids in Scelidosaurus is totally unexpected. The skull of 4 5 Scelidosaurus is quite clearly akinetic, so the snout could not be elevated 6 7 and depressed independently of the skull table – as previously observed 8 by Barrett (2001). The Scelidosaurus epipterygoid has a spire-like dorsal 9 10 process with no articular condyle, so its function is not at all clear (unless 11 12 it anchored some residual protractor pterygoideus musculature). At 13 present I tentatively consider this to be an example of an atavism because 14 15 it is not clear to me what role (if any) this bone might have performed in 16 17 an akinetic skull such as that of Scelidosaurus. 18 19 Comparisons. Within thyreophorans epipterygoids have been 20 reported only in a few ankylosaurs: Euoplocephalus, Saichania, Tarchia 21 For Review Only 22 and juvenile Pinacosaurus (Maryańska 1977, Tumanova 1987, Coombs & 23 24 Maryańska 1990). However, they differ radically from the one described in 25 Scelidosaurus because their shafts have been described as lying obliquely 26 27 against the lateral wall of the braincase and being expanded at both ends. 28 29 30 31 32 Slot on the medial wall of the pterygoid 33 34 The remarkably detailed preservation of the palate bones of the lectotype 35 (NHMUK R1111) reveals the existence of a narrow, slot-like pocket on the 36 37 medioventral wall of the quadrate wing of the pterygoid (Fig. 28B, sl). 38 39 This feature appears to be unique to Scelidosaurus, but its function is 40 unclear. 41 42 43 44 45 Fusion of the occipital plate 46 47 The dorsal margin of the paroccipital process of the lectotype alone 48 49 reveals the presence of a spur-like process (Fig. 33, psp); this process 50 partly encloses an obliquely oriented channel that runs anterolaterally 51 52 from the nuchal region through the occipital plate into the adductor 53 54 chamber. This is, in all probability, a remnant of the post-temporal 55 fenestra (ptf) that permitted venous/lymphatic drainage from the neck 56 57 musculature into the braincase via the vena capitis dorsalis. The occipital 58 59 surface in the only specimen known in which this area is visible (the 60
118 Page 119 of 215 Zoological Journal of the Linnean Society
1 2 3 lectotype – Fig. 33) appears to be completely fused (unless this is an 4 5 artefact of preparation and the application of many layers of consolidant). 6 7 Fusion and a smearing of exostotic bone is considered probable because 8 the occiput supports and secures the two large occipital horns dorsally, as 9 10 well as providing anchorage for a median nuchal osteoderm (visible in the 11 12 referred specimen BRSMG LEGL 0004, and represented by a base-plate in 13 CAMSM X39256 – see Norman, [Part 3]). 14 15 The juvenile specimen (CAMSM X39256 – Fig. 32) reveals part of 16 17 the unfused braincase of Scelidosaurus. This specimen indicates the 18 19 possibility of at least one additional canal penetrating the occiput 20 dorsomedially along the suture between the paroccipital and 21 For Review Only 22 supraoccipital; this canal becomes occluded as the occiput fuses over later 23 24 in ontogeny, judged by the structure of the occiput in the lectotype (Fig. 25 33). 26 27 28 29 30 Dental variation 31 32 The two large individuals of Scelidosaurus (NHMUK R1111 and BRSMG 33 34 LEGL 0004) exhibit remarkable differences in their dentitions. The 35 lectotype (NHMUK R1111: Figs 42-44) exhibits the morphological 36 37 characteristics of what can be considered to be a ‘normal’ dentition in the 38 39 sense that it exhibits teeth at various stages of eruption and replacement 40 along the dentition, as well as a range of abrasion surfaces (tooth-food- 41 42 tooth) created during on-going jaw action. The abrasion facets include 43 44 tooth-tooth occlusion, as revealed by the macro- and micro-wear features 45 described by Barrett (2001). In contrast, the referred specimen (BRSMG 46 47 LEGL 0004: Figs 14, 39) displays a dentition and abrasion characteristics 48 49 that do not conform to those seen in the lectotype. Overall, maxillary and 50 dentary crowns are arranged in the same way as in the lectotype: deeply 51 52 inset dentitions are bowed lingually and sinuous along their length (in 53 54 profile view). Unlike the lectotype, the upper and lower dentitions appear 55 uniformly fully erupted and show little evidence of abrasion. The maxillary 56 57 dentition appears to be pristine (unabraded): all the crowns seem to be 58 59 complete, with their apices and denticulate margins undamaged. It is 60
119 Zoological Journal of the Linnean Society Page 120 of 215
1 2 3 unfortunate that its current state of preservation means that the maxillary 4 5 dentitions cannot be examined lingually. The left dentary dentition (only 6 7 visible in buccal aspect) has a very similar appearance and has only one 8 obvious example of a tooth in ‘replacement mode’. Wear facets are 9 10 present on a small number of crowns, but these are entirely restricted to 11 12 the central mounded area at the base of the crown and appear as smooth, 13 spindle-shaped facets with no development of the oblique lip or ledge 14 15 seen on the dentary crowns of the lectotype. 16 17 How such similar-sized individuals (the lectotype is a little larger 18 19 than the referred specimen) can have such different dentitions has proved 20 to be challenging to explain. Thulborn (1978) proposed seasonal 21 For Review Only 22 aestivation as an opportunity for the mass replacement of 23 24 heterodontosaur dentitions. However, this novel suggestion has been 25 effectively refuted (Hopson 1980) and there is no evidence to support 26 27 such extreme seasonality in the Liassic ecosystem. 28 29 30 31 32 Inferred jaw action in Scelidosaurus 33 34 Structural constraints upon jaw motion. Both upper and lower dentitions 35 are notable for the degree to which they are inset, relative to the lateral 36 37 surface of the skull: there is an obviously deep buccal recess and the oral 38 39 cavity is (comparatively) narrow relative to the overall width of the skull 40 dorsal to the jaws. The tilting of the crowns relative to their roots, 41 42 observed in both maxillary and dentary teeth (both tilted lingually), 43 44 together with the modified conical morphology of the crowns, impose a 45 measure of constraint upon the pattern of occlusion exhibited by 46 47 individual teeth (as well as the mechanics associated with jaw action in 48 49 this species). Adding to the constraint implied by the structure of the 50 individual teeth in opposing dentitions, the upper and lower dentitions are 51 52 markedly bowed lingually. Consequently, the shape of the entire dentition 53 54 acts as a large-scale structural ‘guide’ limiting, or most likely removing, 55 the possibility of palinal (fore-aft) displacement of the lower jaw during 56 57 the later stages of occlusion. Furthermore, along the fore-aft length of the 58 59 dentitions, the occlusal edges display complementary sinuous profiles; this 60
120 Page 121 of 215 Zoological Journal of the Linnean Society
1 2 3 will also restrict the motion of the lower jaw relative to the upper one 4 5 during occlusion. 6 7 8 9 Abrasion features on the crowns of the teeth. Some dental crowns 10 11 (particularly in the lectotype – Barrett 2001) exhibit damage: low-angle 12 13 chips or irregular abrasion on their apices and/or denticulate carinae. 14 Rather than simply attributing such features to accidental damage 15 16 associated with mechanical preparation (less likely in the case of the 17 18 lectotype), it is possible that this minor damage was the result of 19 opposing crown contact during the initiation of jaw closure prior to the 20 21 jaws becoming ‘correctly’For alignedReview for full closure. Only 22 23 The general pattern of tooth abrasion seen in the lectotype 24 25 (reflecting a combination of tooth-tooth and tooth-food-tooth occlusion) is 26 of two general types. Maxillary crowns display zones of predominantly 27 28 high-angle apical abrasion on the lingual surface of individual crowns. The 29 30 majority of these crowns bear single abrasion facets, with a few 31 suggestive of confluent facets shared between two crowns. Dentary 32 33 crowns show the development of equally high-angle abrasion facets on 34 35 the distal half of the crown (predominantly) and, again, a few share 36 37 overlapping facets. As is the case in the maxillary dentition, not all the 38 teeth along the dentition exhibit abrasion: wear is evidently patchily 39 40 distributed along the dentition. The abrasion facets on dentary crowns, 41 42 when well developed, extend on to the swollen base of the crown and a lip 43 (a small, oblique shelf) is usually present on the basal edge of this facet 44 45 creating a partial puncture-crushing feature, as described by Barrett 46 47 (2001). 48 49 50 51 Jaw action. The range of movements available to the lower jaw is 52 53 controlled, or influenced, by a variety of factors. Fundamentally, this is a 54 sauropsid, so the bony framework of the skull and the overall pattern of 55 56 its underlying musculature obliges the jaws to close isognathically (the 57 58 dentitions on both sides of the skull occluded simultaneously, rather than 59 independently on either side – Norman & Weishampel 1985). Some rare 60
121 Zoological Journal of the Linnean Society Page 122 of 215
1 2 3 exceptions to this general rule have been described among crocodilians 4 5 (Ösi & Weishampel 2009) but these are truly exceptional. Within this 6 7 general cranial bauplan a number of additional factors influence the 8 induction of jaw action in this species: a) The presence or absence of 9 10 intracranial kinetism or streptostyly; b) The structure of the jaw 11 12 articulation; c) The orientation (and sum of the lines of action) of the 13 principal adductor muscles, as well as their points of attachment to the 14 15 mandible; d) The positioning of the teeth in the jaws and the overall form 16 17 of the dentition; and, e) The strength and stability of the symphyseal 18 region between the mandibles. 19 20 a. Kinetism/Streptostyly. The skull roof is coated extensively by 21 For Review Only 22 exostoses, and there is no evidence of transverse kinetic hinges between 23 24 any of its component bones (reconfirming the observations of Barrett 25 2001). The quadrate head fits into a deep socket on the squamosal and is 26 27 clamped in position by an elongate tapering process that extends from the 28 29 body of the squamosal down the jugal wing of the quadrate. Additionally, 30 a vertical buttress supports the posterior portion of the quadrate head; 31 32 this buttress is ligamentously bound to the paroccipital process and an 33 34 oblique posterior flange of the squamosal. The jugal arch is firmly sutured 35 36 to the facial bones anteriorly as well as the ventral portion of the jugal 37 wing of the quadrate posteriorly, effectively locking the quadrate in 38 39 position and thereby preventing conventional steptostylic (fore-aft) 40 41 displacement of the latter. The quadrate head is small, smooth and 42 convex (i.e. it looks to be ‘articular’) but the quadrate shaft could not 43 44 swing in a typically streptostylic manner. It is considered to be 45 46 theoretically possible that the quadrate shaft could swing pleurokinetically, 47 i.e. it was able to (flex) transversely, allowing slight/subtle lateral-medial 48 49 displacement of its distal articular condyles and, therefore, of the 50 51 mandible itself. 52 53 b. Jaw articulation. The jaw articulation comprises a nearly 54 spherical medial condyle of the quadrate (ac) that articulates against the 55 56 articular bone (see also Barrett 2001). However, the quadrate condyle 57 58 also projects obliquely laterally and this portion forms a tapering articular 59 surface (sac) partly separated from the medial condyle by a shallow 60
122 Page 123 of 215 Zoological Journal of the Linnean Society
1 2 3 groove. This lateral extension of the condylar surface contacts a facet on 4 5 the medial surface of the surangular, adjacent to the articular facet. The 6 7 total transverse width of the articular surface of the quadrate, its partial 8 subdivision by a groove, and the involvement of the medial wall of the 9 10 surangular creates a partial trochlea. This structural complexity would 11 12 have permitted orthogonal jaw motion (simple, uni-axial hinging) and 13 would have also been capable of resisting medial rotation (torsion) in the 14 15 long axis of the mandible. 16 17 The jaw joint is offset ventral to the plane of occlusion of the main 18 19 posterior portion of the dentition; however, near the mandibular 20 symphysis the plane of occlusion would have lain in line with the jaw 21 For Review Only 22 articulation because the profile of the mandibular dentition can be seen to 23 24 undulate along its length. 25 26 c. Jaw musculature orientation and lines of action. The principal jaw 27 28 adductor muscles (MAMES, MAMEM, MAMEP, MAMP, MPST) have their 29 origin on the lateral walls of the braincase and adjacent temporal arches 30 31 in ornithischian dinosaurs (e.g. Ostrom 1961). The muscles run, from 32 33 their cranial areas of origin, diagonally anteroventrally and/or 34 anterolaterally to insert principally upon the coronoid eminence (either 35 36 directly, or via a common bodenaponeurosis) as well as around the 37 38 adductor fossa. While such muscles are primarily responsible for exerting 39 forces that result in orthogonal (vertical) motion of the jaw, each 40 41 mandible is also subjected to medially directed forces: dorsomedial 42 43 vectors reflecting the sum of the lines of action of these jaw muscles and 44 the dorsolateral location upon the mandible of the muscle insertions. A 45 46 number of attempts have been made to reconstruct the jaw musculature 47 48 of ankylosaurs (Haas 1969, Holliday 2009, Carpenter et al. 2011, Ösi et 49 al. 2014, 2016). Jaw musculature and jaw action in Scelidosaurus will be 50 51 discussed elsewhere (Norman [Part 4]). 52 53 d. Tooth positioning. The dentitions of both upper and lower jaws 54 55 are arranged along the medial edges of the maxilla and dentary 56 57 respectively. In the case of each mandible that means that, in view of the 58 bowing of the dental row, a substantial portion of the dentition is located 59 60 medial to the principal area of attachment of the adductor musculature.
