A new, largely complete eusauropod dinosaur with cranial and postcranial elements from two skeletons, Mierasaurus bobyoungi gen. nov., sp. nov. from the lower Yellow Cat Member (Early Cretaceous) of Utah (USA), is the first recognized member of Turiasauria from North America. Moreover, according to our phylogenetic results, Moabosaurus utahensis from the lower Yellow Cat Member of Utah (USA) is also a member of this clade. This group of non-neosauropod eusauropods, which now includes five genera (Losillasaurus, Turiasaurus, Mierasaurus, Moabosaurus and Zby), was previously known only from the Jurassic of Europe. These recent discoveries in Utah suggest that turiasaurs as a lineage survived the Jurassic-Cretaceous extinction boundary and expanded their known range, at least, into western North America. The revised spatiotemporal distribution of turiasaurs is consistent with the presence of a land connection between North America and Europe sometime during the late Tithonian to Valanginian (c.147-133 Ma). Mierasaurus and Moabosaurus are the only non-neosauropod eusauropods known from North America, despite being younger than the classic neosauropods of the Morrison Formation (c.150 Ma).
Sauropod dinosaurs were the largest vertebrates ever to walk the Earth1,2 and were a diverse and successful group3. Although they had achieved a global distribution by the Middle Jurassic and maintained their dominance until the end of the Cretaceous3,4, their evolutionary history was not without its crises. In particular, they suffered a 60-80% extinction at the end of the Jurassic3,4,5,6,7, which was part of a much wider extinction event affecting many marine and terrestrial groups7,8. This apparently resulted in the extinction of most of the non-neosauropod eusauropods that had dominated Jurassic faunas, and their replacement by neosauropods in the Cretaceous3,4,7. However, recent work has suggested that lower diversity at this time was not the result of a true mass extinction occurring at the Jurassic-Cretaceous boundary itself; rather, it represents a sequence of turnover events that lasted well into the Early Cretaceous6,7. The non-neosauropod eusauropod lineage Turiasauria, previously only known from the Late Jurassic of Europe1,9,10 and possibly the Middle Jurassic of Africa11, was therefore a potential victim of an extinction at the Jurassic-Cretaceous boundary4,5,7. However, here we report the unexpected occurrence of a new turiasaur from North America, dating to the late Berriasian-early Aptian (c.142-124 Ma) a time interval in which sauropod remains are rare globally5,12. This discovery supports a more complex pattern of turnover among sauropods during the Jurassic-Cretaceous transition and is consistent with previously proposed palaeobiogeographic and palaeogeographic connections between Europe and North America after the Tithonian13,14,15,16. Thus, during the late Tithonian-early Aptian (c.147-124 Ma), non-neosauropod turiasaurs were unexpectedly present, at least, in the Utah region of North America. Importantly, this relatively rare group of sauropods are represented in North America by a new taxon described here as Mierasaurus bobyoungi gen. nov, sp. nov. (Figs 1–5), which is closely related to the Kimmeridgian-Tithonian turiasaurs (Losillasaurus, Turiasaurus and Zby) of Europe1,9,10 and the recently published Early Cretaceous taxon (from the lower Yellow Cat Member) Moabosaurus utahensis 17, after the revised phylogenetic placement of the latter proposed in this work (Fig. 6 and Supplementary Figs S4 and S5).
Dinosauria Owen, 1842
Saurischia Seeley, 1887
Sauropoda Marsh, 1878
Eusauropoda Upchurch, 1995
Turiasauria Royo-Torres, Cobos and Alcalá, 2006
Mierasaurus bobyoungi gen. nov., sp. nov.
Genus named for Bernardo de Miera y Pacheco, Spanish cartographer and chief scientist for the 1776 Domínguez-Escalante Expedition: the first European scientist to enter what is now Utah. The species name acknowledges the importance of the underappreciated research by Robert Young on the Early Cretaceous of Utah18.
