Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
Review
. 2013 Oct 30;8(10):e78572.
doi: 10.1371/journal.pone.0078572. eCollection 2013.

The articulation of sauropod necks: methodology and mythology

Kent A Stevens  1
Affiliations
Review

The articulation of sauropod necks: methodology and mythology

Kent A Stevens. PLoS One. .

Abstract

Sauropods are often imagined to have held their heads high atop necks that ascended in a sweeping curve that was formed either intrinsically because of the shape of their vertebrae, or behaviorally by lifting the head, or both. Their necks are also popularly depicted in life with poses suggesting avian flexibility. The grounds for such interpretations are examined in terms of vertebral osteology, inferences about missing soft tissues, intervertebral flexibility, and behavior. Osteologically, the pronounced opisthocoely and conformal central and zygapophyseal articular surfaces strongly constrain the reconstruction of the cervical vertebral column. The sauropod cervico-dorsal vertebral column is essentially straight, in contrast to the curvature exhibited in those extant vertebrates that naturally hold their heads above rising necks. Regarding flexibility, extant vertebrates with homologous articular geometries preserve a degree of zygapophyseal overlap at the limits of deflection, a constraint that is further restricted by soft tissues. Sauropod necks, if similarly constrained, were capable of sweeping out large feeding surfaces, yet much less capable of retracting the head to explore the enclosed volume in an avian manner. Behaviorally, modern vertebrates generally assume characteristic neck postures which are close to the intrinsic curvature of the undeflected neck. With the exception of some vertebrates that can retract their heads to balance above their shoulders at rest (e.g., felids, lagomorphs, and some ratites), the undeflected neck generally predicts the default head height at rest and during locomotion.

PubMed Disclaimer

Conflict of interest statement

Competing Interests: The author has declared that no competing interests exist.