123 Zoological Journal of the Linnean Society Page 124 of 215
1 2 3 As the jaws close and the teeth occlude the teeth in the mandible 4 5 encounter orthal (vertical) and medial force vectors: a combination of 6 7 adductor muscle vectors and the reaction forces generated by the 8 pressure generated on the occlusal surfaces of the crowns if the teeth are 9 10 to achieve an effective power-stroke and shearing action. 11 12 Structurally, the medial (lingual) bowing of both dentitions requires 13 14 a precise positioning of the mandibular dentition against the maxillary 15 arcade for effective (unobstructed) occlusion between each dentition. 16 17 Propalinal movement (anterior or posterior) would misalign the two dental 18 19 arcades and generate potentially destructive mal-occlusions and might 20 even lock the dentitions together (at least temporarily). The inability of 21 For Review Only 22 the quadrate shaft to pivot anteroposteriorly on the squamosal cotylus, as 23 24 well as the essentially uni-axial hinge structure of the mandibular joint, 25 minimises propalinal displacement. 26 27 28 e. Mandibular symphysis. The only example currently known of the 29 mandibular symphysis and adjacent structures is seen in an immature 30 31 individual (BRSMG Ce12785: Fig. 40). Although the crowns of the teeth 32 33 are not preserved, the roots and alveoli show clearly that the dentitions 34 did not converge in order to meet in the midline above the symphysis. The 35 36 dorsal edges of the anterior ends of the dentaries straighten, producing 37 38 parallel dentitions beneath which the dentary rami twist and expand 39 medially in order to meet at a sutural surface. The latter has an 40 41 undulatory structure and shows no obvious sign of fusion between the 42 43 mandibular rami. Because this was a juvenile specimen, mandibular fusion 44 would not necessarily have been expected. In addition, there is 45 46 circumstantial evidence supporting the existence of a predentary (whether 47 48 cartilaginous, fibrous or bony) that capped and presumably stablilised the 49 mandibles by forming a clamp across the tips of both dentaries. 50 51 52 53 54 Jaw action summarised 55 56 Anteroposterior pivoting of the quadrate (conventional streptostyly) did 57 58 not occur. Slight movement was possible of the quadrate head within the 59 squamosal cotylus that would have allowed the mandibular jaw joint to be 60
124 Page 125 of 215 Zoological Journal of the Linnean Society
1 2 3 displaced mediolaterally. This motion may have facilitated subtle passive 4 5 re-positioning of the mandibles during occlusion. The structure of the 6 7 mandibular joint indicates that each mandible hinged uni-axially against 8 the quadrate condyle. The lateral expansion of the condylar surface, which 9 10 contacted the surangular, resisted medial rotation (torsion) that was 11 12 induced in the mandible during jaw closure. The oblique orientation of the 13 principal mandibular adductor muscles and their insertion on the coronoid 14 15 eminence promoted principally orthogonal motion of the mandible along 16 17 with a lesser (medial) rotatory couple. The positioning of the teeth 18 (medially offset, relative to the longitudinal axis of the tooth-bearing 19 20 bones) as well as the shape of the opposing dentitions had several 21 For Review Only 22 predictable effects. The over-bite of the maxillary dentition imposed a 23 medially-directed force (torsional couple) on each mandible during 24 25 isognathic occlusion. The bowing seen along the length of both dentitions 26 27 necessitated accurate (fore-aft) positioning of mandibular and maxillary 28 dentitions during biting, to ensure that opposing dentitions engaged 29 30 correctly. The sinuous profile of each dentition altered the mechanics of 31 32 occlusion along the length of the dentition because the anterior dentition 33 is at the level of the jaw articulation (facilitating a scissors-like action). In 34 35 contrast, the posterior dentition is dorsally displaced: so the posterior 36 37 portion of the dentitions would have tended to occlude simultaneously as 38 a block, as observed in specialist herbivores (Norman & Weishampel 39 40 1985). This configuration suggests a partial differentiation of function: the 41 42 anterior dentition provided an orthal slicing apparatus whereas the 43 posterior dentition was more involved in pulping and crushing food prior 44 45 to swallowing. Repeated occlusion particularly at the rear of the mouth 46 47 would have necessitated the presence of fleshy cheeks to enclose the 48 buccal recess and restrict food loss during each cycle of occlusion (Lull & 49 50 Wright 1942, Ostrom 1961, Galton 1973, Norman & Weishampel 1985). 51 52 The symphyseal region of the mandible was unfused in 53 54 ontogenetically immature specimens and it is considered unlikely that 55 56 fusion occurred in adults because the inferred dynamics of jaw closure 57 imply some passive axial mandibular torsion to accommodate the reaction 58 59 forces at the occlusal surfaces. Torsion that occurred during biting would 60
125 Zoological Journal of the Linnean Society Page 126 of 215
1 2 3 generate elastic strain energy in the ligaments binding the mandibles at 4 5 the symphysis and would be released during abduction, so the mandibles 6 7 would be restored to their pre-occlusal configuration. The undulatory 8 pattern seen on the symphysis of the immature specimen tends to 9 10 support the hypothesized mobility of this joint. 11 12 Comparative observations. Rybczynski & Vickaryous (2001) 13 14 described the jaw action in the ankyosaur Euoplocephalus in a study that 15 also employed visualisation of the microwear on the abrasion facets on its 16 17 teeth. They revealed, contrary to previous interpretations (e.g. Russell 18 19 1940, Haas 1969, Coombs & Maryańska 1990) that Euoplocephalus 20 utilised a power-stroke that involved muscle-powered lower jaw 21 For Review Only 22 retraction. The plane of shear between opposing teeth ran along a line 23 24 parallel to the long axis of the mandible. The power-stroke also involved a 25 small degree of bilateral medial pivoting of the mandibles that was made 26 27 possible by having a comparatively loose (mobile) dentary-predentary 28 29 joint. 30 31 Although microwear features on their teeth do not support a 32 33 retractive power-stroke in Scelidosaurus, the possibility that both 34 Scelidosaurus and Euoplocephalus used long-axis passive mandibular 35 36 pivoting as part of their jaw closure mechanism is at least suggestive of a 37 38 shared inheritance. However, it is considered likely that this latter 39 mechanism may be widespread in the clade Ornithischia. 40 41 42 43 44 Epistyloid bones 45 46 The orientation and positioning of the distal portion of the left epistyloid 47 48 (Fig. 46, l.styl) is suggestive of an articulation point close to the 49 ventrolateral wall of the braincase. One of the notable features of the 50 51 braincase of Scelidosaurus is the presence of flattened and oblique 52 53 basioccipital tuberosities (Figs 10, 31, bot) and, immediately dorsal to 54 these, are enlarged opisthotic pedicles (Figs. 10, 31, op. ped). It seems 55 56 plausible to envisage these conjoined facets as potential articular/sutural 57 58 attachment sites for the epistyloid bones. This supposition needs to be 59 60
126 Page 127 of 215 Zoological Journal of the Linnean Society
1 2 3 confirmed by preparation of the original specimen, or non-invasive CT 4 5 scanning of this specimen. 6 7 Styloid bones per se are associated with the support of the tongue 8 9 and hyoid muscles in human anatomy and may be pathologically enlarged 10 in some rare individuals, but these bones are rarely reported among 11 12 vertebrates. In this particular instance the bones referred to as epistyloids 13 14 are unusually long well ossified bones that have a precise structure. What 15 their role or function may have been is at present uncertain. They lie 16 17 dorsolateral to the hyoid apparatus and their bladed distal ends flank the 18 19 anterior pharynx-neck region adjacent to the area of the neck where the 20 cervical ribs are short and deflected posteriorly (Norman [Part 2]). It is of 21 For Review Only 22 course reasonable to argue that the epistyloids referred to here are a part 23 24 of the hyoid apparatus. However, the orientation of these very long bones 25 and their positioning relative to the 2nd ceratobranchial (found attached to 26 27 the posterior corner of the left mandible of the same individual) render 28 29 such an argument unlikely. 30 31 The nearest comparison that might be made is to the epibranchials 32 33 reconstructed as part of the hyoid apparatus in the ankylosaurid 34 Pinacosaurus by Hill et al. (2015: fig. 4). However, these latter are very 35 36 small rod-shaped bones described as having a compressed and spatulate 37 38 anterior end and tapering to a narrow cylindrical tip posteriorly. These 39 latter bones were somewhat disarticulated in the original specimen, so it 40 41 is possible that their orientation as described is a post-mortem artefact. 42 43 However, these are also small bones positioned (and reconstructed) 44 posterolateral to the main body of the hyobranchial apparatus. In marked 45 46 contrast, in Scelidosaurus the proximal ends of the epistyloid bones 47 48 occupy a dorsal position, ventrolateral to the braincase, and project 49 posteroventrolaterally into the anterolateral precincts of the neck before 50 51 terminating in spatulate-bladed distal ends. Hill et al. (2015) suggested 52 53 that there were a number of examples of misinterpreted parts of the hyoid 54 apparatus in dinosaurs. It is perhaps worth noting that two particularly 55 56 well preserved articulated theropods Scipionyx samniticus (Dall Sasso & 57 58 Signore, 1998), and Syntarsus kayentakae (Tykoski, 1998), also preserve 59 very elongate rod-like structures positioned posteroventral to their skulls. 60
127 Zoological Journal of the Linnean Society Page 128 of 215
1 2 3 These bones were identified as ‘hyoids’. There is a remote possibility that 4 5 these bones may be more similar to the bones referred to as epibranchials 6 7 by Hill et al. (2015) or, perhaps, the epistyloids of Scelidosaurus. 8 9 The orientation of the semi-articulated epistyloids in the referred 10 skeleton (BRSMG LEGL 0004) results in their being shafts being directed 11 12 toward the pectoral girdle, roughly in line with that of the cucullaris 13 14 cervicis musculature of extant diapsids (Theis 2010) – perhaps more 15 familiar as the sternocleidomastoids of mammalian anatomy. This portion 16 17 of the cucullaris muscle complex takes its origin on the clavicle-scapula 18 19 near the midline and extends diagonally forward toward the base of the 20 braincase; this musculature is involved in head positioning and neck 21 For Review Only 22 flexure. The bladed posterior ends of the epistyloid bones may be 23 24 interpreted as offering a site for insertion of cucullaris cervicis muscles. 25 Neck musculature with this orientation would have been capable of flexing 26 27 the head on the neck (against the tension created by the ligamentous 28 29 connections between the osteoderm arrays in the nuchal region dorsally – 30 Norman [Part 3]) for (re)positioning the head while feeding on browse. 31 32 Equally, the ability to flex and rotate the head would have allowed the 33 34 animal to expose its occipital horns more prominently, which may have 35 36 been linked to postural signaling and (unknowable) aspects of the 37 behaviour of these dinosaurs. 38 39 40 41 42 Pneumatic system remnants? 43 44 The undoubted juvenile specimen (CAMSM X39256: Fig. 29C) displays a 45 deep pit medial to the quadrate foramen. This structure resembles a pit 46 47 seen in the same position on the quadrate of Heterodontosaurus (Norman 48 49 et al. 2011) and interpreted as a remnant of cranial pneumatism in the 50 latter taxon. This feature corresponds positionally to the posterior 51 52 pneumatic foramen found in Recent avians and avian theropods, as 53 54 described by Hendrickx et al. (2015). It could be argued that this pit 55 simply represents a vascular foramen for providing nutrients necessary for 56 57 growth of the quadrate in a juvenile. Indeed several bones of small 58 59 (immature) individuals of Scelidosaurus show isolated foramina (see Fig. 60
128 Page 129 of 215 Zoological Journal of the Linnean Society