A partial skeleton of a single individual (UMNH.VP.26004), comprising disarticulated cranial and postcranial elements from the type site (Doelling’s Bowl). We regard this as subadult because it is a relatively large animal with unfused vertebral centra and neural arches in some dorsal vertebrae. This individual includes a partial skull and jaw, teeth, atlas, 8 cervical vertebrae, 11 cervical ribs, 11 dorsal vertebrae, 6 dorsal ribs, 6 sacral ribs, 15 caudal vertebrae, two chevrons, right scapula and partial left scapula, left radius, left ulna, left manus, complete pelvic elements, both femora, left tibia, left fibula, left astragalus and left pes (Figs 1–5; Supplementary Tables S1 and S2 and Supplementary Fig. S6).
Other disarticulated sauropod specimens from the type locality include a possible juvenile dentary (UMNH.VP.26010) and juvenile femur (UMNH.VP.26011) (Fig. 1).
Type locality and horizon
All Mierasaurus remains discussed herein are from Doelling’s Bowl bonebed (Fig. 1 and Supplementary Figs S1-S3), UMNH VP.LOC.1208 (Utah Loc. 42Gr0300v) within the lower Yellow Cat Member (below the marker calcrete), Cedar Mountain Formation, on lands managed by the U.S. Bureau of Land Management in northern Grand County, east-central Utah. In a preliminary report in a conference abstract, detrital zircon dating indicates maximum ages ranging from ~136.4 ± 1.1 Ma and ~132 Ma for lower Yellow Cat Member19 and ~137.2 ± 2.0 Ma for upper Yellow Cat Member19 (Supplementary Figs S1-S3). The combination of these data with the upper Berriasian-Valanginian age based on ostracods20 and charophyte fauna21 for the upper Yellow Cat Member indicates a conflict in the maximum age traditionally considered for it 124.2 ± 2.6 Ma17,22 (see Supplementary Information). All of these data suggest a potential age of late Berriasian-early Aptian (c.142-124 Ma) for the Yellow Cat Member.
Doelling’s Bowl has produced the iguanodont cf. Iguanocolossus sp., a new species of polacanthid ankylosaur, a large allosauroid theropod (teeth), and the dromaeosaur Yurgovuchia doellingi 23,24,25. Exact locality information will be provided to qualified researchers on request through the Natural History Museum of Utah or the Utah Geological Survey.
A turiasaurian sauropod possessing the following features (autapomorphies marked by *): *the otosphenoidal ridge extends from the anterior surface of the paroccipital process, near its ventral margin and is restricted to the medialmost part of the latter process (Fig. 2b); *the occipital condyle has a pair of rounded ridges extending dorsoventrally, one on either lateral face of the condylar articular surface (Fig. 2a–d); *the atlantal intercentrum (Fig. 3e,f) bears a pair of depressions in the medial surface, facing posteromedially, each of which receives the anterolateral margin of the odontoid process; a well-developed spinoprezygapophyseal lamina extends anteriorly (Fig. 3a–d, g–m) as a low ridge onto the lateral surface of the prezygapophyses roofing a lateral fossa on prezygapophyses in middle and posterior cervical vertebrae (shared with the diplodocine Kaatedocus 26); *cervical ribs bear a ridge or bulge on the lateral surface of the tuberculum, immediately posterior to the base of the anterior process (Fig. 3n–v); dorsal neural arches lack posterior centroparapophyseal laminae; *lateral depression on the distal ramus of haemal arches (Fig. 5g,h); metacarpal I longer than metacarpal IV, shared with Macronaria27 (Fig. 5m); a very short ischium compared to pubis length (ischium:pubis length ratio = 0.75) (Fig. 5d–f); the midpoint of the fourth trochanter placed in the proximal part of the femur (Fig. 5a); femur with subequal distal condyles (Fig. 5k); and pedal unguals 2 and 3 compressed dorsoventrally (Fig. 5l,o).