Figures

Figure 1
Figure 1. Intrinsic neck curvature starts with the bones.
In (A), cervical vertebrae C4 and C5 of Giraffatitan brancai specimen SI are shown articulated and undeflected, i.e., in osteologically neutral pose (ONP). Their vertebral axes, shown in red, naturally create a slight downward bend in ONP, contributing to the subtle ventral osteologically induced curvature (OIC) likely shared with other sauropod necks cranially (Figure 5). In (B) the giraffe Giraffa camelopardalis, cervical vertebrae C6 and C7 are shown also in ONP, revealing the naturally-ascending slope characteristic of giraffe necks at the base. Note the similarity in their opisthocoelous central articulations compared to the sauropod above. Vertebrae to scale; scale bar equals 10 cm. Giraffatitan photographs courtesy Christopher McGowan; giraffe photographs courtesy Brian Curtice.
Figure 2
Figure 2. The life and death of Euhelopus zdanskyi.
In 1929, Wiman illustrated this sauropod in life with a decidedly giraffe-like pose, rising at a slope of 38° (vertebral axes indicated by the solid red line in A, derived from [4:fig. 3 and pl. 3], see also [32:fig. 9]). In the life reconstruction the base of the neck was given the same curvature as the opisthotonic pose in which the original specimen was found (C). It has subsequently been depicted with a steeper slope (dashed red line in B, from [14]) that even exceeds the death pose in which bone already contacts bone (indicated by the red arrows in D). While the neck has also been regarded as more moderately curved , Euhelopus may in fact have had a straight neck in the cervico-dorsal region in ONP –. Photographs courtesy Valérie Marin-Rolland of the E. zdanskyi specimen PMU 24705, Paleontological Museum of Uppsala University.
Figure 3
Figure 3. Impressive sculpture.
The Giraffatitan brancai mount at the Humboldt Museum of Natural History has been restored with an extraordinarily steep neck at the base, with an ascending neck that appears to be in ONP. While the neural arches in the cervico-dorsal region were not preserved, the centra were, and the sculpture in the mounted skeleton deviates significantly from the actual fossil material (see Figure 4). Photographs by the author.
Figure 4
Figure 4. The iconic swan neck of Giraffatitan brancai.
Janensch (A) illustrated the original fossil material in the cervico-dorsal vertebrae (C10 to D2) as they were found, in articulation, and despite their missing the neural spines, the centra are collinear and appear close to ONP based on their central articulations. Janensch's skeletal reconstruction (B), however, does not reflect this osteology; instead a gracefully-curved swan neck was illustrated, complete with restoring the vertebrae at the base of the neck as if wedge-shaped to formed that elegant rising curve in ONP. The slope of the neck increased further in some later illustrations, e.g., the red curve (C) is drawn from Paul's reconstruction [14; 32:fig. 6]. The centra at the base of the neck are straight, elongated cylinders with parallel anterior and posterior central margins (A) and not wedge-shaped with convergent margins (as inevitably, mis-represented) like those of a giraffe, there is no osteologically-induced bend at the base of the neck. Substituting an ONP reconstruction of the complete vertebral series from C3 to D2 based entirely on Janensch's individual vertebral illustrations (see text) two alternatives are presented (D and E). In D the slope of the anterior column matches that of the original skeletal reconstruction by Janensch , which has relatively high placement of the pectoral girdles upon the ribcage (but lower placement than Paul illustrated, which caused his reconstruction to have a lower vertebral column at the base of the neck). If the scapulocoracoids are reconstructed as closely separated medially and more ventrally placed upon the ribcage, the resultant slope of the anterior dorsals rises necessarily. This raises the head height to 10 m, while the Berlin mount goes to 11, or more. Scale bar is 10 m. The horizontal line represents the ground plane according to revised appendicular reconstructions.
Figure 5
Figure 5. Estimation of sauropod ONP from illustrations.
Composite figures are assembled into approximate ONP for partial or complete presacral columns for various sauropods: A: Apatosaurus louisae , B: Giraffatitan brancai , C: Dicraeosaurus hansemanni , D: Cetiosaurus oxoniensis, E: Euhelopus zdanskyi , F: Diplodocus carnegii , and G: Mamenchisaurus young . Note that some exhibit a slight dorsal OIC cranially, and all are straight caudally. Cetiosaurus illustrations courtesy John Martin. The reconstructions are not to scale, however, the individual vertebrae within a column were adjusted as necessary to the same scale within each vertebral column , .
Figure 6
Figure 6. Revised Skeletal Reconstruction of Apatosaurus louisae.