1 2 3 22) that are not seen in the equivalent bones of larger and more mature 4 5 specimens. However, these latter foramina are surface features, rather 6 7 than ones associated with a pronounced surrounding fossa (as often seen 8 in examples of pneumatic foramina). 9 10 11 12 13 Exostoses (accretionary bone) and osteoderms 14 15 More derived ankylosaurs (nodosaurids and ankylosaurids) exhibit 16 17 substantial cranial ornamentation (including osteoderms) that mask the 18 endochondral skeleton; this degree of masking is considerably greater in 19 20 extent than anything seen in Scelidosaurus. In ankylosaurs the dominant 21 For Review Only 22 cranial covering is in the form of bony osteoderms that adhere firmly to 23 the external skull bones and often form what can be considered to be 24 25 systematically distinctive mosaic patterns (e.g. Coombs 1978, Arbour & 26 27 Currie 2016, Arbour & Evans 2017). Cranial osteodermal plating can be so 28 pervasive that it may occlude the temporal fenestrae, form bony eyelids, 29 30 cover-plates (cheek plates) lateral to the buccal emargination and also 31 32 become elaborated into conical horns that project from the posterior 33 margin and flanks of their skulls. 34 35 Vickaryous, Russell & Currie (2001) investigated the developmental 36 37 origin of this extreme form of cranial armour-plating using a combination 38 39 of evidence from some immature ankylosaurid specimens of 40 Euoplocephalus and Pinacosaurus, and some examples of extant lizards 41 42 that develop analogous patterns of cranial bony plating. Their approach 43 44 was an attempt to discriminate between two generalised models of the 45 development of ankylosaur cranial ornament that had been discussed in 46 47 prior literature (Coombs 1971; Coombs & Maryańska 1990): accretion 48 49 (growth by elaboration of bone on the surface of the underlying dermal 50 skull bones), as opposed to fusion (of osteodermal plates that formed 51 52 independently within the overlying dermis to the underlying dermal skull 53 54 roof). Their conclusion was that there were osteological correlates, based 55 on the growth and development of cranial armour in extant lizards, 56 57 indicating that accretionary bone growth and fusion of dermally derived 58 59 osteoderms occurred in the skulls of ankylosaurs. 60
129 Zoological Journal of the Linnean Society Page 130 of 215
1 2 3 Scelidosaurus exhibits exostoses (periosteal accretionary growth of 4 5 bone) on the external surface of its skull and mandible. It is probable that 6 7 this rugose superficial bone acted as a base for the attachment and 8 growth of an epidermal shields formed by keratinous scales. The 9 10 development and fusion of true dermal osteoderms (caputegulae) to the 11 12 surface of endochondral skull bones is considered to be a largely derived 13 ankylosaur condition. This may have been presaged, in Scelidosaurus, by 14 15 the elaboration of the supraorbital bones (regarded as osteodermal in 16 17 origin), an osteoderm on the lateral surface of the postorbital and the 18 osteoderm horns fused to the posterodorsal surface of the occipital plate. 19 20 21 For Review Only 22 23 Occipital horns 24 25 The occipital horns are constructed from two discrete components: a 26 27 thick, disc-like base-plate (ba) and, fused to the base-plate, a more 28 superficially placed sub-conical osteoderm (see Norman [Part 3]). The 29 30 base-plate osteoderm has a very coarse and porous texture. In contrast, 31 32 the surface texture of the horn-shaped osteoderm is far smoother, 33 although still porous, being shallowly excavated by many small foramina 34 35 and shallow grooves suggestive of an anastomosing network of blood 36 37 vessels that supplied nutrients for the growth of keratinous tissue that 38 coated its surface. 39 40 Similar bony horn-core textures are seen, for example, in extant 41 42 bovids (pers. colln: Ovis aries). More recently, the remarkable 43 44 preservation of epidermal tissue on the exoskeleton of the ankylosaurs 45 Borealopelta (Brown et al. 2017) and Zuul (Arbour & Evans 2017) 46 47 confirms the existence of keratinous sheaths capping some of their 48 49 osteoderms (as suggested for Scelidosaurus by Martill et al. 2000). 50 51 It is considered most probable that the exostoses and genuine 52 53 osteoderms on the skull, and some of the cap-shaped osteoderms 54 distributed across the torso and appendicular skeleton of Scelidosaurus, 55 56 were sheathed by keratin. A more detailed description of the dermal 57 58 skeleton of Scelidosaurus is presented elsewhere (Norman [Part 3]). 59 60
130 Page 131 of 215 Zoological Journal of the Linnean Society
1 2 3 CONCLUSIONS (PART 1) 4 5 The cranial anatomy of Scelidosaurus harrisonii has been re-described 6 7 using the acid-prepared remains of the lectotype skull (NHMUK R1111), 8 9 and supplementary information provided by three important referred 10 specimens (BRSMG Ce12785, BRSMG LEGL 0004, and CAMSM X39256). 11 12 13 The surface of the chondrocranium is coated with periosteally 14 derived bony tissues (exostoses) with a granular or fibrous texture. These 15 16 textured surfaces are likely to have anchored epidermal keratinous scales 17 18 akin to those seen plating the surface of the skull in many extant 19 testudines. A mound-like exostosis is also centred on the angular bone of 20 21 the mandible and Forbony strands Review radiate from Only the centre of the angular and 22 23 encroach upon the surangular and dentary. There is evidence of a 24 comparatively shallow, crater-shaped osteoderm that is fused to the 25 26 lateral surface of the postorbital. There is a prominent pair of large horn- 27 28 shaped keratinous osteoderms anchored to the occiput by block-shaped 29 base-plates. The tip of the snout is enveloped by a small keratinous 30 31 rhamphotheca that grew from a discrete, rugose pad of exostotic bone on 32 33 the tip of the premaxillae. The anterior keratinous premaxillary beak was, 34 by inference, complemented by an equivalent structure on the tip of the 35 36 lower jaw. A small crescentic predentary (membranous or cartilaginous, 37 38 and perhaps bony in mature specimens) can be inferred to have capped 39 the symphyseal area of the mandibles and provided an attachment and 40 41 growth site for the mandibular beak. 42 43 Each premaxilla is narrow and supports five marginal teeth (rather 44 45 than six, as previously reported). The anterior premaxillary teeth (P1-P3) 46 47 have crowns that have bulbous bases that support carinate, slightly 48 recurved cones (the carinae bear small rounded denticles in some 49 50 instances); the last two teeth P4 & P5 transition structurally to the more 51 52 symmetrical, trapezoidal, crenulate morphology seen in the maxillary 53 dentition. The maxilla of comparatively mature individuals would have had 54 55 at least 22 tooth positions and the dentary at least 27 positions (by 56 57 inference from the number of teeth and alveoli in presently known 58 specimens). The crowns of teeth in both dentitions are tightly clustered, 59 60 so that functional crowns are arranged en echelon. Both dentitions are
131 Zoological Journal of the Linnean Society Page 132 of 215
1 2 3 markedly bowed lingually and have a gently sinuous anteroposterior 4 5 profile; this arrangement acted as a passive guidance system that 6 7 ensured jaw closure was strictly orthogonal. Tooth abrasion, although 8 somewhat discontinuous along the dentition produces some clusters of 9 10 well-defined wear facets that are high-angle (nearly vertical). The tooth 11 12 eruption/replacement pattern exhibited by one individual created a 13 dentition of uniformly erupted teeth with little evidence of abrasion. The 14 15 development of this latter dentition defies explanation at present. 16 17 The premaxillae and nasals bordered a comparatively small, fully 18 19 enclosed, external naris. The nasal cavities run posteriorly and form 20 simple, slightly arched tunnels subdivided anteriorly in the midline by a 21 For Review Only 22 deep, narrow, fused vomerine plate; the dorsal edges of the vomers 23 24 appear to have anchored thin, slightly-arched bones, referred to as 25 epivomers. There is a narrow interpterygoid vacuity between the bones of 26 27 the palate posterior to the vomers. A vertically positioned epipterygoid 28 29 with a laterally flattened base and conical apex is attached (probably 30 sutured) to the dorsolateral edge of the quadrate wing of the pterygoid; 31 32 the osseus apical portion of this bone is pointed and does not appear to 33 34 have articulated against the sidewall of the braincase. The ossified portion 35 36 of the stapes (columella auris) is described and illustrated. The anterior 37 ventrolateral portion of the basioccipital is expanded, creating a pair of 38 39 oblique rugose facets that are bounded dorsally by pedicles derived from 40 41 the opisthotics. Elongate blade-ended bones, referred to herein as 42 epistyloids, have been found in semi-articulation at the back of the skull of 43 44 the referred skeleton (BRSMG LEGL 0004) and as a single, incomplete, 45 46 isolated element in the smaller referred skeleton (CAMSM X39256). It is 47 considered possible that these bones articulated against the prominent 48 49 facets on the basioccipital and were attachment sites for the cucullaris 50 51 cervicis musculature that was anchored to the pectoral arch (clavicle- 52 scapula) was involved in head-neck flexure (broadly equivalent to the 53 54 sternocleidomastoid of mammals). These structures, and their associated 55 56 musculature (if correctly interpreted), may have been involved not only in 57 head repositioning while feeding on low browse, but had the potential to 58 59 60
132 Page 133 of 215 Zoological Journal of the Linnean Society
1 2 3 be involved in posturing and display of the occipital horns as part of the 4 5 behavioural repertoire of these animals. 6 7 Sclerotic ossicles supported the sclera in the eyeball, and a set of 8 9 three (dermally derived) supraorbital bones border the dorsolateral 10 margin of the orbit, creating a prominent brow ridge. The latter may have 11 12 had a protective function and may also have been associated with some 13 14 unknown behavioural aspect of this animal’s way of life. 15 16 17 18 19 20 21 AcknowledgementsFor. It was Review Alan Jack Charig’s Only (1927-1997) dying wish 22 that I finish the work on Scelidosaurus with which he had been involved. 23 24 This study would not have been possible without the assistance of Angela 25 26 Milner and Paul Barrett (former and subsequent curators of the relevant 27 NHM collections) and latterly Susannah Maidment (curator of the dinosaur 28 29 collections) as well as Sandra Chapman and Lorna Steel (assistant 30 31 curators). They all granted access to the collections and provided facilities 32 necessary for the examination of the fragile lectotype skull and other 33 34 historically important remains of Scelidosaurus in the fossil collections of 35 36 the Natural History Museum, London. I am also grateful to the staff of the 37 Lyme Regis Museum (formerly the Philpot Museum) for access to the 38 39 original paratype material of Scelidosaurus that is in their care. David Sole 40 41 (Lyme Regis) has discovered, collected and assembled very important 42 scelidosaur material from the cliffs and foreshore at Charmouth over the 43 44 past two decades and encouraged and facilitated my work on these 45 46 specimens by donating these important specimens to the Bristol City 47 Museum. He also kindly assisted my work by providing background 48 49 information concerning these discoveries and their collection. At Bristol 50 51 City Museum and Art Gallery, the late Roger Vaughn assisted in a first 52 phase of preparation and study of the intermediate-sized skeleton 53 54 (BRSMG Ce12785). Somewhat later Deborah Hutchinson (Curator) and 55 56 David Singleton (Conservator) at Bristol City Museum provided 57 considerable support and assistance that facilitated the removal from 58 59 exhibition for study of three partial skeletons of Scelidosaurus. At the 60
133 Zoological Journal of the Linnean Society Page 134 of 215
1 2 3 Sedgwick Museum, Cambridge, Daniel Pemberton (Collections Manager) 4 5 and Matthew Riley (Palaeontological Assistant Curator) facilitated the 6 7 study of the partial skeleton of Scelidosaurus (CAMSM X39256). Dudley 8 Simons (Department of Earth Sciences, Cambridge) assisted with the 9 10 preparation of many photographic images of scelidosaur bones. Matt 11 12 Carrano (Smithsonian Institution, Washington DC) provided access and 13 study facilities to allow study of some Stegosaurus material in their 14 15 collections and checked the anatomy of some of Gilmore’s exhibited skull 16 17 material. 18 Two largely unsung heroes: Ron Croucher (formerly superintendent 19 20 of the fossil preparation laboratory at the Natural History Museum) and 21 For Review Only 22 the late David Costin (Lyme Regis) devoted decades of their lives to the 23 preparation of several partial skeletons of Scelidosaurus – without their 24 25 diligent work this monograph would not have been possible. 26 27 I must applaud the generosity of Kevin Padian (University of 28 California, Berkeley) and David Weishampel (Johns Hopkins University, 29 30 Baltimore) for undertaking the largely thankless task of reading and 31 32 commenting upon a pre-submission version of this monograph. Their 33 wisdom and incisive commentaries improved the quality of the work 34 35 presented here. I am also grateful to anonymous reviewers (and Atilla Ösi 36 37 – Hungarian Academy of Sciences, Budapest) working on behalf of the 38 Linnean Society, for the time and effort expended on their reviews, as well 39 40 as Nick Fraser (the Associate Editor – National Museums of Scotland, 41 42 Edinburgh) for his forbearance. All the comments and critical observations 43 have improved this account of the skull of this fascinating and yet 44 45 enigmatic animal. Any persistent errors and misinterpretations are the 46 47 fault of the author alone. 48 This work was made possible through the award of a Visiting 49 50 Scholarship awarded by St John’s College Oxford, as well as grants from 51 52 the Prism Fund administered by the Science Museum London and Trinity 53 College Cambridge. An Asher Tunis Distinguished Research Fellowship 54 55 awarded by the Smithsonian Institution aided preliminary work on this 56 57 long-term research project, as did a year of study leave (during the 58 academic year 2017-18) that was granted by the University of Cambridge 59 60 and the Governing Body of Christ’s College Cambridge.