The braincase of Mierasaurus (Fig. 2a–d) resembles those of non-titanosauriform and non-diplodocoid neosauropods3,9,27. The frontal forms part of the anterior margin of the supratemporal fenestra (the plesiomorphic state seen in many non-sauropod sauropodomorphs, Shunosaurus and probably in Turiasaurus), whereas most frontals of other sauropods exhibit the derived condition by not participating in this fenestra9. The parietal dorsal surfaces are inclined such that they face slightly posteriorly, the supraoccipital faces posterodorsally, and the occipital condyle as a whole projects posteroventrally. The skull roof is essentially flat without domes or sulci. The supratemporal fenestra is large and transversely wide. The parietal extends laterally and twists to produce a vertically tall plate forming the posterior margin of the supratemporal fenestra. The squamosal is more slender and ventrally curved, and less flared and dorsally curved, than in Camarasaurus 28. The nuchal crest on the supraoccipital is very low. A pair of transversely widened proatlantal facets is present on the occiput, dorsolateral to the foramen magnum. The paroccipital process is anteroposteriorly compressed and directed laterally and moderately ventrally as in non-neosauropods3. The occipital condyle bears a strongly convex articular surface with two rounded ridges laterally, differing from Turiasaurus where the occipital condyle has small dimples and convexities9. A central, ventrally projecting area links the neck of the occipital condyle to the region of the basal tubera, with moderately deep grooves on either side below the otosphenoidal ridge. There appear to have been ridges extending along the lateral margins of the central, ventral projecting portions of the tubera. A clear pit is present in the basisphenoid fossa between and anterior to the basal tubera. The latter are relatively widely separated, and there is no clear anteroposteriorly thin wall linking their medial margins and partitioning the basisphenoid fossa from the basioccipital fossa (such a thin partition is present in Moabosaurus 17). The basal tubera are moderately compressed anteroposteriorly, similar to those in Turiasaurus 9, and are not sheet-like as occurs in titanosaurs and some rebbachisaurs3,26,29. The anterolateral portion of the tip of each basal tuber projects anteriorly, giving the free distal surface an unusual L-shaped profile in ventral view, a character state shared with the turiasaur Moabosaurus 17. The quadrate is curved such that its proximal head projects posteriorly and there is a well-developed deep posterior fossa. The dentary (based on two left elements of different size) increases in dorsoventral height towards the symphysis (Fig. 2s), thereby distinguishing it from Jobaria 30 but likening it to Camarasaurus 28 and an unpublished new turiasaur specimen (RD-28) from Riodeva (Teruel, Spain). As in flagellicaudatans, and to a lesser extent Camarasaurus, the ventral margin of the dentary forms a sharp, chin-like transverse ridge in lateral view27. There are approximately 13 alveoli in the dentary. The teeth are spatulate in labial view, and D-shaped in horizontal cross section. Each crown possesses a bulge on its labial surface between the apicobasally-directed mesial and distal grooves. There are no denticles (a possible synapomorphy uniting Turiasauria and Neosauropoda)9,27,31. The most posterior teeth possess the diagnostic heart-shape (Fig. 2v,w) seen in European turiasaurs1,9,31. The mesial-most teeth of Mierasaurus differ from those of Camarasaurus in being more pointed (Fig. 2o-p); those of the latter genus also have shorter apices and a basal bulge on the lingual face28,31. This kind of broad crown morphology is common in non-neosauropods, and is very different from the narrow-toothed condition seen in most neosauropods such as diplodocoids and titanosauriforms3.
The atlas is well preserved (Fig. 3e,f). The intercentrum is anteroposteriorly short with a wide, concave anterior face and convex posterior face. The posterolateral surface of the odontoid process fits into two posterolaterally facing depressions on the atlas intercentrum, unlike Camarasaurus 28, Diplodocus 32 or Moabosaurus 17. The atlantal neural arch bears a pair of medially directed projections that partially separate the neural canal and neurapophyses (above) from the notch in the dorsal surface of the centrum that receives the odontoid process (below). The centrum lacks the robust parapophysis described for the atlas of Moabosaurus 17.