In the original 1936 reconstruction (A) of Apatosaurus louisae (CM 3018) the pectoral girdles were positioned quite dorsally upon the ribcage, which created a downward slope to the anterior dorsal column at the shoulders, and hence a downward slope at the base of the neck. Reconstruction of the vertebral column from individual illustrations [B] corresponds closely to the skeletal illustration and was used as one check of the dimensional accuracy of a fully-articulated digital model of the specimen CM 3018 (C). All elements modeled individually to scale, based on archival sources plus photographs and personal observation of the original material during its reassembly and remounting at Phil Fraley Studio, and scale orthographic drawings courtesy Philip Platt [pers. comm.]. In articulating and posing the digital model, the orientation and placement of the pectoral girdles and the angulation of the ribs incorporate many current contributions of studies of the articular skeleton, in particular the placement of the pectoral girdles –[32, Phil Platt, pers. comm.].
Figure 7
Figure 7. Diplodocids swept out a huge feeding surface, despite their relative inflexibility.
Apatosaurus (A) and Diplodocus (B) are shown in extreme lateroventral flexion, reaching down and laterally to ground level, and in C and D in extremes of dorsal flexion (including dorsiflexion at the cranio-cervical joint) as if to reach as high as possible (see also overall feeding envelope visualization in Figure 20). Despite the enormous sweep of these necks, the vertebral joints, especially at the base of the neck of Diplodocus (C13 and C14) permit limited flexion prior to disarticulation (see Figure 9). While both necks sweep out a huge surface area, Apatosaurus, with its larger posterior cervical zygapophyses, could reach higher despite having a somewhat shorter neck than Diplodocus.
Figure 8
Figure 8. Sauropod intervertebral separations.
Examples of articulated sauropod cervical columns with condyles deeply inserted into their associated cotyles, leaving intervertebral gaps of only a few centimeters (see arrows). Camarasaurus lentus (DNM 28, A and D) and Barosaurus (CM 11984, B, C, and E). Photographs by the author and J. Michael Parrish.
Figure 9
Figure 9. Sauropod necks did not have avian flexibility.
Cervical vertebra C13 of the ostrich Struthio camelus (A) and C13 of Apatosaurus louisae (B) are scaled to equal vertebral axis length. The heterocoelous central articulation (A) and the opisthocoelous articulation (B), both have geometrically-defined centers of rotation defined by their centers of curvature in the sagittal plane. The ostrich postzygapophyses (red) are both relatively larger and closer to the center of rotation (white arrows) than those of the sauropod. The geometric consequence is that for any value of ZSF applied equally to the ostrich and to the sauropod, the former will have a greater range of motion. C and D show two articulated cervical vertebrae, C13 and C14, near the base of the neck of Diplodocus carnegii (CM 84) in maximum lateroventral flexion to the left (C) and maximum laterodorsal flexion (D), i.e., diagonal extremes of the range of motion. Note that the postzygapophyses (red regions) in C and D barely overlap their associated prezygapophyses (the ZSF is about 0.5). Struthio image courtesy John Martin; Apatosaurus image courtesy Virginia Tidwell. Supplemental material: Movie S1.
Figure 10
Figure 10. Osteological stops.
The posterior cervicals of camel Camelus dromedarius (A) show pronounced depressions (see arrow) where the postzygapophyses make contact just posterior to the associated prezygapophyses with which they articulate. At the limit of travel in dorsiflexion the zygapophyses remain in overlap (contra [80]) and compression forces can be transmitted through the zygapophyses as the neck becomes effectively rigid and stable at the extremes of dorsiflexion. Pronounced osteological stops are also exhibited in many birds, such as the Greater Rhea Rhea americana (B, see arrows). Photographs by the author; rhea specimen at the Zoology Museum, University of Cambridge, access courtesy Matthew Lowe, and the camel vertebrae are at the Condon Museum, University of Oregon.
Figure 11
Figure 11. Bracing at the base of the giraffe's neck.
The base of the giraffe's neck is braced to protect the intervertebral joints from excessive strain on their synovial capsules and to rigidify the neck as it reaches the limits of range of motion. As the neck is raised at the base (A), the postzygapophyses of C7 travel posteriorly until they wedge into depressions in the neural spines of T1 just behind the prezygapophyses (see arrow). Another bracing scheme applies when the neck is deflected laterally (B), In defecting the neck to the left, for example, C7 bears against the left postzygapophysis of T1, see arrow. In either dorsal or lateral flexion the two vertebrae progressively lock up firmly and stably. At these extremes the zygapophyses maintain substantial overlap (roughly a ZSF of roughly 0.5). CT data provided courtesy American Museum of Natural History and Timothy Rowe, University of Texas. Supplemental material: Movie S2, Movie S3.
Figure 12