134 Page 135 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 For Review Only 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
135 Zoological Journal of the Linnean Society Page 136 of 215
1 2 3 FIGURE LEGENDS 4 5 6 7 Figure 1. The coastline and adjacent topography of the Charmouth area of 8 Church Bay (Lyme Bay). The unstable and degraded cliff area known as 9 10 The Spittles-Black Ven, situated between Charmouth and Lyme Regis, has 11 12 been subjected to intermittent minor rotational slips and landslides 13 (Brunsden 2002, Gallois 2008). Scelidosaur remains are found either 14 15 weathering out in broken marlstone nodules on the cliff face, in the scree, 16 17 or as eroded water-rolled nodules on the foreshore beneath the cliff. 18 Dark speckled areas: foreshore. Breaks of slope: multiple landslips on The 19 20 Spittles-Black Ven, as well as on the cliffs of Cain’s Folly directly beneath 21 For Review Only 22 Stonebarrow. 23 24 25 Figure 2. A. Theropod femur (NHMUK OR109560). Part of the hypodigm of 26 27 Scelidosaurus harrisonii Owen, 1861 (paratype following Lydekker 1888). 28 An isolated, incomplete proximal portion of a left femur belonging to an 29 30 indeterminate theropod dinosaur that was collected near Charmouth by 31 32 James Harrison and illustrated in Owen (1861: Tab. I). Shown here in 33 anterolateral view, posteromedial and posterior views. B. Theropod knee 34 35 joint (NHMUK OR39496) and an isolated ungual phalanx (NHMUK 36 37 OR109561). Part of the hypodigm of Scelidosaurus harrisonii Owen (1861: 38 Tab. II). The knee joint was arbitrarily designated the “type” of the 39 40 species by Lydekker (1888). Medial view of the distal end of a right femur 41 42 articulated (by matrix) to the proximal tibia (fibula not visible) belonging 43 to another indeterminate theropod dinosaur. The two smaller illustrations 44 45 are views of a partly eroded theropod ungual phalanx. Scale bars = 46 47 centimetres. 48 49 50 Figure 3. Scelidosaur bones (LYMPH 1998 6.1-6.7). Part of the hypodigm 51 52 of Scelidosaurus harrisonii Owen, 1861 (a paratype, following Lydekker 53 54 1888). An assortment of views of hindlimb bones and vertebrae that are 55 commensurate and have their own distinctive preservational condition – 56 57 all the limb bones are crushed flat; these bones represent associated 58 59 parts of a juvenile scelidosaur (Owen (1861: Tab. III). 60
136 Page 137 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 Figure 4. Scelidosaurus harrisonii. Part of the hypodigm of Scelidosaurus 6 7 harrisonii Owen, 1861 (paratype following Lydekker 1888); this skull and 8 its associated skeleton (NHMUK R1111) was later designated the 9 10 lectotype. The articulated skull and lower jaws embedded in marlstone 11 12 matrix as illustrated by Owen (1861: Tabs. IV-VI). Note the tip of the 13 snout is missing as are the upper parts of the orbital margin, temporal 14 15 arcade and occiput. This damage might have occurred during quarrying 16 17 but seems more likely to have occurred after a period of exposure and 18 water-rolling of the skull-containing marlstone nodule on the foreshore. A. 19 20 Right lateral view. B. Left lateral view. C. Partial internal view of the right 21 For Review Only 22 mandible and palate. D. Dorsal aspect of the skull. It is evident that the 23 rugose texturing on the surface of the skull bones was visible in its 24 25 original state of preservation. Scale bar = centimetres. 26 27 28 Figure 5. Scelidosaurus harrisonii (NHMUK R1111). The assembled 29 30 skeleton of Scelidosaurus as exhibited for many years in its own glass 31 32 enclosure in one of the galleries at the Natural History. This sketch was 33 created from an image on an old photographic postcard dating from the 34 35 1950s. Anatomical note: the skull and shoulder nodules are unnaturally 36 37 close together, which reflects the fact that the majority of the cervical 38 region was never recovered. Length of mount: roughly 4 metres. 39 40 41 42 Figure 6. Hylaeosaurus. Sketch of the supposed life-sized iron-framed, 43 44 concrete and tiled restoration (1854) of a Wealden armoured dinosaur, 45 created for the opening of Crystal Palace Park in South London. 46 47 Approximately 7 metres long. Illustration prepared by Sharon Capon 48 49 (Department of Earth Sciences, Cambridge). 50 51 52 Figure 7. Stratigraphic column focused on the exposures seen on The 53 54 Spittles-Black Ven, Charmouth. Based, in part, on an original stratigraphic 55 log recorded by Hesselbo & Jenkyns (1995). The bed naming and 56 57 numbering system was devised by Lang (1924) and was informed by the 58 59 nomenclature adopted by local quarrymen. The Obtusum zone log 60
137 Zoological Journal of the Linnean Society Page 138 of 215
1 2 3 illustrated here comprises a basal sequence of carbonate-enriched 4 5 mudrock with occasional benches of purer marly limestone; this is 6 7 succeeded by an organic-rich set of shales with rare lenticular marlstones. 8 Finally, the upper part of the sequence reverts to the carbonate-rich 9 10 mudrock with thin, occasionally lenticular, marly limestone beds. The beds 11 12 that have been reported to have yielded scelidosaur remains are 83 – 13 Stonebarrow Flatstones (Ian West - 14 15 http://www.southampton.ac.uk/~imw/Lyme-Regis-to-Charmouth.htm) 16 17 and 85 – Topstones (David Sole, pers. comm. 2018). 18 19 20 Sedimentary log conventions: pale even tone – carbonate-rich mudstones, 21 For Review Only 22 fine horizontal shading – shale/mudstones, blocky units and lenticles: 23 marlstone/limestone. Scale bar in metres. 24 25 26 27 Figure 8. Scelidosaurus harrisonii Owen. Skull and lower jaw 28 reconstructed in lateral view (excluding occipital osteoderms and 29 30 epistyloid bones). Based upon the lectotype (NHMUK R1111), and the 31 32 referred specimens (BRSMG Ce12785 and BRSMG LEGL 0004). 33 34 35 Abbreviations: aaf – anterior antorbital foramen, af – antorbital fenestra, 36 37 amf – anterior maxillary foramen, An – Angular, aof – antorbital fossa, apf 38 39 – anterior premaxillary foramen, Ar – Articular, be – buccal emargination 40 (rim of), Co – Coronoid, D – Dentary, do – (postorbital) osteoderm, Ec – 41 42 Ectopterygoid, ed – edentulous tip of premaxilla, Ju – Jugal, La – 43 44 Lacrimal, Mx – Maxilla, N – Nasal, nf – external narial fenestra, Pd? – 45 predentary cartilage or bone (implied), P – Parietal, Pf – Prefrontal, Pm – 46 47 Premaxilla, p.os – periosteal ossifications (exostoses), Po – Postorbital, Pp 48 49 – palpebral bone, poc – paroccipital process, pr – parasphenoid rostrum 50 (cultriform process), Pro – Proötic, pso – posterior supraorbital, Pt – 51 52 Pterygoid, Q – Quadrate, qf – quadrate (paraquadratic) foramen, Qj – 53 54 Quadratojugal, Sa – Surangular, sac – surangular condyle (of the 55 quadrate), saf – surangular foramen, Sq – Squamosal. 56 57 58 59 60
138 Page 139 of 215 Zoological Journal of the Linnean Society
1 2 3 Figure 9. Scelidosaurus harrisonii Owen. Skull reconstructed in dorsal 4 5 aspect based upon the lectotype (NHMUK R1111), and the referred 6 7 specimen (BRSMG LEGL 0004). 8 9 10 Abbreviations: apf – anterior premaxillary foramen, Bo – Basioccipital, Ex 11 12 – Exoccipital, F – Frontal, gap – unossified slot, isl – islets of bone 13 spreading across the surface of the postorbital, Ls – Laterosphenoid, mso 14 15 – middle supraorbital bone, Mx – Maxilla, oc.f – facet for attachment of 16 17 the occipital horn base-plate, Op – Opisthotic, P – Parietal, Pf – Prefrontal, 18 Pm – Premaxilla, Po – Postorbital, Pp – palpebral bone, poc – paroccipital 19 20 process, pso – posterior supraorbital, Pt – Pterygoid, So – Supraoccipital, 21 For Review Only 22 Sq – Squamosal, vcd – fenestra for the vena capitis dorsalis. 23 24 25 Figure 10. Scelidosaurus harrisonii Owen. Skull and lower jaw 26 27 reconstructed in ventral view based upon the lectotype (NHMUK R1111). 28 29 30 Abbreviations: ac – articular condyle (of the quadrate), apf – anterior 31 32 premaxillary foramen, Ar – Articular, be – buccal emargination (rim of), 33 34 Bo – Basioccipital, bot – basioccipital tuberosity, Bs – Basisphenoid, Ec – 35 Ectopterygoid, epv – epivomer, Ju – Jugal, La – Lacrimal, mwp – maxillary 36 37 wall of the palatine, Mx – Maxilla, op.ped – opisthotic pedicle, p.os – 38 39 periosteal ossification, Pal – Palatine, pf – palatine foramen, Pm – 40 Premaxilla, Po – Postorbital, Pp – palpebral bone, poc – paroccipital 41 42 process, pso – posterior supraorbital, Pt – Pterygoid, Q – Quadrate, qf – 43 44 quadrate (paraquadratic) foramen, Qj – Quadratojugal, sac – surangular 45 condyle (of the quadrate), sf – special foramina, Sq – Squamosal, st – 46 47 stapes, V – Vomer. 48 49 50 Figure 11. Scelidosaurus harrisonii Owen. Skull and lower jaw 51 52 reconstructed in occipital view based upon the lectotype (NHMUK R1111), 53 54 and the referred specimen (CAMSM X39256). 55 56 57 Abbreviations: ac – articular condyle (of the quadrate), Bo – Basioccipital, 58 59 bot – basioccipital tuberosity, bpt – basipterygoid process, Bs – 60
139 Zoological Journal of the Linnean Society Page 140 of 215
1 2 3 Basisphenoid, dep – depression in the medial wall of the pterygoid, Ex – 4 5 Exoccipital, fm – foramen magnum, Ju – Jugal, oc.f - facet for attachment 6 7 of the occipital horn base-plate, Op – Opisthotic, op.ped – opisthotic 8 peduncle, P – Parietal, poc – paroccipital process, Pt – Pterygoid, ptf – 9 10 post-temporal fenestra/foramen, qf – quadrate (paraquadratic) foramen, 11 12 Qj – Quadratojugal, sac – surangular condyle (of the quadrate), sl – 13 narrow, pocket-like slot on the medial wall of the pterygoid, sg – sagittal 14 15 groove on the parietals, So – Supraoccipital, st – stapes. 16 17 18 Figure 12. Scelidosaurus harrisonii Owen, 1861. Premaxilla and articulated 19 20 lacrimal of the lectotype (NHMUK R1111). A. Left maxilla and lacrimal in 21 For Review Only 22 lateral view. B. Left maxilla, palatine and lacrimal in medial view; lacrimal 23 shown in outline and semi-transparent in articulation with the lacrimal 24 25 suture. C. Lacrimal in medial view. 26 27 28 Abbreviations: aaf – anterior antorbital foramen, af – antorbital fenestra, 29 30 aof - antorbital fossa, ap – alveolar parapet, be – buccal emargination 31 32 (boundary of), br – broken surfaces, ecs – ectopterygoid suture, es – 33 eroded surface, jp – jugal process, js – jugal suture, La – Lacrimal, ld – 34 35 lacrimal duct (opening for), ls – lacrimal suture, mw – medial wall of 36 37 lacrimal, mwp – maxillary wall of the palatine, Mx – Maxilla, mxs – 38 maxillary shelf, p.os – periosteal ossification, Pal – Palatine, Pm – 39 40 Premaxilla, pp? – fragment of the palpebral, sf – special foramina. 41 42 Even tone: broken/eroded surfaces. 43 44 45 Figure 13. Scelidosaurus cf. harrisonii Owen, 1861. Facial skeleton of the 46 47 referred specimen (BRSMG Ce12785). A. Photographic image of the right 48 49 lateral view of the partial, laterally compressed skull. B. Interpretative 50 drawing of the same image. C. Close-up of the premaxillae in right 51 52 ventrolateral aspect (see Fig. 41A). D. Premaxilla (left) in lateral aspect, 53 54 the dorsal process is sheared off at its base. 55 56 57 Abbreviations: aaf – anterior antorbital foramen, af – antorbital fenestra, 58 59 aof - antorbital fossa, apo – palatal opening of the anterior premaxillary 60
140 Page 141 of 215 Zoological Journal of the Linnean Society
1 2 3 foramen, dp – dorsomedial process of premaxilla, gr – grooves, Ju – 4 5 Jugal, La – Lacrimal, nf – external narial fenestra, Pm – Premaxilla, pms – 6 7 premaxillary suture, p.os – periosteal ossification, pos – postorbital 8 suture, pp – posterior premaxillary process, sf – special foramina. 9 10 Even tone: matrix and broken surfaces. 11 12 13 Figure 14. Scelidosaurus harrisonii Owen, 1861. Facial skeleton of the 14 15 referred specimen (BRSMG LEGL 0004) in left lateral view. Photographic 16 17 image of the original and an interpretative sketch of the same. 18 19 20 Abbreviations: af – dental abrasion facets, amf – anterior maxillary 21 For Review Only 22 foramen, An – Angular, aof – antorbital fossa, br – breakage, cb2 – 2nd 23 ceratobranchial, do – postorbital osteoderm, Ju – jugal, La – Lacrimal, lN 24 25 – left Nasal, p.os – periosteal ossification, Pf – Prefrontal, Pm – 26 27 Premaxilla, Po – Postorbital, Pp – palpebral, pso – posterior supraorbital, 28 Qj – Quadratojugal, rN – right Nasal, Sa – Surangular. 29 30 Even tone: matrix/broken surfaces. 31 32 33 34 Figure 15. Scelidosaurus cf. harrisonii Owen, 1861. The isolated nasals of 35 the referred specimen (BRSMG Ce12785). A. Left lateral aspect. B. Right 36 37 lateral aspect. 38 39 40 Abbreviations: mf – medial flange, nf – external narial fenestra, p.os – 41 42 periosteal ossifications, pfs – prefrontal suture, pms – premaxillary suture 43 44 (for the median dorsal process of the premaxilla). 45 Even tone: matrix. 46 47 48 49 Figure 16. Scelidosaurus cf. harrisonii Owen, 1861. The skull roof of the 50 referred specimen (BRSMG LEGL 0004) in dorsal aspect. Photographic 51 52 image of the original specimen and interpretative sketch of the same. 53 54 55 Abbreviations: aof – antorbital fossa, ba – base-plate osteoderm, cro – 56 57 cervical ring epiosteoderm, F – Frontal, gap – unossified space, lN – left 58 59 Nasal, l.oc.ost – left occipital epiosteoderm (horn), lPf – left Prefrontal, Ls- 60