The internal tissue structure of the presacral vertebrae and thoracic ribs is solid, lacking camerae, camellae or other pneumatic cavities: this suggests that Mierasaurus is not a titanosauriform sauropod27,33,34. The cervical vertebra neurapophyses are long curved plate-like elements that closely approach each other without actually meeting on the midline above the neural canal. The cervical centra are strongly opisthocoelous (Fig. 3a-d) and have lateral pneumatic openings that are sometimes divided by a septum. The postzygodiapophyseal laminae of the middle and posterior cervicals extend far laterally along the posterior margins of the diapophyses, unlike most other sauropod taxa where these laminae fade out closer to the neural spine (Fig. 3a-d). All preserved cervicals possess pre-epipophyses35 and post-epipophyses as in taxa such as Euhelopus, Jobaria, Turiasaurus and Moabosaurus 17. Two character states in the cervical neural arch are regarded as local autapomorphies: first, a well-developed spinoprezygapophyseal lamina (SPRL) extends anteriorly onto the lateral surface of the prezygapophysis, and second, a lateral fossa lies below the SPRL. These character states are shared with some diplodocids (e.g. Kaatedocus)26. The cervical neural spines are single or have a very shallow bifurcation (Fig. 3k), differing from Turiasaurus which has more deeply bifid spines1. The cervical ribs are simple, differing from Turiasaurus and Moabosaurus 17 wherein the ribs have an accessory process projecting posterodorsally from the dorsal margin of the distal shafts1. The cervical ribs of Mierasaurus bear a bulge on the lateral surface of the tuberculum, immediately posterior to the base of the anterior process (a potential autapomorphy).
The dorsal centra of Mierasaurus lack complex lateral pneumatic openings and are amphicoelous in posterior elements of the series, differentiating Mierasaurus from Moabosaurus 17, Camarasaurus and titanosauriforms wherein these vertebrae are strongly opisthocoelous3,27. Middle and posterior dorsal neural spines have a posterior spinodiapophyseal lamina (post.SPDL) (Fig. 4h-l) that bifurcates dorsally such that a thinner posterior branch extends to the aliform process as usual, and a stouter anterior branch extends anterodorsally to the posterior surface of the anterior spinodiapophyseal lamina (ant.SPDL) (Fig. 4h-l): this additional anterior branch thus divides the spinodiapophyseal fossa (SDF) (Fig. 4h-l) into upper and lower portions, with the upper portion forming a dorsally opened fossa. The sacrum is only represented by sacral ribs formed from thin plates of bone with an expanded sacral articulation. Five different sacral rib morphotypes are present, suggesting that there were at least five vertebrae in the sacrum of Mierasaurus. The orientation of the dorsally-directed plate of bone on each sacral rib, and the posterior extension of the sacral articular area, demarcates a fossa on the ventromedial part of the dorsal plate, situated between the latter and the sacral articulation.
The anterior caudal vertebral centra are strongly procoelous, differing from the amphicoelous caudals of Haplocanthosaurus, Camarasaurus and many other sauropods, and are more similar to those of the non-neosauropods Mamenchisaurus, Turiasaurus, Losillasaurus 1,3,27, Moabosaurus 17 and most titanosaurs3,27. The caudal vertebral centra of Mierasaurus lack a ventral fossa and lateral pneumatic openings, differentiating them from those of diplodocids26,29. Caudal neural arches are situated on the anterior half of each centrum, (a derived state characterising most titanosauriforms36) with the neural spines directed posterodorsally. The anterior haemal arches are generally unbridged proximally, a derived state that occurs in Shunosaurus, Haplocanthosaurus, rebbachisaurids and macronarians3,27 although one element is nearly closed (Fig. 5g,h). The length of the haemal canal versus chevron total length is 41%, a derived condition shared with Titanosauriformes34,37.
The scapula has a rounded proximal plate with smooth acromial ridge and a long, fairly straight blade. The latter is slightly expanded distally (Fig. 5n). The sternal plate has a straight lateral margin and convex medial one, and is longer anteroposteriorly than transversely. The radius and ulna were preserved in articulation (Fig. 5j). These elements are straight and have transversely expanded distal ends. On the ulna, beside the fossa for the radius, there is a second deep proximolateral fossa on the lateral surface, similar to the derived condition seen in Turiasaurus, Losillasaurus (R.R.T. personal observation) and Dystrophaeus 38. The metacarpals are closely appressed to each other (Fig. 5m). Metacarpal I has two distal condyles, unlike Giraffatitan 33, and it is longer than metacarpal IV (as in many macronarians27).
