A photo documentation of bipedal ornithischian dinosaurs from the Upper Jurassic Morrison Formation, USA

Bipedal ornithischian dinosaurs from the Upper Jurassic Morrison Formation are rare, forming only about 15% of the dinosaur specimens. Nevertheless, one of them was among the first dinosaurs named from what was then the ‘’Atlantosaurus Beds’’ of Colorado. Collecting and restudy for 140 years has in­creased the diversity from the initial 1877 discovery to the currently valid four genera and six species, viz., Fruitadens haagaroum, Nanosaurus agilis, Camptosaurus dispar, C. aphanoecetes, Dryosaurus altus, and D. elderae, which we briefly review. We demonstrate that the enigmatic Nanosaurus agilis is the senior name for Drinker nisti, Othnielosaurus consors, and Othnielia rex. In addition, a new species, Dryosaurus elderae is proposed for the Dryosaurus specimens from Dinosaur National Monument that are characterized by elongate cervical verebrae and a long, low ilium among other features.


INTRODUCTION
The late Paleozoic shark record from Utah is poorly known, but what is known was summarized by Sumida and others (1999). They report teeth of the freshwater xenacanth Orthacanthus from the Lower Permian Halgaito Shale (Cutler Group) and an unnamed partial chondrichthyean tooth from the nonmarine Lower Permian Organ Rock Formation (Cutler Group). In addition, Miller (1981) reported on occurrences of the form genus Cladodus sp. from the Mississippian Great Blue Limestone and the Manning Canyon Shale in north central Utah. One of us (Ottinger) amassed a small collection of marine shark teeth during the 1970s from the area between the potash mine and evaporation ponds in the Shafer Basin southwest of Moab, Utah. This is the same general area that Lohman (1974) reported the cochliodont shark Deltodus sp. and the petalodont shark Petalodus. Lohman's specimens have not been figured or described and their current location are unknown.
Owing to the importance of documenting the new collection, we describe and illustrate the material below. Original specimens or casts of specimens are curated at the Prehistoric Museum, Utah State University Eastern, Price, Utah. Exact locality information for each specimen is no longer known.

GEOLOGICAL SETTING
The specimens were collected in the Shafer Basin, which is located between Cane Creek and Shafer anticlines, southwest of Moab Utah (figure 1). Here, the Permian strata are well exposed along the Colorado River below the cliffs of Dead Horse Point State Park (figure 2A). Matrix adhering to the teeth show that all but one came from the Shafer limestone bed, an informal name applied to a prominent limestone in the Shafer Basin. This limestone caps a 90 m interval (figure 2B) of alternating marine limestones and muddy sandstones containing invertebrate fossils and burrows (fig-ure 2C), and nonmarine arkosic sandstones and aeolian sandstones containing rhizoliths (figure 2D), including a possible lycopod root (figure 2E).
These strata have been variously called or mapped in this basin as the Rico Formation (Prommel, 1923;Baker and others, 1927;McKnight, 1974;Hinrichs and others, 1967;Loope, 1984), Rico transition facies (Wengerd and Matheny, 1958), Elephant Canyon Formation (Terrell, 1972;Campbell, 1987), the lower Cutler beds (Condon, 1997;Doelling and others, 1994;Doelling and Chidsey, 2009;Dubiel and others, 2009), lower Cutler Group (Doelling, 2004), or simply considered as part of the Cutler Formation undifferentiated (Mattox and Brand, 1974;Huntoon and others, 2002). We refer these marine-continental transition strata as the "lower Cutler beds" as used by Loope and others (1990), with the Shafer limestone as the top of this unit. Petalodus is a distinctive form genus characterized by a labio-lingually compressed, vertically short, broadbased, triangular crown that is convex anteriorly and concave posteriorly. The base of the crown is bordered by a band or cingulum composed of imbricated ridges that are especially well developed on the lingual side (Leidy, 1856, first referred to the structure as a cingulum and is followed here; Robb, 2003 refers to it as the "distal crown tongue"). The crown is situated on a long, tapering root (base). The five teeth vary in size and in the height and width of the crowns; their measurements are given in table 1. Part of these differences may be ontogenetic; wear and location in the jaws as has long been recognized (e.g., Newberry and Worthen, 1866;Eastman, 1896). The specimens are assigned to the species P. ohioenesis because of the vertically narrow cingulum around the crowns; the cingulum is much wider in P. acuminatus. Hansen (1985) notes that P. ohioenesis is the only species in the Pennsylvanian and Lower Permian of the United States.