141 Zoological Journal of the Linnean Society Page 142 of 215
1 2 3 Pr.s – Laterosphenoid-Proötic suture surface, mso – middle supraorbital 4 5 bone, P – Parietal, Po – Postorbital, Pp – palpebral (anterior supraorbital), 6 7 pso – posterior supraorbital, rN – right Nasal, r.oc.ost – right occipital 8 epiosteoderm (horn), rPf – right Prefrontal, Sq – Squamosal, vcd – 9 10 foramen for the vena capitis dorsalis. 11 12 Even tone: matrix. 13 14 15 Figure 17. Scelidosaurus cf. harrisonii Owen, 1861. Orbital elements 16 17 assembled in the referred specimen (BRSMG Ce12785). A. Lateral view of 18 the right postorbital, the posterior supraorbital, middle supraorbital and 19 20 prefrontal. B. Postorbital, posterior supraorbital and middle supraorbital in 21 For Review Only 22 articulation – note the pitting/foramina on the dorsal surface of the middle 23 supraorbital, the entire bone has a spongy texture. C. Postorbital and 24 25 posterior supraorbital in anterior view. D. Middle supraorbital in ventral 26 27 view. E. Prefrontal in medial view. 28 29 30 Abbreviations: do – postorbital osteoderm, fs – frontal suture, ls – 31 32 lacrimal suture, mso – middle supraorbital, ns – nasal suture, os – orbital 33 roof, Pf – Prefrontal, Po – Postorbital, pos – postorbital suture, ppf – 34 35 palpebral facet, Pr – Prefrontal, prs – prefrontal suture, pso – posterior 36 37 supraorbital, sq.s – squamosal suture, vas – vascular openings. 38 39 40 Figure 18. Scelidosaurus cf. harrisonii Owen, 1861. The orbitotemporal 41 42 region of the skull of the referred specimen (BRSMG LEGL 0004). 43 44 45 Abbreviations: aof – antorbital fossa, do – postorbital osteoderm, j.s – 46 47 jugal suture, Ju – Jugal, La –Lacrimal, Mx – Maxilla, mx.s – maxillary 48 49 suture, N – Nasal, P – Parietal, p.os – periosteal ossification, Pf – 50 Prefrontal, Po –Postorbital, pop – paroccipital, Pp – palpebral, pso – 51 52 posterior supraorbital, Qj – quadratojugal, rug – rugose ledge, Sq – 53 54 Squamosal. 55 Even tone: highlighting the supraorbital elements. 56 57 58 59 60
142 Page 143 of 215 Zoological Journal of the Linnean Society
1 2 3 Figure 19. Scelidosaurus cf. harrisonii Owen, 1861. A complete right 4 5 postorbital belonging to the referred specimen (BRSMG Ce12785) with a 6 7 fused osteoderm. A. Lateral view. B. Anterior view. C. Medial view. D. 8 Dorsal view. 9 10 11 12 Abbreviations: bwl – bowl-shaped suture surface, do – postorbital 13 osteoderm, fs – frontal suture, hk – hook-shaped portion of the 14 15 osteoderm, jp – jugal process, j.s – jugal suture, sqp – squamosal 16 17 process, sqs – squamosal suture, vas – vascular foramina. 18 19 20 Figure 20. Stegosaurus sp. USNM 6645. Isolated left postorbital with a 21 For Review Only 22 posterior supraorbital sutured to its external surface. In: A. Left 23 dorsolateral view. B. Medial view. 24 25 26 27 Abbreviations: br – broken bone surface, fs – frontal suture surface, itf – 28 infratemporal fenestra, js – jugal suture, mso – sutural surfaces for the 29 30 middle supraorbital, Po – postorbital, pso – posterior supraorbital, stf – 31 32 supratemporal fenestra. 33 34 Even tone: highlighting the posterior supraorbital. 35 36 37 Figure 21. Scelidosaurus harrisonii Owen, 1861. The left jugal, and 38 39 associated bones, of the lectotype (NHMUK R1111). A. Left lateral view of 40 the jugal, quadratojugal and quadrate with fragments of the postorbital 41 42 and squamosal. B. Medial view of the jugal with fragments of the 43 44 quadratojugal and postorbital. C. Sketch anterior view of the jugal and 45 postorbital. 46 47 48 49 Abbreviations: br – breakage, do – postorbital osteoderm (eroded), ecs – 50 ectopterygoid suture, Ju –Jugal, ls – lacrimal suture, Mx – fragment of the 51 52 maxilla, p.os – periosteal ossifications, pal.s – palatine suture, Po – 53 54 Postorbital, ptw – pterygoid wing, Q – Quadrate, Qj – Quadratojugal, sc.o 55 – sclerotic ossicles, Sq – Squamosal, vas – vascular foramina. 56 57 Even tone: eroded surfaces. 58 59 60
143 Zoological Journal of the Linnean Society Page 144 of 215
1 2 3 Figure 22. Scelidosaurus harrisonii Owen, 1861. The right quadratojugal of 4 5 the lectotype (NHMUK R1111) in: A. Medial view. B. Lateral view. The 6 7 right quadratojugal of the referred specimen (BRSMG Ce12785) in: C. 8 Medial view. D. Lateral view. 9 10 11 12 Abbreviations: for – vascular foramina, j.s – jugal suture, qs – quadrate 13 suture, sqs – squamosal suture. 14 15 16 17 Figure 23. Scelidosaurus harrisonii Owen, 1861. The vomers and 18 associated palatal elements of the lectotype (NHMUK R1111). A. Vomers 19 20 in left lateral view, sketched in 1999. B. Vomers in left lateral view, 21 For Review Only 22 sketched in 2009 showing deterioration of these bones. C. Vomers in 23 dorsal view. D. Vomers in right lateral view. E. Vomers and left pterygoid 24 25 in articulation. 26 27 28 Abbreviations: bev – bevelled edge, bif – bifurcated end, br – broken 29 30 surface, Ep – Epipterygoid, lip – small notch and lip that secured the 31 32 premaxilla against the vomer, Pm – Premaxilla, pms – suture for the 33 34 premaxilla, Pt – Pterygoid, qs – sutural surface for the quadrate, V – 35 Vomer, vfr – fragment of vomer adhering to the pterygoid (the ‘notch’ 36 37 shows the place where the vfr fits). 38 39 40 Figure 24. Scelidosaurus cf. harrisonii Owen, 1861. The isolated snout 41 42 nodule of the referred specimen (BRSMG LEGL 0004) in right lateral view. 43 44 Photographic image of the original specimen and an interpretative sketch 45 of the same. 46 47 48 49 Abbreviations: Mx – maxilla, Pm – premaxilla, V – vomers. 50 Even tone (pale) – matrix. Even tone (dark) – tooth abrasion facet. 51 52 53 54 Figure 25. Scelidosaurus harrisonii Owen, 1861. An inked copy of rough 55 pencil sketches of the left and right palatal region of the lectotype 56 57 (NHMUK R1111). These were drawn by the author in the late 1970s, while 58 59 acid preparation of the lectotype skull was proceeding. 60
144 Page 145 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 Abbreviations: Bs – Basisphenoid, bpt – basipterygoid process, Ec – 6 7 Ectopterygoid, ?epv – possible epivomer remnants, Mx – Maxilla, Pal – 8 Palatine, pr – parasphenoid rostrum (cultriform process), Pt – Pterygoid, 9 10 st – Stapes, svm – crest for attachment of sub-vertebral musculature, V – 11 12 Vomer, vid – vidian canal. 13 14 15 Figure 26. Scelidosaurus harrisonii Owen, 1861. Two bones referred to 16 17 here as epivomers of the lectotype (NHMUK R1111). A,B. Dorsal and 18 ventral views. C,D. Dorsal and ventral views of the epivomers positioned 19 20 as ‘best-fit’ on the bevelled surface of each vomer. E. Anterior view of 21 For Review Only 22 vomers and epivomers in articulation. 23 Abbreviations: bev – bevelled edge, epv – Epivomer, nc – nasal cavity, 24 25 Pm – Premaxilla, sut – suture surface, V – Vomer. 26 27 28 Figure 27. Scelidosaurus harrisonii Owen, 1861. The palatines, 29 30 ectopterygoids and associated palatal elements of the lectotype (NHMUK 31 32 R1111). A,B. Right palatine in dorsal and ventral views. C,D. Left 33 34 ectopterygoid in dorsal and ventral views. E. Ectopterygoid and pterygoid 35 in ventral view. F. Left maxilla and palatine in dorsal view. G. Palatine and 36 37 ectopterygoid in ventral view and in articulation with the posterior maxilla. 38 39 40 Abbreviations: adf – adductor fenestra, aof – antorbital fossa, Ec – 41 42 ectopterygoid, ecs – ectopterygoid suture, fen – fenestra, js – jugal 43 44 suture, ls – lacrimal suture, mxu – maxillary suture, mwp – maxillary wall 45 of the palatine, mxf – maxillary fossa, mxs – maxillary shelf, Pal – 46 47 palatine, pal.s – palatine suture, pf – palatine fenestra, Pt – Pterygoid, 48 49 pt.s – pterygoid suture, sl – slot on medial wall of the pterygoid, sf – 50 special foramina 51 52 53 54 Figure 28. Scelidosaurus harrisonii Owen, 1861. The central portions of 55 the palatal complex of the lectotype (NHMUK R1111). A. Left lateral 56 57 reconstruction. B. Right medial reconstruction. 58 59 60
145 Zoological Journal of the Linnean Society Page 146 of 215
1 2 3 Abbreviations: ac – articular condyle, b.ar – basal articulation, bev – 4 5 bevelled edge, ecs – ectopterygoid suture, Ep – Epipterygoid, fl – 6 7 pterygoid flange, ms – medial shelf, pal.s – palatine suture, pil – pillar-like 8 structure on pterygoid, Q – Quadrate, qh – articular head of quadrate, Qj 9 10 – Quadratojugal, qs – quadrate suture, qw – quadrate wing, sl – narrow 11 12 slot-like pocket on the medial surface of the pterygoid, Sq – Squamosal, V 13 – Vomer, vfr – fragment of vomer, vp – vomerine process, vs – 14 15 intervomer suture, w – warped leading edge of the epipterygoid. 16 17 18 Figure 29. Scelidosaurus harrisonii Owen, 1861. The left quadrate of the 19 20 lectotype (NHMUK R1111) in A. Posterior view. B. Anterior view. C. The 21 For Review Only 22 left distal half of the quadrate of the referred specimen (CAMSM X39256) 23 I lateral view. 24 25 26 27 Abbreviations: ac – articular condyle, cb – condylar buttress, es – eroded 28 surface, jw – jugal wing, pn – pneumatic opening (remnant of?), ptw – 29 30 pterygoid wing, pt.s – pterygoid suture, qf – quadrate (paraquadratic) 31 32 foramen, qh- quadrate head (condylar surface), Qj – Quadratojugal, sac – 33 surangular condyle, Sq - Squamosal. 34 35 Even tone: eroded surface. 36 37 38 39 Figure 30. Scelidosaurus cf. harrisonii Owen, 1861. The right quadrate of 40 the referred specimen (BRSMG Ce12785). A. Lateral view. B. Posterior 41 42 view. C. Quadrate and quadratojugal in articulation. 43 44 45 Abbreviations: ac – articular condyle, cb – condylar buttress – eroded 46 47 surface, dep – depression in medial wall of pterygoid wing, js – jugal 48 49 suture, jw – jugal wing, ptw – pterygoid wing, pt.s – pterygoid suture, qf 50 – quadrate (paraquadratic) foramen, qh- quadrate head (condyle), Qj – 51 52 Quadratojugal, qjs – quadratojugal suture, sac – surangular condyle. 53 54 55 Figure 31. Scelidosaurus harrisonii Owen, 1861. A sketch reconstruction of 56 57 the neurocranium of the lectotype (NHMUK R1111). A. Left lateral view. B. 58 59 Ventral view. 60
146 Page 147 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 Abbreviations: au.rec – auditory (stapedial) channel, Bo – Basioccipital, 6 7 bot – basioccipital tuberosity, bpt – basipterygoid process, Bs – 8 Basisphenoid, Ex – Exoccipital, f.ov – fenestra ovalis, jug – foramen for 9 10 jugular vein, Ls – Laterosphenoid, n [roman] – cranial nerve number, oc – 11 12 occipital condyle, Op – Opisthotic, op.ped – opisthotic pedicle, Os – 13 Orbitosphenoid, ped – pedicle of the basisphenoid, pob – postorbital boss, 14 15 poc – paroccipital process, Pr – Proötic, pr – parasphenoid rostrum 16 17 (cultriform process), ps – parietal suture, psp – paroccipital spur, ptf – 18 post-temporal fenestra, So – Supraoccipital, svm – attachment ridges for 19 20 the sub-vertebral musculature, vid – Vidian canal. 21 For Review Only 22 23 Figure 32. Scelidosaurus cf. harrisonii Owen, 1861. Sketch of the 24 25 preserved remnant of the braincase of the referred specimen (CAMSM 26 27 X39256). A. Lateral view. B. Occipital (posterior) view. C. Oblique view. 28 29 30 Abbreviations: au.rec – auditory (stapedial) channel, Bo – Basioccipital, 31 32 bot – basioccipital tuberosity, can – canal between opisthotic and 33 34 supraoccipital, en.f – endocranial floor, Ex – Exoccipital, ex-op.s – 35 exoccipital-opisthotic suture, for – cranial nerve and vascular foramina, oc 36 37 – occipital condyle, op.ped – opisthotic pedicle, poc – paroccipital process, 38 39 so.s – supraoccipital suture, svm – crest for the attachment of sub- 40 vertebral muscles. 41 42 43 44 Figure 33. Scelidosaurus harrisonii Owen, 1861. The occiput of the 45 lectotype (NHMUK R1111) as preserved in posterodorsal view. 46 47 48 49 Abbreviations: Bo – Basioccipital, bot – basioccipital tuberosity, Ex – 50 Exoccipital, n.VI – foramina for abducens nerves, op.ped – opisthotic 51 52 pedicle, poc – paroccipital process, psp – paroccipital spur, ptf – post- 53 54 temporal fenestra (foramen), otc – otic capsule, sc – ligamentous 55 scarring, So – Supraoccipital, Sq – Squamosal, sr – sagittal ridge. 56 57 58 59 60
147 Zoological Journal of the Linnean Society Page 148 of 215