The preacetabular process of the ilium has a subtriangular outline in lateral view (Fig. 5e), which is the plesiomorphic state that occurs in non-titanosauriforms36. As in neosauropods, this process projects anterolaterally36. The pubis lacks a prominent ambiens process. The ischia are short (Fig. 5d), with an ischium:pubis length ratio of 0.75, which has otherwise only been observed in titanosaurs36. The short ramus and small acetabular margin of the Mierasaurus ischium are also reminiscent of those of titanosaurs such as Rapetosaurus 37. The femur has a straight shaft (Fig. 5a) and its proximal head projects medially. The fourth trochanter is represented by a ridge and a notch on the posteromedial margin, located very proximally (i.e., the most ventral point of the fourth trochanter lies at 40% of femur length from the proximal end, whereas in other sauropods it is usually located just above midlength3). The distal femoral condyles are subequal in size (Fig. 5k). The absence of a lateral femoral bulge supports the placement of Mierasaurus outside Titanosauriformes3,27,36. The tibia has an anteroposteriorly elliptical proximal head and the cnemial crest projects anterolaterally. The posterior rim of the ascending process of the astragalus terminates a short way anterior to the posterior margin of its main body, which is the plesiomorphic state that occurs in non-neosauropod eusauropods3,27. The medial part of the astragalus forms a distinct subtriangular process in dorsal or ventral view, similar to the astragalus of Turiasaurus. This means that the medial one-third of the astragalus narrows in anteroposterior width such that it is about one-third of the width of the middle and lateral portion. The pedal phalangeal formula for Mierasaurus is 2-3-3-2-?0. Pedal unguals on digits II and III are unusual because they are compressed dorsoventrally; such a morphology has been observed in a few other sauropods, e.g. Vulcanodon 39 and possibly Sanpasaurus 40, but in most forms, these unguals are laterally compressed41.
Phylogenetic results and comparison
Phylogenetic analysis (Fig. 6 and Supplementary Figs S4-S5) places both Mierasaurus and the previously established taxon Moabosaurus within the Turiasauria (a clade outside of Neosauropoda) and indicates that, these represent the only known non-neosauropods to have survived into the Cretaceous and the only turiasaurs known from North America. The results support the monophyly of Turiasauria that includes Losillasaurus, Turiasaurus and Zby from the Late Jurassic of the Iberian Peninsula1,9,10 and Mierasaurus and Moabosaurus from the Early Cretaceous of Utah, USA. The phylogenetic analyses identify several synapomorphies of Turiasauria in Mierasaurus (see Supplementary Information for a full list). This view is strengthened by other morphological features that have not yet been incorporated into phylogenetic data sets. For example, a deep proximolateral fossa in the ulna is present in Turiasaurus, Losillasaurus, Zby, Moabosaurus (R.R.T. personal observation in 2016) and Mierasaurus. Turiasaurus and Mierasaurus both possess a pair of small pits or foramina on the supraoccipital, one on either side of the dorsal end of the nuchal crest9, and the medial one-third of the astragalus narrows in anteroposterior width to form a distinct subtriangular process. However, Mierasaurus can be clearly distinguished from other turiasaurs via a suite of autapomorphies, as well as differences in other character states. For example, none of the cervical ribs in Mierasaurus are bifid, differing from Moabosaurus and Turiasaurus; the anterior cervical ribs of Mierasaurus have a small dorsal process projecting from the main distal shaft (i.e., BBGI-5D), potentially representing the incipient phase of development of bifid cervical ribs or the loss of this character, seen in some anterior ribs of Turiasaurus riodevensis paratype1 and in anterior and middle cervical ribs of Moabosaurus 17.