CEUM 81502 (figures 3A to 3C)
The most complete tooth in the collection, it has a crown that is convex on the labial surface and concave on the lingual side just above the root. The concavity may accommodate the convex surface of the succeeding replacement tooth (figure 4A) in a manner similar to those reported by Davis (1883) for the petalodontiforms Petalorhynchus and Glossodus. Lucas and others (2011) suggested that the concavity accommodated the crown apex of the opposing tooth (figure 4B). In profile view, the root and crown are slightly angled relative to one another. The cutting edges of the crown terminate in a  slightly acuminate asymmetrical apex. A narrow band on each side of the edges shows tiny parallel dentine tubules. The labial band shows irregular wear, which extends onto the crown and exposes the underlying tangled-fibered enameloid as described and illustrated by Lund (1989, figure 16). A broad, slightly developed medial ridge extends from the apex towards the base were it merges. The labial side of the crown base is sigmoid and shows faint traces of ridges where the enameloid is preserved. In profile, the crown overhangs the root, which becomes thicker distally. The root is D-shaped in horizontal cross-section, with the lingual side flat. The acuminate and slight asymmetry of the crown indicate this is an anteromedial tooth (Robb, 2003).

CEUM 81503 (figure 3D)
The smallest of the teeth, it is also the most worn. Much of the upper part of the crown reveals the enameloid on the surface well below the parallel dentine tubule band on the labial side. The apex is worn to a notch and the medial ridge is absent. The low, elongate crown indicate that it is a lateral tooth (Robb, 2003).

CEUM 81504 (figure 3E)
Unlike the other teeth, this one was recovered from a white to light grey, arkosic, coarse sandstone, which Terrell (1972) states only occurs in the middle of the upper limestone (unit 20) in Shafer Basin. The tooth is white, rather than gray to reddish colored. The medial crown ridge is absent. The enameloid surface is etched and in places has remnants of pink feldspar and white quartz embedded. The parallel dentine tubules are accentuated from wear. The root is missing, but the low, elongate crown shows that it is a lateral tooth (Robb, 2003).

CEUM 81643 cast (figure 3F)
Much of the crown is damaged, either lacking enameloid or missing exposing the inner core. The impression shows that the apex was not acuminate, possible due to wear. A trace of the medial ridge is present. The root is incomplete, but enough remains to show that it was expanded distally in profile. The low, elongate crown, with low apex suggests it was lateral to the medial teeth, but not far posterolaterally in the jaw.

CEUM 81644 cast (figure 3G)
The crown is also damaged, with all of the enameloid eroded exposing the inner core. The lingual side of the tooth is exposed. The distal end of the root is damaged, but enough remains to suggest it curved labially much like one of the specimens figured by Lucas and others (2011; figure 3D). The high crown suggests this was an anteromedial tooth.

CEUM 81505 tooth (figure 5H)
The tooth is heavily damaged, having an incomplete mid-crown and a base of one side crown attached to an incomplete base. It is assigned to the form genus on the basis of the large median cusp that is convex on the lingual side, flat on the labial side, and the broken base of a smaller, rounded in cross section lateral cusp. This specimen is the youngest occurrence of Cladodus in Utah. The two others reported by Miller (1981) are from the Mississippian Great Blue Limestone and from the Manning Canyon Shale. Adhering matrix is a dull, reddish-brown, fine, micaceous sandstone.