Figure 12. Giraffe flexibility is predicted by their joint geometry.
The ability of a giraffe to reach vertically and to flex laterally to just reach its flanks is closely replicated by a digital model based on CT scan data of a recent giraffe (see also closeup in Figure 11). The zygapophyses remain in articulation with substantial overlap when they reach osteological stops at the base). CT data provided courtesy American Museum of Natural History and Timothy Rowe, University of Texas. Supplemental material: Movie S4.
Figure 13
Figure 13. ONP for various birds.
The avian neck has a sigmoidal curve that is formed intrinsically by its osteology when the vertebrae are articulated in ONP. The alert resting head height for the ostrich Struthio camelus (top) is higher than predicted by ONP (and the ostrich often further retracts the head during locomotion [98]). Many other birds, however, do assume a pose close to ONP as their characteristic alert resting posture: Cape Penguin Spheniscus demersus (bottom left), Flightless Cormorant Phalacrocorax harrisi (middle), and Kiwi Apteryx australis (bottom right). Note inflection points (arrows). Photographs by the author and John Martin; specimens at the Zoology Museum, University of Cambridge, access courtesy Matthew Lowe.
Figure 14
Figure 14. ONP for various reptiles.
The Nile Monitor Varanus nilotictus (A) and Komodo Dragon Varanus komodoensis (C) have very straight necks in ONP. Head elevation, if any, is primarily through the slope of the anterior dorsals. In contrast, the crocodilians Alligator mississippiensis (B) and Crocodylus acutus (D) have gently rising necks in ONP. The Seychelles tortoise Testudo elephantina (E) has an inflection in curvature; note that its characteristic head elevation arises in ONP. The cryptodiran snapping turtle Chelydra serpentine (F) curves monotonically from a vertical descent caudally to nearly straight cranially. Photographs by the author and John Martin of specimens at the Zoology Museum, University of Cambridge, access courtesy Matthew Lowe.
Figure 15
Figure 15. ONP for various mammals.
Mammals have more or less dorsally-curved necks that tend to raise the head intrinsically. ONP is characteristic of mammals in alert rest and locomotion (an exception is exemplified by the Brown Hare Lepus europaeus (A) which assumes ONP for locomotion and exploratory behavior [97: fig. 17-3] but not in alert rest , . The giraffe Giraffa camelopardalis naturally rises in ONP (B, note the deep insertion of condyles within cotyles consistent with dissections [15]), and assumes approximately this pose in locomotion and alert rest , , , . The nearly straight necks of the Giant Anteater Myrmecophaga tridactyla (C), also mounted in ONP, is characteristic of habitual alert resting pose of alert rest, locomotion pose and feeding. The horse Equus caballus (D) and camel Camelus dromedarius (E) also hold their heads close to ONP in alert rest and locomotion. Note that the cranio-cervical joint is undeflected (arrow) as well as the entire cervical column. Photographs by the author (camel photograph by J. Michael Parrish); hare, anteater and horse specimens at the Zoology Museum, University of Cambridge, access courtesy Matthew Lowe. The camel is at the Field Museum of Natural History. The giraffe is a 3D digital model placed in ONP based on CT data courtesy American Museum of Natural History.
Figure 16
Figure 16. Camarasaurus had a swan neck taphonomically, but not in life.
The 1925 skeletal reconstruction of the juvenile Camarasaurus (A) accurately replicates the curvature of the neck as found (B), but the zygapophyses are illustrated misleadingly as if they were aligned, in ONP, suggesting that the upward curve is intrinsic and ‘built in’. The original specimen, however, is obviously contorted into a dramatic opisthotonic pose, with the zygapophyses disarticulated throughout much of the neck. Red indicates the exposed postzygapophyses (compare with nearly identical opisthotonic pose in the larger specimen USNM 13786-310D, Figure 17). Disregard for this extreme opisthotonic distortion in subsequent skeletal depictions, some portraying the neck comfortably achieving a near vertical pose , has resulted in a nearly universal expectation that Camarasaurus had a natural swan-like curve to the neck. Photograph of Camarasaurus lentus CM 11338, by the author.
Figure 17
Figure 17. Another Camarasaurus lentus in opisthotonic pose.
A partly-prepared block, USNM 13786-310D, reveals a ‘death pose’ with curvature very close to that of the more familiar juvenile specimen CM 11338 (Figure 16). In both specimens the postmortem dorsiflexion disarticulated the zygapophyses such that it was preserved in a pose that was unlikely attainable in life. Red indicates exposed postzygapophyses, and the white line segments indicate the extreme displacement of the zygapophyseal pairs from ONP. Photographs by the author.
Figure 18
Figure 18. This Diplodocus has a false kink in the neck.