A comparison with other sauropods from the Early Cretaceous of the USA suggests that Mierasaurus is a distinct taxon. For example, Mierasaurus lacks the pneumatic openings in the proximal parts of the thoracic ribs that occur in Astrodon from the Arundel Formation of Maryland42. The anterior middle caudal neural spines of Cedarosaurus and Venenosaurus (Cedar Mountain Fm, Utah) are deflected anteriorly43,44,45 whereas in Mierasaurus they are deflected posteriorly. Sauroposeidon differs from Mierasaurus because the latter has shorter cervical vertebrae46,47. Compared to Abydosaurus and Sonorasaurus, Mierasaurus differs in the absence of camellate pneumaticity in its cervical and dorsal vertebrae45,48. Brontomerus has a longer and more rounded preacetabular process in the illium (a derived state characteristic of Titanosauriformes47,49), whereas Mierasaurus retains the plesiomorphic shorter and sub-triangular process. In addition, a detailed comparison of Mierasaurus and Moabosaurus 17, reveals numerous differences in the skull and postcranial skeleton: 1,) Moabosaurus lacks an occipital condyle with a pair of rounded ridges which is a diagnostic character for Mierasaurus; 2,) the tooth crowns in lingual view have vertical ridges in Moabosaurus that are not developed in Mierasaurus; 3,) the atlas of Moabosaurus has a robust parapophysis that is not present in Mierasaurus; 4,) the atlas intercentrum of Moabosaurus lacks the two posterolaterally facing depressions seen in Mierasaurus; 5,) Mierasaurus lacks bifid cervical ribs; 6,) Mierasaurus lacks the ventral hollows and keel on the ventral surface of the cervical centra that occur in Moabosaurus 17; 7,) the development of the CPRLs on anterior cervical vertebrae is different - convex in Moabosaurus and straight in Mierasaurus; 8,) The cervical ribs of Moabosaurus lack ridges or bulges on the lateral surface of the tuberculum, immediately posterior to the base of the anterior process; 9,) the PRSL is present in the middle and posterior dorsal vertebrae of Moabosaurus and absent in all dorsals of Mierasaurus; 10,) the morphology of the hyposphene differs in posterior dorsal vertebrae, rectangular in Moabosaurus, and triangular in Mierasaurus; 11,) all dorsal vertebrae of Moabosaurus are opisthocoelous whereas they are amphicoelous in the posterior dorsal vertebrae of Mierasaurus; and 12,) Moabosaurus has a slight lateral bulge on the proximolateral femur, which is completely absent in Mierasaurus. The other diagnostic characters of Mierasaurus can not be compared with Moabosaurus at present because of missing data, damage and/or lack of anatomical overlap.
These two North American Early Cretaceous turiasaurs share two synapomorphies that are not observed in their European relatives: 1,) a distinct, small, anteriorly projecting bump is present on the anterolateral part of each basal tuber, just above the distal tip, giving the tips of the tubera a fat L-shape in ventral view; and 2,) middle and posterior dorsal neural spines bear a posterior spinodiapophyseal lamina that bifurcates towards the summit, sending a posterior branch to the aliform process and an anterior branch to the anterodorsal margin of the spine. Therefore, Mierasaurus and Moabosaurus appear to be closely related but distinct, members of the Turiasauria.
There is some evidence that populations of sauropods undertook seasonal migrations of several hundred kilometers in order to locate appropriate environments and access food and water resources over a wide region50, and migrations allowed dinosaurs to explore new geographic areas51. In this way, it seems likely that sauropod groups were able to disperse widely provided that suitable land routes were available52.