A REVIEW OF SOME KEY SPECIMENS OF PETALODUS
The five teeth of Petalodus from Utah are the first described and illustrated from the state. At least 21 species of Petalodus have been named (Hansen, 1985), the majority during the 1800s when the range of variation of tooth shape was less well known. The consensus today is that the majority of the species either belong to other genera or are morphological variants based on position within the jaws (e.g., Hansen, 1985;Lucas and others, 2011;Ginter and others, 2015). Unfortunately, there is no agreement as to whether there is only a single species (Lucas and others, 2011) or two species that are chronostratigraphical distributed (Hansen, 1985). As a result, it was necessary for us to look closely at several key species and specimens in order to determine the correct species name for the Utah specimens. The genus was proposed by Richard Owen (1840-1845 as Petalodus hastingsii, p. 61; also given as P. hastingsae, figure caption). However, there is a problem regarding this specimen that apparently has not been discussed before in the literature. Namely, that the description and figures of type tooth do not remotely look anything like the teeth traditionally referred to Petalodus. Owen (1840Owen ( -1845 3) described the tooth as "A thin lamella, slightly concave like a finger-nail …which I have, on that account, named Petalodus" (Greek petalon leaf, and odus tooth). His more detailed description (p. 61-62) refers to the tooth being "lamelliform, " i.e., thin plate form, plate-like or scale-like. The tooth (NHMUK PV P613) was illustrated as more complete than it actually is (figures 5A to 5C vs 5D to 5G). Woodward (1889) referred to the specimen as a fragment, so was probably not more complete when described by Owen. It is most likely that the tooth was reconstructed by mirroring the fragment to give the impression of how an entire tooth would look (E. Barnard, Natural History Museum, London, England, written communication, 2017). Such a technique was used, for example, by Yale paleontologist O.C. Marsh (e.g., 1896, pl. 21, Brontosaurus excelsus dorsal). Despite Owen misleading reconstruc-tion, Louis Agassiz correctly referred a nearly complete tooth from his collection to the genus Petalodus, thus establishing the sub-rhomboidal or subtriangular morphology by which it is known.
This version of Petalodus, rather than the rectangular shape given by Owen, was established in Agassiz's multi-year (1833Agassiz's multi-year ( -1845, 10 volume, limited edition "Recherches sur les Poissons Fossiles. " The multi-year publication was due to his constant revisions (18 non-consecutive revised printings of the 10 volumes, Jeannet 1928). In text volume 3, he named a new genus, Chomatodus, and three species for several fossil fish teeth he states are from the Bristol Museum, England (but also stated as received from Roderick Murchison, p. 108). One of these teeth was named (text: Agassiz, 1833-1845a, p. 108-109) and figured (atlas: Agassiz, 1833-1845b, plate 19, figures 11 to 13) as Chomatodus acuminatus (figures 5S and 5T), which Woodward (1889) states were published in 1838 and confirmed by Brown (1890) and Jeannet (1928, p. 120). Later (revised text: Agassiz, 1833-1845a, p. 159), Agassiz writes that the tooth differs too much from the other two species of Chomatodus "since it is a tooth with a cutting blade…It is in the new genus Petalodus, by Mr. Owen. " He reiterates this in a footnote (p. 174), which Woodward gives as published in 1843 and confirmed by Brown (1890) and Jeannet (1928, p. 122). The "cutting blade" is apparently the only justification that Agassiz gives as his reason for considering Chomatodus acuminatus as belonging to the genus Petalodus. The specimen was damaged (figure 5U) sometime between 1838 and 1927 as first reported by Jeannet (1927, p. 109) "Exemplaire mutilé depuis qu'il a été figuré" ("Specimen mutilated since it was figured"). Agassiz had a second tooth (figure 5V) also acquired from the Bristol Museum that he never discussed or figured.
Despite glaring differences between the Owen and Agassiz illustrations and descriptions, subsequent authors were quick to accept Agassiz's synonymy. The same year that Agassiz referred C. acuminatus to Petalodus, geologist Joseph Portlock (1843) described and illustrated a tooth (figure 5W) that he referred to as "Petalodus Hastingsii (Agassiz)" [sic]. His reasons for this identification was never stated. Portlock was followed a few years later by Frederick M'Coy (1848), who named  (2011) for the same convexo-concave feature. The serial stacking is more in keeping with the serial replacement of teeth seen in sharks.