1 2 3 Figure 34. Scelidosaurus harrisonii Owen, 1861. The isolated left stapes of 4 5 the lectotype (NHMUK R1111). A,B. Two views. 6 7 8 Abbreviations: est – contact surface for the extrastapedial cartilage, ftp – 9 10 stapedial footplate. 11 12 13 Figure 35. Scelidosaurus harrisonii Owen, 1861. Isolated orbitosphenoid 14 15 and laterosphenoid of the lectotype (NHMUK R1111). A. Orbitosphenoid in 16 17 anteroventral view. B. Laterosphenoid in left lateral view. 18 19 20 Abbreviations: c,c* – contact points between orbitosphenoid and 21 For Review Only 22 laterosphenoid, fen – fenestra, for – foramen, Ls-Pr.s – laterosphenoid- 23 proötic suture, n.V – margin of trigeminal fossa, pob – postorbital boss, Ps 24 25 – parietal suture. 26 27 28 29 Figure 36. Scelidosaurus harrisonii Owen, 1861. The left mandible of the 30 lectotype (NHMUK R1111). A. Lateral view. B. Medial view. C. Dorsal view. 31 32 33 34 Abbreviations: An – Angular, ap – alveolar parapet, Ar – Articular, asf – 35 anterior surangular foramen, be – margin of the buccal emargination, Co 36 37 – Coronoid, D – Dentary, es – erosion surface, for – foramina, gl – 38 39 glenoid, imf – inferior mandibular (Meckelian) fenestra, Mg – Meckelian 40 groove, p.os – periosteal ossification, Pa – Prearticular, rp – retroarticular 41 42 process, rug – rugose bony strands, Sa – Surangular, saf – surangular 43 44 foramen, sf – special foramina, Sp – Splenial. 45 Even tone: broken/abraded surfaces. 46 47 48 49 Figure 37. Scelidosaurus harrisonii Owen, 1861. The right mandible of the 50 lectotype (NHMUK R1111). A. Lateral view. B. Medial view. C. Dorsal view. 51 52 53 54 Abbreviations: add.f – adductor fossa, An – Angular, ans – angular suture 55 surface, ap – alveolar parapet, Ar – Articular, asf – anterior surangular 56 57 foramen, be – margin of the buccal emargination, br – broken surface, Co 58 59 – Coronoid, D – Dentary, es – erosion surface, for – foramina, gl – glenoid 60
148 Page 149 of 215 Zoological Journal of the Linnean Society
1 2 3 (on articular and surangular), imf – inferior mandibular (Meckelian) 4 5 fenestra, Mg – Meckelian groove, p.os – periosteal ossification, Pa – 6 7 Prearticular, rp, retroarticular process, rug – bony strands, Sa – 8 Surangular, saf – surangular foramen, sf – special foramina, Sp – 9 10 Splenial. 11 12 Even tone: broken surfaces. 13 14 15 Figure 38. Scelidosaurus cf. harrisonii Owen, 1861. The posterior portion 16 17 of the left mandible of the referred specimen (CAMSM X39256) in medial 18 view (A); and the isolated surangular of the referred specimen (BRSMG 19 20 Ce12785) in lateral view (B). 21 For Review Only 22 23 Abbreviations: An – Angular, an.s – angular suture, Ar – Articular, br – 24 25 broken surface, gl – surangular portion of the mandibular glenoid, Pa – 26 27 Prearticular, pa.s – prearticular sutural surface undercutting the articular, 28 Sa – Surangular, saf – surangular foramen, rp – surangular portion of the 29 30 retroarticular process. 31 32 33 34 Figure 39. Scelidosaurus cf. harrisonii Owen, 1861. Lateral view of a 35 portion of the right side of the skull of the referred specimen (BRSMG 36 37 LEGL 0004). 38 39 40 Abbreviations: af – abrasion facet, An – Angular, br – breakage, D – 41 42 Dentary, for – foramina (neurovascular), js – jugal suture, Mx – Maxilla, 43 44 p.os – periosteal ossification, Pm – Premaxilla, Sa – Surangular, saf – 45 surangular foramen. 46 47 48 49 Figure 40. Scelidosaurus cf. harrisonii Owen, 1861. The anterior portion of 50 the left dentary of the referred specimen (BRSMG Ce12785). A. Lateral 51 52 view. B. Medial view. C. Dorsal view. 53 54 55 Abbreviations: alv – alveolar bone, ap – alveolar parapet, br – broken 56 57 surface, df – anterior dentary foramen, Mg – Meckelian groove, rt – roots 58 59 60
149 Zoological Journal of the Linnean Society Page 150 of 215
1 2 3 of teeth, rug – rugose surface, sf – special foramina, sh – shelf-like 4 5 projection, sut – sutural surface, sym – dentary symphysis. 6 7 8 Figure 41A. Scelidosaurus cf. harrisonii Owen, 1861. The premaxillary 9 10 dentition of the referred specimen (BRSMG Ce12785). Lateral view of the 11 12 left premaxillary dentition with a photographic image of the original and 13 an interpretative sketch of the same. 14 15 16 17 Abbreviations: apo – anterior premaxillary opening, M [numbered] 18 maxillary tooth number, P [numbered] – premaxillary tooth number, pms 19 20 – interpremaxillary suture, pm-mx.s – premaxillary-maxilla suture, re – 21 For Review Only 22 replacement crown, sf – special foramina. 23 24 25 Figure 41B. Scelidosaurus cf. harrisonii Owen, 1861. Medial view of the 26 27 right premaxillary dentition (BRSMG Ce12785). Photographic image of the 28 original specimen paired with an interpretative sketch of the same. 29 30 31 32 Abbreviations: M [numbered] maxillary tooth number, P [numbered] – 33 premaxillary tooth number, pm-mx.s – premaxillary-maxilla suture, re – 34 35 replacement crown. 36 37 38 39 Figure 42. Scelidosaurus harrisonii Owen, 1861. The dentition of the left 40 maxilla and dentary of the lectotype (NHMUK R1111). A. Lingual view of 41 42 the maxillary dentition. B. Buccal view of the maxillary dentition. C. Buccal 43 44 view of the dentary dentition. 45 46 47 Abbreviations: D [numbered] dentary tooth number, M [numbered] 48 49 maxillary tooth number. 50 51 52 Figure 43. Scelidosaurus harrisonii Owen, 1861. Tooth abrasion as 53 54 exposed on the left maxilla and dentary dentitions of the lectotype 55 (NHMUK R1111). The maxillary dentition, as preserved in lingual view, 56 57 opposing the dentary dentition. The dentary dentition has been reversed 58 59 60
150 Page 151 of 215 Zoological Journal of the Linnean Society
1 2 3 and portrayed as transparent so that the relationship between opposing 4 5 wear facets can be more readily visualised. 6 7 8 Abbreviations: af – abrasion facets, att – attritional facet, D [number] – 9 10 dentary tooth number, M [number] – maxillary tooth number. 11 12 Even tone: abrasion facets. Black tone: attrition facet created by friction 13 resulting from movement between adjacent crowns. 14 15 16 17 Figure 44. Scelidosaurus harrisonii Owen, 1861. Replacement pattern 18 exhibited on the medial surface of the lectotype maxilla (NHMUK R1111). 19 20 21 For Review Only 22 Abbreviations: ap – alveolar parapet, ft – functional tooth, g – vascular 23 groove, rep – replacement crowns, rt – replacement tooth, sf – special 24 25 foramina. 26 27 28 Even tone: highlights the replacement crowns. 29 30 31 32 Figure 45. Scelidosaurus harrisonii. Isolated 2nd ceratobranchials of: A. 33 34 The referred specimen (CAMSM X39256) and B. The lectotype (NHMUK 35 R1111). See also Figure 14. 36 37 38 39 Abbreviations: art – articular surfaces, br – broken surface. 40 41 42 Figure 46. Scelidosaurus cf. harrisonii Owen, 1861. Epistyloid bones. 43 44 Photographic image paired with an interpretative illustration. The posterior 45 portion of the skull and associated cervical osteoderm arrays of BRSMG 46 47 LEGL 0004 in left lateral view. Two elongate, distally bladed bones are 48 49 preserved projecting diagonally from the medio-ventral region of the skull. 50 These cannot be atlas ribs, one of these latter is preserved on this block 51 52 (at.r) and is much shorter and smaller. The right epistyloid is preserved 53 54 dislocated and rotated to a position beneath the left epistyloid; this 55 displacement seems to be in general accord with the rotation imposed 56 57 upon the neck/posterior skull region during burial and subsequent 58 59 compression of the skeleton. 60
151 Zoological Journal of the Linnean Society Page 152 of 215
1 2 3 4 5 Abbreviations: at.int – atlas intercentrum, at.na – atlas neural arch (left), 6 7 at.r – atlas rib (left), ba – base-plate osteoderm, l.styl – left epistyloid 8 bone, ost – epiosteoderms, pop – paroccipital process, Q – shaft of the 9 10 left quadrate, r.styl – right epistyloid bone. 11 12 Even tone: matrix. 13 14 15 Figure 47. Scelidosaurus harrisonii Owen, 1861. Skull reconstruction with 16 17 the left occipital horn and epistyloid in position. The latter bones are 18 highlighted in a pale, even tone. 19 20 21 For Review Only 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60
152 Page 153 of 215 Zoological Journal of the Linnean Society
1 2 3 REFERENCES 4 5 6 7 Agnolin FL, Rozadilla S. 2018. Phylogenetic reassessment of 8 9 Pisanosaurus mertii Casamiquela, 1967, a basal dinosauriform from 10 the Late Triassic of Argentina Journal of Systematic Palaeontology 11 12 16: 853-879. 13 14 Arbour VM, Currie PJ. 2016. Systematics, phylogeny and 15 16 palaeobiogeography of the ankylosaurid dinosaurs. Journal of 17 18 Systematic Palaeontology 14: 385-444. 19 20 Arbour VM, Evans DC. 2017. A new ankylosaurine dinosaur from the 21 For Review Only Judith River Formation of Montana, USA, based on an exceptional 22 23 skeleton with soft tissue preservation. Royal Society Open Science 24 25 4: 28. 26 27 Baron MG, Norman DB, Barrett PM. 2017a. Postcranial anatomy of 28 29 Lesothosaurus diagnosticus (Dinosauria: Ornithischia) from the 30 Lower Jurassic of southern Africa: implications for basal 31 32 ornithischian taxonomy and systematics. Zoological Journal of the 33 34 Linnean Society 179: 125-168. 35 36 Baron MG, Norman DB, Barrett PM. 2017b. A new hypothesis of 37 38 dinosaur relationships and early dinosaur evolution. Nature 543: 39 501-506. 40 41 Baron MG, Norman DB, Barrett PM. 2017c. Brief Communications: 42 43 Baron et al. reply. Nature 551: E4-E5. 44 45 Barrett PM. 2001. Tooth wear and possible jaw action of Scelidosaurus 46 47 harrisonii Owen and a review of feeding mechanisms in other 48 49 thyreophoran dinosaurs. In: Carpenter K, ed. The Armored 50 Dinosaurs. Bloomington Indianapolis: Indiana University Press. 25- 51 52 52. 53 54 Berman DS, McIntosh JS. 1986. Description of the lower jaw of 55 56 Stegosaurus (Reptilia: Ornithischia). Annals of the Carnegie 57 Museum 55: 29-40. 58 59 60
153 Zoological Journal of the Linnean Society Page 154 of 215
1 2 3 Blows WT. 2001. Dermal armor of the polacanthine dinosaurs. In: 4 5 Carpenter K, ed. The Armored Dinosaurs. Bloomington Indianapolis: 6 7 Indiana University Press. 363-385. 8 9 Breeden BT. 2016. Observations on the osteology of Scutellosaurus 10 lawleri Colbert, 1981 (Ornithischia: Thyreophora) on the basis of 11 12 new specimens from the Lower Jurassic Kayenta Formation of 13 14 Arizona. Unpublished Master of Science Thesis, University of Texas. 15 16 Brown B. 1908. The Ankylosauridae, a new family of armoured dinosaurs 17 18 from the Upper Cretaceous. Bulletin of the American Museum of 19 Natural History 24: 187-201. 20 21 For Review Only Brown CM. 2017. An exceptionally preserved armored dinosaur reveals 22 23 the morphology and allometry of osteoderms and their horny 24 25 epidermal coverings. PeerJ 5:4066; DOI 10.7717/peerj.4066: 26 39pp. 27 28 Brown CM, Henderson DM, Vinther J, Fletcher I, Sistiaga A, 29 30 Herrera J, Summons RE. 2017. An exceptionally preserved 31 three-dimensional armored dinosaur reveals insights into coloration 32 33 and Cretaceous predator-prey dynamics. Current Biology 27: 2514- 34 35 2521. 36 37 Brunsden D. 2002. Geomorphological roulette for engineers and 38 39 planners: some insights into an old game. Quarterly Journal of 40 Engineering Geology and Hydrology 35: 101-142. 41 42 43 Butler RJ. 2010. The anatomy of the basal ornithischian dinosaur 44 Eocursor parvus from the lower Elliot Formation (Late Triassic) of 45 46 South Africa. Zoological Journal of the Linnean Society 160: 648- 47 48 684. 49 50 Butler RJ, Smith RMH, Norman DB. 2007. A primitive ornithischian 51 dinosaur from the Late Triassic of South Africa, and the early 52 53 evolution and diversification of Ornithischia. Proceedings of the 54 55 Royal Society of London B 274: 2041-2046. 56 57 58 59 60
154 Page 155 of 215 Zoological Journal of the Linnean Society
1 2 3 Butler RJ, Upchurch P, Norman DB. 2008. The phylogeny of the 4 5 ornithischian dinosaurs. Journal of Systematic Palaeontology 6: 1- 6 7 40. 8 9 BZN. 1994. Opinion 1788. Scelidosaurus harrisonii Owen, 1861 (Reptilia, 10 Ornithischia): lectotype replaced. Bulletin of Zoological 11 12 Nomenclature 51: 288. 13 14 Cadbury, D. 2000. The Dinosaur Hunters. Fourth Estate Ltd: London. 15 16 Carpenter K, DiCroce T, Kinneer B, Simon R. 2013. Pelvis of 17 18 Gargoyleosaurus (Dinosauria: Ankylosauria) and the origin and 19 20 evolution of the ankylosaur pelvis. PLoS ONE 8(11): 13. 21 For Review Only 22 Carpenter K, Hayashi S, Kobayashi Y, Maryańska R, Barsbold R, 23 24 Sato K, Obata I. 2011. Sachania chulsanensis (Ornithischia, 25 Ankylosauridae) from the Upper Cretaceous of Mongolia. 26 27 Palaeontographica Abteil A. 293: 1-61. 28 29 Carpenter K, Kirkland JI, Burge D, Bird J. 2001. Disarticulated skull 30 31 of a new primitive ankylosaurid from the Lower Cretaceous of 32 eastern Utah. In: Carpenter K, ed. The armored dinosaurs. 33 34 Bloomington Indianapolis: Indiana University Press. 211-238. 35 36 Carpenter K, Miles C, Cloward K. 1998. Skull of a Jurassic ankylosaur 37 38 (Dinosauria). Nature 393: 782-783. 39 40 Casamiquela RF. 1967. Un nuevo dinosaurio ornitisquio triasico 41 42 (Pisanosaurus mertii: Ornithopoda) de la formacion Ischigualasto, 43 44 Argentina. Ameghiniana 4: 47-64. 45 46 Charig AJ. 1972. The evolution of the archosaur pelvis and hind limb: 47 an explanation in functional terms. In: Joysey KA and Kemp TS, 48 49 eds. Studies in Vertebrate Evolution. Edinburgh: Oliver & Boyd. 50 51 121-155. 52 53 Charig AJ. 1979. A new look at the Dinosaurs. British Museum [Natural 54 55 History]/ Heinemann: London. 56 57 Charig AJ, Horsfield B. 1975. Before the Ark. British Broadcasting 58 Corporation: London. 59 60