The identification of Mierasaurus and Moabosaurus as turiasaurs has palaeobiogeographic implications. During the Late Jurassic, neosauropods were very abundant in North America: the Morrison Formation has yielded more than 430 specimens53. Virtually all of these belong to Diplodocoidea or Macronaria3,27. Even Late Jurassic North American taxa with uncertain affinities, such as Haplocanthosaurus, have been placed within Neosauropoda in most previous analyses3,27,28 and by the current study, so no confirmed non-neosauropod sauropods have ever been described in the Morrison Formation3,27,53. The development of a terrestrial dispersal route between Europe and North America after the late Tithonian could explain the absence of turiasaurs in the rich sauropod faunas of the Late Jurassic (Morrison Formation) of North America, and their presence during the Early Cretaceous (Yellow Cat Member of the Cedar Mountain Formation) in North America. The Jurassic/Early Cretaceous fossil record of eastern North America (Appalachia) has produced currently no evidence of turiasaurs: in fact, sauropod taxa from the Early Cretaceous of the eastern North America are very different, as explained above. Post -Tithonian-pre-Aptian sauropod faunas from North America have only been described from the Lakota Formation (Berriasian-Barremian) on the basis of a left humerus and the first metacarpal I (FMNH PR 487), both identified as Neosauropoda indet54. Thus, the pre-Aptian Cretaceous fossil record for sauropods is particularly poor47,54. This emphasises the value of the insights which Mierasaurus provides into Early Cretaceous North American sauropod faunas. Moreover, this poor fossil record makes it impossible to determine whether the sudden appearance of turiasaurs in Utah in the Early Cretaceous involved a dispersal from an Appalachian fauna during the Earliest Cretaceous or even Late Jurassic, though any such dispersal from Iberia is likely to have passed through this region. Future fieldwork must test the presence or absence of turiasaurs in eastern North America for the Late Jurassic and Early Cretaceous in terrestrial sediments.
During the early Mesozoic, North America and Europe were connected to form the western and central part of the supercontinent Laurasia. With the opening of the North Atlantic during the Late Jurassic and earliest Cretaceous, the region fragmented into smaller tectonic units55,56,57,58 that might have periodically formed contiguous land masses57,58 during sea level low-stands, characterized by exposure of continental margins and draining of continental interior seaways55,56,57,58. The presence of a land bridge between North America and Europe during the late Kimmeridgian-earliest Tithonian13 might explain similarities in some dinosaurs such as Stegosaurus 59 and Allosaurus 60. However, the absence of turiasaurs in the Morrison Formation suggests that this clade dispersed into western North America after the mid-Tithonian. Thus, unless we wish to argue that turiasaurs were actually present in the Morrison Formation but have not been found yet, a second later land connection between Europe and North America is required. Specifically, during the mid-Valanginian (c.137 Ma), a dramatic drop in sea level7,55,56 potentially reconnected Europe and North America and so permitted a final exchange of vertebrate clades across the Atlantic. This Valanginian event could also explain apparent faunal exchanges between these continents proposed for the late Hauterivian-early Barremian13. Taxa that seem to disperse in the Hauterivian/Barremian might actually have done so during the Valanginian lowstand, but gaps in the fossil record mean that the results of these dispersals might not be detected until we access better sampled post-Valanginian deposits. Aside from Mierasaurus, this connection is supported by the discovery of other groups in Utah, for example, a new haramiyidan mammaliaform (in a preliminary report in a conference abstract)61, a clade now documented in North Africa and Europe during the Jurassic–Cretaceous transition.
Thanks to recent discoveries, there has been a rapid growth in our knowledge of the turiasaur sauropods, both in terms of their evolution (e.g., the existence of a group of gigantic sauropods in Europe) and paleobiogeography (e.g., turiasaurs dispersing from Europe to North America in the Early Cretaceous). Many Early and Middle Jurassic sauropod faunas were dominated by various non-neosauropod eusauropod clades (broad-tooth-crown clades) and, in the Late Jurassic, the neosauropod diplodocoids and macronarians. At, or close to, the Jurassic-Cretaceous boundary, sauropods underwent a major extinction4,5,6,7. Cretaceous sauropod faunas were different from those of the Jurassic, the former being dominated by narrow-tooth-crowned clades (diplodocoids and titanosauriforms) from the late Early Cretaceous onwards5,48. In North America, the diplodocoids, so abundant and diverse in the Morrison Formation, disappeared at the Jurassic-Cretaceous boundary and the Cedar Mountain Formation is dominated by “brachiosaurid” titanosauriforms and camarasauromorphs that might have been descended from Morrison taxa. Turiasaurs, therefore, represent an interesting anomaly, demonstrating that a clade of broad-toothed non-neosauropods persisted beyond the Jurassic-Cretaceous extinction and found at least temporary refuge in North America.