Petalodus rhombus for a complete tooth, noting its more rhomboid crown as compared with P. acuminatus (he says nothing of P. hastingsii). However, in 1854, M'Coy (1854) synonymized P. rhombus with P. acuminatus and illustrated the specimen (figures 5X to 5Z). The synonymy was based on what he thought were a range of variation in several new specimens and marks the first time that morphological variation was taken into account in Petalodus taxonomy. Although the low profile of the crown is also seen in some Petalodus teeth (e.g., figures 5N and 5O), the tooth lacks the ridges along the base of the crown and cingulum that typify Petalodus. It is therefore possible that this tooth does not belong to that genus.
Woodward (1889) 5E). Most importantly, these teeth are from the same locality and horizon as Owen's holotype P. hastingsii (Woodward, 1889). The teeth share with the holotype, NHMUK PV P613, the unique, wide (deep) ridged band that occupies the lower half or more of the crown on the lingual side (figures 5D and 5E arrow). In all other species of Petalodus, this ridged band occupies a narrow zone at the base of the crown, where it is angled ventroposteriorly and protrudes so that a cingulum is formed (e.g., figures 5S and 5T). In addition, the lingual side of P. hastingsii is only slightly concave in side view (figures 5F and 5G) compared to most other Petalodus (e.g., figure 5S), and the root is proportionally short compared to crown height. Other specimens referred to Petalodus share with NHMUK PV P5342 an acuminate apex crown that is convex and subtriangular on the labial side where a well-developed anteroventrally facing, inverted W-shaped ridged cingulum lies at the base of the crown. We therefore conclude that P. hastingsii is a valid species contrary to Woodward (1889). In addition, P. hastingsii is known from the Flechado Formation (Desmoinesian, Middle Pennsylvanian) of New Mexico (Zidek and Kietzke, 1993) (figures 5P to 5R).
Owing to the importance of the Owen and Agassiz holotype specimens to the taxon Petalodus, their stratigraphic position needs to be established (Interna-tional Code on Zoological Nomenclature Article 76). Woodward (1889) reports that the teeth here referred to Petalodus hastingsii were collected from the "Upper Carboniferous Limestone" at the village of Ticknall in South Derbyshire, England. This stratum is now called the Ticknall Limestone (Monteleone, 1973;Carney and others, 2001) and was an important source of building stone. The quarries produced numerous fossils (Parsons, 1917;Monteleone, 1973), which date the strata as Brigantian (upper Visean, a.k.a. upper Middle Mississippian) (Waters and others, 2009). Ginter and others (2015) described non-petalodontiform shark teeth from these beds, which also supported a Visean age. Agassiz obtained the holotype Petalodus acuminatus (Agassiz), MHNN-FOS 171, from British geologist Roderick Murchison, who collected the specimen from the Carboniferous Limestone near the town of Whorlton in County Durham, England. This stratum is identified by the British Geological Survey (2017) as the Stainmore Formation (middle Carboniferous or Namurian, a.k.a. Upper Mississippian to Lower Pennsylvanian). The formation is characterized by numerous limestone beds known to have been historically quarried as building stone (King, 2012).
The Petalodus teeth from the lower Cutler beds better compare with Petalodus acuminatus than to P. hastingsii chiefly in lacking the wide ridged band on the lingual side. However, unlike P. acuminatus, the lingual cingulum is very narrow and the crown taller relative to width. They are more similar to the tooth named and illustrated as Getalodus ohioensis (figures 5A' and 5B') by James M. Safford (1853). The generic name is most certainly a typographical error (Hay, 1895), either due to misinterpretation of Safford's handwritten manuscript by the typesetter, or Safford's misunderstanding of Louis Agassiz' heavy Swiss French accent during conversations Safford (1853) states he had with Agassiz about the tooth at the 1851 meeting of American Association for the Advancement of Science. The tooth was subsequently referred to as the holotype of Petalodus ohioensis (e.g., Hay, 1895;Lucas and others, 2011; we were unable to locate this specimen to re-illustrate). Safford reports that the specimen was collected from near Cambridge, Ohio, and Condit (1912) that it came from the Cambridge Limestone, which is in the Conemaugh Forma-Permo-Pennsylvanian shark teeth from the Lower Cutler beds near Moab, Utah Carpenter, K., and Ottinger, L.. Geology of the Intermountain West 2018 Volume 5 Figure 5. Caption on following page.