155 Zoological Journal of the Linnean Society Page 156 of 215
1 2 3 Charig AJ, Newman BH. 1992. Scelidosaurus harrisonii Owen, 1861 4 5 (Reptilia, Ornithischia): proposed replacement of inappropriate 6 7 lectotype. Bulletin of Zoological Nomenclature 49: 280-283. 8 9 Colbert EH. 1981. A primitive ornithischian dinosaur from the Kayenta 10 Formation of Arizona. Museum of Northern Arizona Bulletin 53: 1- 11 12 61. 13 14 Conybeare WD. 1822. Additional notices on the fossil genera 15 16 Ichthyosaurus and Plesiosaurus. Transactions of the Geological 17 18 Society of London VI: 103-123. 19 20 Conybeare WD. 1824. On the discovery of an almost perfect skeleton of 21 For Review Only the Plesiosaurus. Transactions of the Geological Society of London 1 22 23 (2nd series): 381-389. 24 25 Conybeare WD, De la Beche HT. 1821. Notice on the discovery of a 26 27 new fossil animal, forming a link between the Ichthyosaurus and 28 29 the crocodile, together with general remarks on the osteology of 30 the Ichthyosaurus. Transactions of the Geological Society of London 31 32 V: 558-594. 33 34 Coombs WP. 1971. The Ankylosauria. Unpublished PhD Thesis, Columbia 35 36 University. 37 38 Coombs WP. 1978. The families of the ornithischian dinosaur Order 39 40 Ankylosauria. Palaeontology 21: 143-170. 41 42 Coombs WP, Maryańska T. 1990. Ankylosauria. In: Weishampel DB, 43 44 Dodson P and Osmólska H, eds. The Dinosauria. Berkeley: 45 University of California Press. 456-483. 46 47 Coombs WP, Weishampel DB, Witmer LM. 1990. Basal Thyreophora. 48 49 In: Weishampel DB, Dodson P and Osmólska H, eds. The 50 51 Dinosauria. Berkeley: University of California Press. 427-434. 52 53 Crompton AW, Charig AJ. 1962. A new ornithischian from the Upper 54 55 Triassic of South Africa. Nature 196: 1074-1077. 56 57 Dall Sasso C, Signore M. 1998. Exceptional soft-tissue preservation in a 58 theropod from Italy. Nature 392: 383-387. 59 60
156 Page 157 of 215 Zoological Journal of the Linnean Society
1 2 3 Desmond AJ. 1975. The hot-blooded dinosaurs. A revolution in 4 5 palaeontology. Blond & Briggs: London. 6 7 Eaton TH. 1960. A new armored dinosaur from the Cretaceous of 8 9 Kansas. University of Kansas Paleontological Contributions: 10 Vertebrata 8: 1-24. 11 12 13 Edmund AG. 1960. Tooth replacement phenomena in the lower 14 vertebrates. Contributions to the Life Sciences Division of the Royal 15 16 Ontario Museum, Canada 52: 1-190. 17 18 Edmund AG. 1969. Dentition. In: Gans C, ed. Biology of the Reptilia. 19 20 London & New York: Academic Press. 117-200. 21 For Review Only 22 Ensom PC. 1987. Scelidosaur remains from the Lower Lias of Dorset. 23 24 Dorset Natural History and Archaeological Society Proceedings. 25 108: 203-205. 26 27 Gallois RW. 2008. Geological controls on the failure mechanisms within 28 29 the Black Ven-Spittles landslip complex, Lyme Regis, Dorset. 30 Geoscience in South West England 12: 9-14. 31 32 Galton PM. 1973. The cheeks of ornithischian dinosaurs. Lethaia 6: 67- 33 34 89. 35 36 Galton PM. 1974. The ornithischian dinosaur Hypsilophodon from the 37 38 Wealden of the Isle of Wight. Bulletin of the British Museum 39 40 (Natural History) Geology 25: 1-152. 41 42 Galton PM. 1990. Stegosauria. In: Weishampel DB, Dodson P and 43 44 Osmolska H, eds. The Dinosauria. Berkeley: University of California. 45 435-455. 46 47 Galton PM. 2001. Endocranial casts of the plated dinosaur Stegosaurus 48 49 (Upper Jurassic, western USA): a complete uindistorted cast and 50 51 the original specimens of Othniel Charles Marsh. In: Carpenter K, 52 ed. The Armored Dinosaurs. Bloomington Indianapolis: Indiana 53 54 University Press. 103-129. 55 56 Galton PM, Upchurch P. 2004. Stegosauria. In: Weishampel DB, 57 58 Dodson P and Osmólska H, eds. The Dinosauria, Second Edition. 59 60 Berkeley: University of California Press. 343-362.
157 Zoological Journal of the Linnean Society Page 158 of 215
1 2 3 Gauthier JA. 1986. Saurischian monophyly and the origin of birds. 4 5 Memoir of the California Academy of Sciences 8: 1-55. 6 7 Gilmore CW. 1914. Osteology of the armored Dinosauria in the U.S. 8 9 National Museum, with special reference to Stegosaurus. United 10 States National Museum Bulletin 89: 1-136. 11 12 13 Haas G. 1969. On the jaw muscles of ankylosaurs. American Museum 14 Novitates Issue 2399: 1-11. 15 16 Haubold H. 1990. Ein neuer Dinosaurier (Ornithischia, Thyreophora) aus 17 18 dem unteren Jura des nördlichen Mitteleuropa. Revue de 19 20 Paléobiologie 9: 149-177. 21 For Review Only 22 Hendrickx C, Araújo R, Mateus O. 2015. The non-avian theropod 23 24 quadrate 1: standardized terminology with an overview of the 25 anatomy and function. PeerJ 3: 39pp. 26 27 Hennig E. 1915. Stegosauria Fossilium Catalogus. 1. Animalia. 1-15. 28 29 30 Hesselbo SP, Jenkyns HC. 1995. A comparison of the Hettangian to 31 Bajocian successions of Dorset and Yorkshire. In: Taylor PD, ed. 32 33 Field Geology of the British Jurassic. London: The Geological Society 34 35 of London. 105-150. 36 37 Hill RV, D’Emic MD, Bever GS, Norell MA. 2015. A complex 38 39 hyobranchial apparatus in a Cretaceous dinosaur and the antiquity 40 of the avian paraglossalia. Zoological Journal of the Linnean Society 41 42 175: 892-909. 43 44 Holliday CM. 2009. New insights into dinosaur jaw muscle anatomy. The 45 46 Anatomical Record 292: 1246-1265. 47 48 Hopson JA. 1980. Tooth function and replacement in early Mesozoic 49 50 ornithischian dinosaurs: implications for aestivation. Lethaia 13: 51 93-105. 52 53 Horner JR, Weishampel DB, Forster CA. 2004. Hadrosauridae. In: 54 55 Weishampel DB, Dodson P and Osmólska H, eds. The Dinosauria, 56 57 Second Edition. Berkeley: University of California Press. 438-463. 58 59 60
158 Page 159 of 215 Zoological Journal of the Linnean Society
1 2 3 Hulke JW. 1881. Polacanthus foxii, a large undescribed dinosaur from 4 5 the Wealden Formation in the Isle of Wight. Philosophical 6 7 Transactions of the Royal Society of London 1881: 653-662. 8 9 Huxley TH. 1869. On the Dinosauria of the Trias, with observations on 10 the classification of the Dinosauria. Nature 1: 146. 11 12 13 Huxley TH. 1870. On the classification of the Dinosauria with 14 observations on the Dinosauria of the Trias. Quarterly Journal of the 15 16 Geological Society of London 26: 32-51. 17 18 Lang WD. 1924. The Blue Lias of the Devon and Dorset coasts. 19 20 Proceedings of the Geologists' Association 35: 169-185. 21 For Review Only 22 Lang WD. 1947. James Harrison of Charmouth, geologist (1819-1864). 23 24 Proceedings of the Dorset Natural History and Archaeological 25 Society 68: 103-118. 26 27 Lee Y-N. 1996. A new nodosaurid ankylosaur (Dinosauria: Ornithischia) 28 29 form the Early Cretaceous (late Albian) of Texas. Journal of 30 31 Vertebrate Paleontology 16: 232-245. 32 33 Leidy J. 1856. Notice of remains of fossil reptiles and fishes, discovered 34 35 by Dr F.V. Hayden in the badlands of the Judith River, Nebraska 36 Territory. Proceedings of the Academy of Natural Sciences, 37 38 Philadelphia 8: 72-73. 39 40 Leidy J. 1859a. An account of Hadrosaurus and its discovery. 41 42 Proceedings of the Academy of Natural Sciences of Philadelphia 10: 43 44 215-218. 45 46 Leidy J. 1859b. Hadrosaurus foulkii, a new saurian from the Cretaceous 47 of New Jersey related to the genus Iguanodon. American Journal of 48 49 Science 27: 266-270. 50 51 Lucas SG. 1996. The thyreophoran dinosaur Scelidosaurus from the 52 53 Lower Jurassic Lufeng Formation, Yunnan, China. In: Morales M, ed. 54 55 The Continental Jurassic. Flagstaff: Museum of Northern Arizona 56 Bulletin 60. 81-85. 57 58 Lull RS, Wright NE. 1942. Hadrosaurian dinosaurs of North America. 59 60 Geological Society of America. Special Paper: 1-242.
159 Zoological Journal of the Linnean Society Page 160 of 215
1 2 3 Lydekker R. 1888. Catalogue of the fossil Reptilia and Amphibia in the 4 5 British Museum (Natural History), London (Part 1). Trustees of the 6 7 British Museum (Natural History): London. 8 9 Maidment SCR, Norman DB, Barrett PM, Upchurch P. 2008. 10 Systematics and phylogeny of Stegosauria (Dinosauria: 11 12 Ornithischia). Journal of Systematic Palaeontology 6: 364-407. 13 14 Maidment SCR, Porro LB. 2010. Homology of the palpebral and origin 15 16 of supraorbital ossifications in ornithischian dinosaurs. Lethaia 43: 17 18 95-111. 19 20 Mantell GA. 1833. The geology of the South East of England. Longman, 21 For Review Only Rees, Orme, Brown, Green & Longman: London. 22 23 24 Mantell GA. 1852. On the structure of the Iguanodon, and on the fauna 25 and flora of the Wealden Formation. Proceedings of the Royal 26 27 Institution of Great Britain 1: 141-146. 28 29 Marsh OC. 1877. New order of extinct Reptilia (Stegosauria) from the 30 31 Jurassic of the Rocky Mountains. American Journal of Science 14: 32 513-514. 33 34 35 Marsh OC. 1889a. Comparison of the principal forms of Dinosauria of 36 Europe and America. Geological Magazine VI: 204-210. 37 38 39 Marsh OC. 1889b. Notice of gigantic horned Dinosauria from the 40 Cretaceous. American Journal of Science 38: 173-175. 41 42 Martill DM. 1991. Organically preserved dinosaur skin: taphonomic and 43 44 biological implications. Modern Geology 16: 61-68. 45 46 Martill DM, Batten DJ, Loydell DK. 2000. A new specimen of the 47 48 thyreophoran dinosaur cf. Scelidosaurus with soft tissue 49 50 preservation. Palaeontology 43: 549-559. 51 52 Maryańska T. 1977. Results of the Polish-Mongolian palaeontological 53 expeditions. Part VII. Ankylosauridae (Dinosauria) from Mongolia. 54 55 Palaeontologia Polonica 37: 85-151. 56 57 Maryańska T, Osmólska H. 1974. Results of the Polish-Mongolian 58 59 palaeontological expeditions. Part V. Pachycephalosauria, a new 60
160 Page 161 of 215 Zoological Journal of the Linnean Society
1 2 3 suborder of ornithischian dinosaurs. Palaeontologia Polonica 30: 45- 4 5 102. 6 7 Mateus O, Maidment SCR, Christiansen NA. 2009. A new long-necked 8 9 'sauropod mimic' stegosaur and the evolution of the plated 10 dinosaurs. Proceedings of the Royal Society of London. Series B 11 12 276: 1815-1821. 13 14 Nesbitt SJ. 2011. The early evolution of archosaurs: relationships and 15 16 origin of major clades. Bulletin of the American Museum of Natural 17 18 History 352: 292 pp. 19 20 Newman BH. 1968. The Jurassic dinosaur Scelidosaurus harrisoni, 21 For Review Only Owen. Palaeontology 11: 40-43. 22 23 24 Nopcsa FB. 1902. Notizen über Cretacische Dinosaurier. Sitzungsbericht 25 der Akademie der Wissenschaften, mathematische- 26 27 naturwissenschaften Klasse, Wien 111: 93-114. 28 29 Nopcsa FB. 1905. Notes on British Dinosaurs. Part II: Polacanthus. 30 31 Geological Magazine (new series) Decade V. Vol. II: 241-250. 32 33 Nopcsa FB. 1915. Die dinosaurier des siebenbürgischen Landesteile 34 35 Ungarns. Mitteilungen aus dem Jahrbuche der Königlich 36 ungarsichen Geologischen Reichsanstalt, Budapest XXIII: 3-24. 37 38 39 Norman DB. 1980. On the ornithischian dinosaur Iguanodon 40 bernissartensis from Belgium. Mémoires de l'Institut Royal des 41 42 Sciences Naturelles de Belgique 178: 1-105. 43 44 Norman DB. 1984. A systematic reappraisal of the reptile order 45 46 Ornithischia. In Reif W-E and Westphal F, eds. Proceedings of the 47 Third Symposium on Mesozoic Terrestrial Ecosystems. Tübingen: 48 49 Attempto Verlag, 157-162. 50 51 Norman DB. 1985. The illustrated encyclopedia of dinosaurs. 52 53 Salamander Books: London. 54 55 Norman DB. 1986. On the anatomy of Iguanodon atherfieldensis 56 57 (Ornithischia: Ornithopoda). Bulletin de l'Institut Royal des Sciences 58 Naturelles de Belgique 56: 281-372. 59 60
161 Zoological Journal of the Linnean Society Page 162 of 215
1 2 3 Norman DB. 2000a. Henry De la Beche and the plesiosaur's neck. 4 5 Archives of Natural History 27: 137-148. 6 7 Norman DB. 2000b. Professor Richard Owen and the important, but 8 9 neglected, dinosaur Scelidosaurus harrisonii. Historical Biology 14: 10 235-253. 11 12 13 Norman DB. 2001. Scelidosaurus, the earliest complete dinosaur. In: 14 Carpenter K, ed. The armored dinosaurs. Bloomington Indianapolis: 15 16 Indiana University Press. 3-24. 17 18 Norman DB. 2017. Dinosaurs: A very short introduction. Oxford 19 20 University Press: Oxford. 21 For Review Only 22 Norman DB. submitted. Scelidosaurus harrisonii Owen, 1861 from the 23 24 early jurassic of dorset, england: Part 2. Postcranial endoskeleton. 25 26 Zoological Journal of the Linnean Society. 27 28 Norman DB. submitted. Scelidosaurus harrisonii Owen, 1861 from the 29 early jurassic of dorset, england: Part 3. The dermal skeleton. 30 31 Zoological Journal of the Linnean Society. 32 33 Norman DB. in preparation. Scelidosaurus harrisonii Owen, 1861 34 35 (dinosauria: ornithischia) from the early jurassic of dorset, england. 36 37 Part 4. Biology, systematics and phylogenetics. 38 39 Norman DB, Porro LB. in preparation. Digital visualisation of the 40 anatomy of the skull of Scelidosaurus harrisonii Owen, 1861. 41 42 43 Norman DB, Barrett PM. 2002. Ornithischian dinosaurs from the Lower 44 Cretaceous (Berriasian) of England. In: Milner AR and Batten DJ, 45 46 eds. Special Papers in Palaeontology: Life and Environments in 47 48 Purbeck Times. Oxford: Blackwells. 161-189. 49 50 Norman DB, Butler RJ, Maidment SCR. 2007. Reconsidering the status 51 and affinities of the ornithischian dinosaur Tatisaurus oehleri 52 53 Simmons, 1965. Zoological Journal of the Linnean Society 150: 54 55 865-879. 56 57 Norman DB, Crompton AW, Butler RJ, Porro LB, Charig AJ. 2011. 58 59 The Lower Jurassic ornithischian dinosaur Heterodontosaurus tucki 60