The Utah Geological Survey has been excavating the Doelling’s Bowl bonebed (DBBB) in east-central Utah annually since 2005 (Supplementary Figs S1-S3). The DBBB is a multitaxic bonebed that crops out in an area of low relief and is estimated to cover an area of over 5000 m2. Over 1500 individual vertebrate bones have been excavated, mapped (Fig. 1), and collected to date, from several excavation areas totaling 140 m2. The bones are found in a green-gray sandy mudstone with traces of silcrete and sparse root casts along with abundant chert pebbles. In 2010, a partially articulated sauropod dinosaur was discovered eroding out of an arroyo in an area of the bonebed dubbed Gary’s Island. The articulated manus and pes extended through the sediment at an angle below the level of the majority of the skeleton, indicating that an individual sauropod became mired in soft sediment and died in place. After decomposition, many of the elements of the skeleton were scattered over an area of roughly 10 m2. Several shed theropod teeth have been found in association with the sauropod skeleton but no evidence of tooth marks has yet been identified on any of the bones from this part of the quarry. There is minor plastic deformation of many of the bones that, along with sedimentological evidence, suggest that the area was waterlogged. The undersides of the recovered bones show more degradation than the upper surfaces, possibly as a result of grazing by invertebrates. The material has been cleaned and prepared in the laboratory of the Utah Geological Survey. Exact locality information will be provided to qualified researchers on request through the Natural History Museum of Utah or the Utah Geological Survey.
We conducted phylogenetic analyses using two different data matrices62,63 utilizing TNT 1.164 in order to find the most parsimonious trees (MPTs). For the pruned Carballido and Sander matrix62, we implemented a heuristic tree search in TNT. This performed 1000 replications of Wagner trees (using random addition sequences) followed by tree bisection reconnection as the swapping algorithm, saving 10 trees per replicate (Supplementary Fig. S4). For the Mannion et al. analysis63, the pruned data matrix was analysed using the Stabilize Consensus option in the New Technology Search in TNT 1.164. Searches were carried out using sectorial searches, drift, and tree fusing, with the consensus stabilized five times, prior to using the resultant trees as the starting trees for a traditional Search using Tree Bisection-Reconnection (Supplementary Fig. S5). In order to test the support of the phylogenies, the Bremer support and the bootstrap values were obtained (absolute frequencies based on 10,000 replicates).
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
All excavations were conducted under BLM permit UT06-35E-GR. Funding and assistance in the field was provided by UGS, Utah Friends of Paleontology, and BLM. S. Madsen, G. Hunt, N. Toth, J. Cavin, R. Hunt-Foster, and D. Brommel, played critical roles in the field work and in the preparation of the fossils. We are grateful to Dr. B. B. Britt and Dr. R. D. Scheetz for giving access to the material of Moabosaurus in the Museum of paleontology of Brigham Young University. We acknowledge the Willi Hennig Society who support the distribution and use of the phylogenetics package TNT. This research was funded by the Utah Geological Survey, Spanish Ministerio de Economía y Competitividad (CGL2013-41295-P Project DINOTUR), the Departamento de Innovación, Investigación y Universidad, the Fondo Social Europeo (FOCONTUR, Ref. E62), the Departamento de Educación, Cultura y Deporte del Gobierno de Aragón, the Instituto Aragonés de Fomento, the Instituto de Estudios Turolenses. This Project is also part of the IGCP 632 (Continental Crises of the Jurassic: Major Extinction Events and Environmental Changes within Lacustrine Ecosystems). We thank the University of Utah Health Care, and the Hospital Obispo Polanco de Teruel regarding CT-scanning of specimens and related technical assistance. Technical reviews by M. Hayden, G. Willis, and K. Harty are appreciated.