tion (Rice and others, 1994), or the Glenshaw Formation if the Conemaugh is raised to group (e.g., Rollins and others, 1979;Heckel and others, 2011). Conodonts demonstrate a middle to upper Missourian Stage (Middle Pennsylvanian) for the Cambridge (Heckel and others, 2011;Barrick and others, 2013). Oddly, Newberry (1875) says nothing about this specimen in his review of fossil fishes from Ohio, but does refer all Petalodus to P. allegheniensis (see below). P. acuminatus supposedly differs from P. ohioensis in the wide, ridged, lingual cingulum (Hansen, 1997). The specimens of Newberry are important because they appear to represent additional specimens of P. ohioensis and provide more accurate information about this species (attempts to locate these specimens to re-illustrate were unsuccessful). Newberry reports that the specimens are all from the Crinoidal Limestone, which Condit (1912) identifies as the Ames Limestone of the Conemaugh Formation (or the Glenshaw Formation if the Conemaugh is raised to group). Conodonts place the Ames in the lower Virgilian (Heckel and others, 2011); i.e., lower Upper Pennsylvanian, and closer in time to the Petalodus specimens from Utah. Two of the teeth were figured by Newberry (figures 5C' and 5D'), and one of them (figure 5D') shows a tall crown and similar crown profile to the illustration by Safford (figures 5A' and 5B'). These teeth also confirm the one feature that has been used to diagnose P. ohioensis, namely the vertically narrow cingulum band of ridges (e.g., Hansen, 1985;Dalla Vecchia 1988), which is about one-third to one-half that of P. alleghaniensis (see next).
Two years after Safford named Petalodus ohioensis, Leidy (1855) named Sicarius extinctus for a tooth he described as resembling a pangolin scale. It was renamed and figured (figures 5E' and 5G') as Petalodus alleghaniensis the following year (Leidy, 1856); the reason for changing the species name was not given by Leidy. P. extinctus was used only once, by Eastman (1896) and is considered nomen oblitum (ICZN Art. 23.9). The tooth described by Leidy was recovered from the Glenshaw Formation at Bens Creek Station in what is now the Allegheny Portage Railroad National Historic Site (Koch  Zidek and Kietzke, 1993). Note that the ridges occur at the base of the crown on the lingual side and on the cingulum of the labial side. Holotype of Petalodus acuminatus (as Chomatodus acuminatus) as illustrated by Agassiz (1838, pl. 19, figs. 11, 13) (from Shafford 1853, p. 142); no scale given. The whereabouts of this specimen is unknown. Two specimens (C' and D') described by Newberry (1875, plate 58, figures 13 and 13a). Holotype of Petalodus allegeniensis as illustrated by Leidy (1856, plate 16, figures 4 and 5) compared with recent photographs: (E' and F') in labial and (G' and H') lingual views. Arrows point to the ridged basal band or cingulum. The origin of the green dot on specimen D is uncertain, but may have been the code used by Woodward (1889) or Owen (1840Owen ( -1845 to denote it was a holotype. The red dot on specimen F is a blob of wax used by Woodward (1889) to denote that he include the specimen in his catalog of fossil fishes. The blue star is probably the old symbol used to denote a holotype. Scales in cm.
and Santucci, 2004). Conodonts from the Glenshaw are Missourian Stage (Middle Pennsylvanian) (Heckel and others, 2011;Barrick and others, 2013) and thus, P. alleghaniensis is contemporaneous with P. ohioensis. Hay (1895) raised the possibility that P. alleghaniensis might be synonymous with P. ohioensis, although in the end he did not accept that. The holotype of P. alleghaniensis does differ from the holotype of P. ohioensis as figured by Safford (compare figure 5A' and 5B' with 5E' to 5H') in having a vertically wider cingulum of ridges (about twice or more wider). The narrower cingulum of P. ohioensis was cited by Dalla Vecchia (1988), Hansen (1997) and Brusatte (2007) as diagnostic. However, the reliability of this character was questioned by Ivanov and others (2009) and Lucas and others (2011). Ginter and others (2010) and Hansen (1985) seem to accept only P. ohioensis as the valid taxon but do not explicitly say so. Zidek and Kietzke (1993) consider P. alleghaniensis synonymous with P. ohioensis but do not state why. Ivanov and others (2009) suggested that the width of the lingual band is so variable in Petalodus that it may be unreliable as a diagnostic character and that the differences may be due to position in the jaws. We agree that the utility of this character is questionable given that we can find no functional reason for a rigid dichotomy between narrow and wide bands. In addition, we note that the specimens figured by Newberry (figures 5C' and 5D') show variable cingulum width, with one (figure 5C') approaching that of P. alleghaniensis (figures 5E' to 5H'). We therefore conclude that P. alleghaniensis is synonymous with P. ohioensis to which we refer all of the lower Cutler bed specimens.