The Denver Museum of Nature and Science mount of Diplodocus longus DNMS 1494 has a sharp upward bend that appears intrinsic since the vertebrae in the vicinity of C13–C15 appear undeflected. The curvature, however, is an artifact of the restoration of the fragmentary neural spines, and not exhibited by any other diplodocid specimen including the Carnegie Museum of Natural History Diplodocus carnegii CM 84, a cast of the first 10 cervicals of which were used for the Denver mount. Those cervicals caudal to C10 are heavily restored and induced the misleading suggestion of an upturned neck. Taylor, Wedel and Naish claim that “… computerized studies are not as objective as they may appear, since seemingly Stevens and Parrish could not replicate the flexibility of actual specimens” presuming that the entire neck of DMNS 1494. In fact, the flexibility estimates from would have permitted the head to have reached such heights (see and C). The specimen they refer to (A) has a sculpted bend that is not representative of other, more complete specimens of Diplodocus that emerged straight from the shoulders (D). Photographs by the author, access courtesy Kenneth Carpenter. Supplemental material: Figure S1, Figure S2.
Figure 19
Figure 19. Details of the digital modeling of Apatosaurus louisae.
Archosaur vertebral morphology varies smoothly along the axial skeleton, and the gradual changes from one vertebra to the next is amenable digital modeling by ‘blend shapes’ (see text regarding digital modeling). Through a multi-step process, first deformable generic forms are created for all elements then used to create specific variations on that shep. For example, a generic dorsal rib is constructed, then several specific ribs are modeled to match the corresponding original fossil material, with the remaining intervening elements created by interpolation, and finally each element is painstakingly sculpted and adjusted to capture individualities of the original specimen such as the irregularities in the cervical ribs, compared to the original specimen (Figure 5). The process of creating a digital scale model, like sculpting in a more conventional physical medium, shares the same goals of faithfully replicating the morphology and dimensions of the original. Like physical sculptures, it is a matter of judgment as to when the resemblance is sufficient, and as to what is to be regarded as artifactual, such as an apparent distortion due to preservation. Unlike physical sculptures, these models are readily edited and successively refined, and most importantly, readily articulated without need for a physical armatures. As a visualization tool, digital models greatly facilitate the appreciation of design as the bauplan emerges from the aggregation of the component pieces (note that A. louisae is accompanied by a Camarasaurus lentus, to scale).
Figure 20
Figure 20. Long necks, but not swan necks.
In addition to sweeping out a broad ‘feeding envelope’ (a curved surface of maximum reach [28]), sauropod necks are sometimes expected to be able to pull the head back to reach closer to the animal to explore the volume within this surface, (e.g., [113: fig. 12.1]), somewhat in the manner of a swan (A). While Apatosaurus could place its head at any point across an enormous feeding surface (C), the neck was not able to retract the head back towards the body (B, D). Supplemental material: Movie S5, Movie S6, Movie S7, Movie S8.
Figure 21
Figure 21. Some mammals relax in an alert posture by retracting their heads over their shoulders, but most do not.
The sphinx-like alert resting posture in Panthera leo (A) and Sylvilagus nuttallii (B) is achieved by maximum dorsiflexion at the base of the neck (C7-T1) and maximum ventriflexion at the head to keep the head level, as shown by radiographic studies (95–97). But few mammals can achieve this feat. Most quadrupeds hold their heads cantilevered before the shoulders with the intervertebral joints in a relaxed ONP posture and the weight of the head and neck carried by dorsal ligaments and muscles. The horse, for instance, holds its head high in alert rest (as in Figure 15d), with all joints of the cervical column, including the cranio-cervical joint and the C7-T1 junction undeflected, in ONP. Photos by the author.
Figure 22
Figure 22. Inner ear orientation is consistent with subhorizontal sauropod necks.