162 Page 163 of 215 Zoological Journal of the Linnean Society
1 2 3 Crompton & Charig, 1962: cranial anatomy, functional morphology, 4 5 taxonomy and relationships. Zoological Journal of the Linnean 6 7 Society 163: 182-276. 8 9 Norman DB, Weishampel DB. 1985. Ornithopod feeding mechanisms: 10 their bearing on the evolution of herbivory. American Naturalist 11 12 126: 151-164. 13 14 Norman DB, Witmer LM, Weishampel DB. 2004a. Basal Ornithischia. 15 16 In: Weishampel DB, Dodson P and Osmólska H, eds. The 17 18 Dinosauria, Second Edition. Berkeley: University of California Press. 19 325-334. 20 21 For Review Only Norman DB, Witmer LM, Weishampel DB. 2004b. Basal Thyreophora. 22 23 In: Weishampel DB, Dodson P and Osmólska H, eds. The 24 25 Dinosauria, Second Edition. Berkeley: University of California Press. 26 335-342. 27 28 29 Ostrom JH. 1961. Cranial morphology of the hadrosaurian dinosaurs of 30 North America. Bulletin of the American Museum of Natural History 31 32 122: 33-186. 33 34 Ösi A. 2005. Hungarosaurus tormai, a new ankylosaur (Dinosauria) from 35 36 the Upper Cretaceous of Hungary. Journal of Vertebrate 37 38 Paleontology 25: 370-383. 39 40 Ösi A, Barrett PM, Földes T, Tokai R. 2014. Wear pattern, dental 41 function and jaw mechanism in the Late Cretaceous ankylosaur 42 43 Hungarosaurus. Anatomical Record 297: 1165-1180. 44 45 Ösi A, Prondvai E, Mallon J, Bodor ER. 2016. Diversity and 46 47 convergences in the evolution of feeding adaptations in ankylosaurs 48 49 (Dinosauria: Ornithischia). Historical Biology 29: 539-570. 50 51 Ösi A, Weishampel DB. 2009. Jaw mechanism and dental function in 52 the Late Cretaceous basal eusuchian Iharkutosuchus. Journal of 53 54 Morphology 270: 903-920. 55 56 Owen R. 1842. Report on British Fossil Reptiles. Part 2. Report of the 57 58 British Association for the Advancement of Science (Plymouth) XI: 59 60 60-204.
163 Zoological Journal of the Linnean Society Page 164 of 215
1 2 3 Owen R. 1859. Palaeontology Encyclopaedia Britannica. 8th ed. 4 5 Edinburgh: Adam & Charles Black. 91-176. 6 7 Owen R. 1860. Palaeontology, or a systematic summary of extinct 8 9 animals and their geological relations. Adam & Charles Black: 10 Edinburgh. 11 12 13 Owen R. 1861. Monograph of the Fossil Reptilia of the Liassic 14 Formations. Part 1. A monograph of a fossil dinosaur 15 16 (Scelidosaurus harrisonii Owen) of the Lower Lias. 17 18 Palaeontographical Society Monographs XII: 1-14. 19 20 Owen R. 1862. On the dinosaurian reptile (Scelidosaurus Harrisonii), 21 For Review Only from the Lower Lias of Charmouth. Reports of the British 22 23 Association for the Advancement of Science 1862: 121-122. 24 25 Owen R. 1863. A monograph of the Fossil Reptilia of the Liassic 26 27 Formations. Part II. A monograph of a fossil dinosaur 28 29 (Scelidosaurus harrisonii Owen) of the Lower Lias. 30 Palaeontographical Society Monographs XIII: 1-26. 31 32 Owen R. 1875. A monograph on the Fossil Reptilia of the Mesozoic 33 34 Formations. Part II. (Genera Bothriospondylus, Cetiosaurus, 35 36 Omosaurus). Palaeontographical Society Monographs XXIX: 15-93. 37 38 Padian K. 1989. Presence of the dinosaur Scelidosaurus indicates 39 40 Jurassic age for the Kayenta Formation (Glen Canyon Group, 41 northern Arizona). Geology 17: 438-441. 42 43 44 Padian K. 2013. The problem of dinosaur origins: integrating three 45 approaches to the rise of Dinosauria. Earth and Environmental 46 47 Science Transactions of the Royal Society of Edinburgh 103: 423- 48 49 442. 50 51 Porro LB, Witmer LM, Barrett PM. 2015. Digital preparation and 52 osteology of the skull of Lesothosaurus diagnosticus (Ornithischia: 53 54 Dinosauria). PeerJ 3: 44pp. 55 56 Raven TJ, Maidment SCR. 2017. A new phylogeny of Stegosauria 57 58 (Dinosauria: Ornithischia). Palaeontology 60: 401-408. 59 60
164 Page 165 of 215 Zoological Journal of the Linnean Society
1 2 3 Rixon AE. 1949. The use of acetic and formic acids in the preparation of 4 5 fossil vertebrates. The Museums Journal 49. 6 7 Rixon AE. 1968. The development of the remains of a small 8 9 Scelidosaurus from a Lias nodule. Museums Journal 67: 315-321. 10 11 Romer AS. 1927. The pelvic musculature of ornithischian dinosaurs. Acta 12 13 Zoologica 8: 225-275. 14 15 Romer AS. 1956. Osteology of the Reptiles. University of Chicago Press: 16 Chicago. 17 18 19 Romer AS. 1966. Vertebrate Paleontology. University of Chicago Press: 20 Chicago. 21 For Review Only 22 Romer AS. 1968. Notes and comments on Vertebrate Paleontology. 23 24 University of Chicago Press: Chicago. 25 26 Rudwick, MJS. 1992. Scenes from Deep Time: early pictorial 27 28 representations of the prehistoric world. University of Chicago 29 30 Press: Chicago & London. 31 32 Russell LS. 1940. Edmontonia rugosidens (Gilmore), an armoured 33 dinosaur from the Belly River Series of Alberta. University of 34 35 Toronto Studies, Geological Series 43: 3-28. 36 37 Rybczynski N, Vickaryous MK. 2001. Evidence of complex jaw 38 39 movement in the Late Cretaceous ankylosaurid Euoplocephalus 40 41 tutus (Dinosauria: Thyreophora). In: Carpenter K, ed. The Amored 42 Dinosaurs. Bloomington Indianapolis: Indiana University Press. 43 44 299-318. 45 46 Santa Luca AP. 1980. The postcranial skeleton of Heterodontosaurus 47 48 tucki (Reptilia: Ornithischia) from the Stormberg of South Africa. 49 50 Annals of the South African Museum 79: 159-211. 51 52 Santa Luca AP, Crompton AW, Charig AJ. 1976. A complete skeleton 53 of the Late Triassic ornithischian Heterodontosaurus tucki. Nature 54 55 264: 324-328. 56 57 58 59 60
165 Zoological Journal of the Linnean Society Page 166 of 215
1 2 3 Schoch RJ. 2007. Osteology of the small archosaur Aetosaurus from the 4 5 Upper Triassic of Germany. Neues Jahrbuch für Geologie und 6 7 Paläontologie - Abhandlungen 246: 1-35. 8 9 Seeley HG. 1887. On the classification of the fossil animals commonly 10 named Dinosauria. Proceedings of the Royal Society of London 43: 11 12 165-171. 13 14 Sereno PC. 1984. The phylogeny of the Ornithischia: a reappraisal. In 15 16 Reif W-E and Westphal F, eds. Proceedings of the Third Symposium 17 18 on Mesozoic Terrestrial Ecosystems. Tübingen: Attempto Verlag, 19 219-226. 20 21 For Review Only Sereno PC. 1986. Phylogeny of the bird-hipped dinosaurs. National 22 23 Geographic Research 2: 234-256. 24 25 Sereno PC. 1991. Lesothosaurus, 'Fabrosaurids' and the early evolution 26 27 of Ornithischia. Journal of Vertebrate Paleontology 11: 234-256. 28 29 Sereno PC. 2012. Taxonomy, morphology, masticatory function and 30 31 phylogeny of heterodontosaurid dinosaurs. ZooKeys 226: 225pp. 32 33 Sereno PC, Dong Z. 1992. The skull of the basal stegosaur 34 35 Huayangosaurus taibaii. Journal of Vertebrate Paleontology 11: 36 318-343. 37 38 39 Steel R. 1969. Ornithischia. Gustav Fischer Verlag: Stuttgart. 40 41 Theis S. 2010. Origin and development of cucullaris/trapexius- 42 sternocleidomastoid muscles in birds, mammals and reptiles. 43 44 Unpublished Inaugural-Dissertation, Albert-Ludwigs-Universität 45 46 Freiburg im Breisgau. 47 48 Thulborn RA. 1970. The skull of Fabrosaurus australis, a Triassic 49 50 ornithischian dinosaur. Palaeontology 13: 414-432. 51 52 Thulborn RA. 1972. The postcranial skeleton of the Triassic ornithischian 53 dinosaur Fabrosaurus australis. Palaeontology 15: 29-60. 54 55 56 Thulborn RA. 1977. Relationships of the Lower Jurassic dinosaur 57 Scelidosaurus harrisonii. Journal of Paleontology 51: 725-739. 58 59 60
166 Page 167 of 215 Zoological Journal of the Linnean Society
1 2 3 Thulborn RA. 1978. Aestivation among ornithopod dinosaurs of the 4 5 African Trias. Lethaia 11: 185-198. 6 7 Toombs HA, Rixon AE. 1950. The use of plastics in the "Transfer 8 9 Method" of preparing fossils. The Museums Journal 50: 105-107 10 11 Toombs HA, Rixon AE. 1959. The use of acids in the preparation of 12 13 vertebrate fossils. Curator 2. 14 15 Torrens HS. 1995. Mary Anning (1799-1847) of Lyme; 'the greatest 16 fossilist the world ever knew'. British Journal for the History of 17 18 Science 28: 257-284. 19 20 Tumanova TA. 1987. The armored dinosaurs of Mongolia [in Russian]. 21 For Review Only 22 The Joint Soviet-Mongolian Paleontological Expedition 32: 1-80. 23 24 Tykoski RS. 1998ms. The osteology of Syntarsus kayentakae and its 25 26 implications for ceratosaurid phylogeny. Unpublished Master of 27 Science Thesis, University of Texas. 28 29 30 Vickaryous MK, Russell AP, Currie PJ. 2001. Cranial ornamentation of 31 ankylosaurs (Ornithischia: Thyreophora): reappraisal of 32 33 developmental hypotheses. In: Carpenter K, ed. The Armored 34 35 Dinosaurs. Bloomington Indianapolis: Indiana University Press. 36 318-340. 37 38 39 Weishampel DB, Dodson P, Osmólska H eds. 2004. The Dinosauria. 40 Berkeley and Los Angeles: University of California Press. 41 42 Whybrow PJ. 1985. A history of fossil collecting and preparation 43 44 techniques. Curator 28: 5-26. 45 46 Wilford JN. 1985. The Riddle of the Dinosaur. Alfred Knopf: New York. 47 48 49 50 51 52 53 54 55 56 57 58 59 60
167 Zoological Journal of the Linnean Society Page 168 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 169 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Owen's syntypes 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 170 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 171 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Owen's skull lithographs 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 172 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 265x100mm (299 x 299 DPI) 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 173 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 174 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 175 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 176 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 177 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 178 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 179 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 180 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 181 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 182 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Nasals - revised annotation 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 183 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 184 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 185 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 186 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 187 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 188 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 189 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 190 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 191 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 192 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 193 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 194 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 195 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 196 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 197 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 198 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 199 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 200 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 201 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 202 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 203 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 204 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 205 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 206 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (79 x 79 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 207 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 208 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 209 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 Added legend and indication of placement re. 41A 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 210 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 297x209mm (200 x 200 DPI) 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 211 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 212 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 213 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Zoological Journal of the Linnean Society Page 214 of 215
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 210x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60 Page 215 of 215 Zoological Journal of the Linnean Society
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 For Review Only 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 209x297mm (200 x 200 DPI) 48 49 50 51 52 53 54 55 56 57 58 59 60