CONCLUSIONS
The lower Cutler beds in the Shafer Basin have produced several teeth of the petalodontiform shark Petalodus ohioensis and one of ctenacanthid Cladodus. Most of the teeth are from the Shafer limestone, an informal name applied to the widespread limestone at the top of the lower Cutler beds. The Petalodus teeth are the first documented occurrence of this taxon in Utah, and Cladodus the youngest occurrence in the state.
Among the various Petalodus form species, we recognize P. hastingsii as a valid taxon characterized by a crown with a gently concave vertical cross-section, broad band of ridges on the lingual side and a ridged, outward facing cingulum on the labial side. P. acuminatus as having a low crown compared to width, with a more concave vertical cross-section than in P. hastingsii, and imbricated ridges on the ventrally facing cingulum. P. ohioensis is characterized by tall crowns compared to width, having a more concave vertical cross-section than in P. hastingsii, and imbricated ridges on a ventrally facing cingulum. Both P. acuminatus and P. ohioensis are larger taxa than P. hastingsii. "Petalodus" rhombus is either a Petalodus species characterized by the lack of ridges on the crown, or more likely that it represents a distinct genus.

ACKNOWLEDGMENTS
We are extremely grateful to the help we had with the search for several key specimens of Petalodus, as well as providing photographs of key specimens: Emma Bernard, Natural History Museum, London, England; Christian Klug, Paläontologisches Institut und Museum, Zurich, Switzerland; Andreas Müller, ETH-Bibliothek, Zurich, Switzerland; Thierry Malvesy, Conservator, Muséum d'histoire naturelle de Neuchâtel, Switzerland who located the long lost tooth of Petalodus acuminatus, Alain Germond, Muséum d'histoire naturelle de Neuchâtel, Switzerland provided the photograph of the long lost type; Deborah Hutchinson, Bristol City Museum, Bristol, England; Matt Riley, Sedgewick Museum, Cambridge University, Cambridge, England; Ned Gilmore, Academy of Natural Sciences of Drexel University, Philadelphia, PA. Thanks also to Emma Bernard for email discussions regarding Owen' s holotype, and Thierry Malvesy for discussions regarding the Agassiz specimens at the Muséum d'histoire naturelle de Neuchâtel. The teeth described from the "lower Cutler beds" are from a legacy collection made in the 1970s by Ottinger. Fianlly, review comments by James Kirkland, Grant Willis, and Doug Sprinkel are appreciated, especially in discussions on the proper name of the strata the teeth were collected from.