The lateral semicircular canal (LSC) is approximately horizontal in alert birds. The orientation α (see text) is plotted for 32 species of birds [111: fig 7a] as a conventional histogram (A) and polar histogram (B) with 5° intervals (c.f. expanded-scale plot in [132: fig. 2]). When a sauropod cranium is similarly oriented (α=+5°), the rostrum slopes downward (by −15° in Camarasaurus lentus and by −37° in Diplodocus longus) , . The LSC also constrains the slope of the neck cranially. The neural canal passing through the atlas-axis is collinear with the foramen magnum, as illustrated by the solid green line in C and the physical armature in the original specimen (D) of Kaatedocus siberi, SMA 0004 – see also the location of the foramen magnum (indicated in green) in the posterior view (E) of Diplodocus . Consequently, with the cranium oriented relative to gravity as indicated by the LSC, and with the cranio-cervical joint undeflected, the anterior neck is roughly horizontal . Taylor et al. , however, misinterpreting the anatomy, suggest “… the foramen magnum and occipital condyle are [both] at a right angle relative to the long axis of the skull …” so that the atlas-axis inserts posteroventrally to the cranium, and consequently they falsely conclude the anterior neck ascends steeply as indicated by the red dashed line in F, from [12: fig. 4]; they figured an even steeper neck for Camarasaurus. But properly interpreted, the anatomy of the occiput, the atlas-axis, and the LSC, together with observations of habitual head orientating in the EPB, supports the interpretation that the necks were habitually subhorizontal cranially in diplodocids (E) and camarasaurids (as depicted in Figure 23) . The digital reconstruction (C, F, and G) is based on data courtesy Andreas Christian and Gordon Dzemski. Photo (D) by the author. Supplemental material: Movie S9. A turntable movie depicting the spinal cord (red) entering the foramen magnum of Kaatedocus siberi.
Figure 23
Figure 23. Resurrection of a juvenile Camarasaurus lentus.
The iconic swan-like ascending neck of Camarasaurus sp. likely derives from the opisthotonic pose of the remarkably complete specimen CM 11338 (upper left). However, when all elements are modeled individually and placed into ONP, the opisthotonic pose in the neck and the axial twist through the dorsal column is removed revealing that this sauropod had a rather short neck that extends straight from the anterior dorsals, which raised the neck with a slight incline (see also [136]). Red indicates elements that were missing in the original specimen. This model was created for the Carnegie Museum of Natural History, with cranial modeling contributed by Scott Ernst, forelimb modeled with reference to digitization data of AMNH 664 and scapula coracoid of CM 11338, both courtesy Ray Wilhite (see below regarding digital modeling). Supplemental material: Movie S10. Animation of Camarasaurus from its death pose into a life pose near ONP.
Figure 24
Figure 24. Digital articulation.
CT data of individual vertebrae of a recent giraffe Giraffa Camelopardalis are articulated in Autodesk Maya . Cervical vertebra C7 pivots about a center of rotation that closely corresponds to the center of curvature of the roughly hemispherical condyle of T1, confirmed by exploratory manipulation and adjustment, resulting in close intervertebral separations as reported in (see red arrows). In A–C, by alternating between opaque and transparent one can observe osteological bracing dorsiflexion (A) and the ZSF at the limit of ventriflexion. With all intervertebral joints adjusted (D–E), the articulated neck approximates the range of motion observed in life (see also Figures 11, 12). This method applies equally to the similarly opisthocoelous vertebrae –, see Figure 25. CT data provided courtesy American Museum of Natural History.
Figure 25
Figure 25. Creating digital, articulated skeletal models.
In A, the cervical vertebrae of A. louisae CM 3018, modeled by subdivision surfaces (see text) are rigged to form a kinematic chain with joints at the centers of curvature of the condyles (displayed in red), with empirically-determined intervertebral separations that maximize the congruence between condyles and cotyles and associated zygaphophyseal pairs at each intervertebral joint. The articulated skeleton resembles the original specimen (B), but fortunately without the rigid steel armature. In C, a digital model of an ostrich Struthio camelus is shown in ONP, based on published data of joint-by-joint intervertebral separations and flexion limits (in both mediolateral and dorsoventral flexion), and in D, an example of its extraordinary flexibility. Supplemental material: Movie S11, Movie S12, File S1.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Reisdorf AG, Wuttke M (2012) Re-evaluating Moodie's Opisthotonic- Posture Hypothesis in Fossil Vertebrates Part I: Reptiles – the taphonomy of the bipedal dinosaurs Compsognathus longipes and Juravenator starki from the Solnhofen Archipelago (Jurassic, Germany). Palaeobiodiversity and Palaeoenvironments 92: 119–168 10.1007/s12549-011-0068-y - DOI
    1. Tornier G (1909) Wie war der Diplodocus carnegii wirklich gebaut? Sitzungsberichte der Gesellschaft Naturforschender Freunde zu Berlin 1909-4: 193–209.
    1. Hay OP (1910) On the manner of locomotion of the dinosaurs, especially Diplodocus, with remarks on the origin of the birds. Proceedings of the Washington Academy of Sciences. 12: 1–25.
    1. Wiman C (1929) Die Kriede-dinosaurier aus Shantung. Palaeontologica Sinica (Series C) 6: 1–67.
    1. Paul G (1998) Differing bipedal and tripodal feeding modes in sauropods. Journal of Vertebrate Paleontology 69: 70A.

Grants and funding

The author has no support or funding to report.

LinkOut - more resources