Evidence of disease and traumatic injury is commonly observed in the vertebrate fossil record (see Tanke and Rothschild, 2002). In fact, the presence of paleopathologies has lent insights into the behavior (Motani et al., 1999; Peterson et al., 2009), biomechanics and locomotion (Mulder, 2001; Avanzini et al., 2008), and evolution of the disease (Hanna, 2002; Wolff et al., 2009); specimens that demonstrate developmental malformations, however, are much more rare. The metameric vertebral column and associated costal elements are ideal entities for observing these morphological deviations. Although the vertebral column is the defining character of the vertebrates, we still lack a comprehensive understanding of the mechanisms (genetic and cellular interactions) that control early development (i.e., size, patterning, and morphology; Christ et al., 2000; Morin-Kensicki et al., 2002; Fleming et al., 2004).
Studies of embryonic development often employ the use of genetic knockouts or alter gene expression in order to observe developmental anomalies that arise from the over- or under expression of various signaling factors (i.e., Shh, noggin, Pax, or Sox families; Stockdale et al., 2000; Christ and Scaal, 2008). The abnormal development of these manipulated embryos provides a means to interpret developmental pathways. Although similar experimental manipulations are not feasible with extinct organisms, inferences can be drawn from affected bones exhibiting developmental or pathological anomalies.Congenital malformations within the axial skeletons of dinosaurs are rarely observed in the fossil record. However, some developmental abnormalities such as hemivertebrae (Lydekker, 1889; Witzmann, 2007; Witzmann et al., 2008) and congenital block vertebrae (Dingus, 1996; Rothschild, 1997) have been reported in ornithopods and theropods. Although paleopathologies associated with trauma or physiological stresses have been noted in the mid-caudals of diplodocoids (Osborn, 1899; Hatcher, 1901; Rothschild and Berman, 1991), no occurrences of congenital malformations have previously been reported from within sauropoda.
MATERIALS AND METHODS
Birkemeier and Swor (2007) briefly described an incomplete Apatosaurus (cf. A. ajax) specimen from the Laura's Apatosaur (LA) quarry at the Wyoming Dinosaur Center (WDC), Warm Springs Ranch, Hot Springs Co, WY. An undescribed mid-anterior caudal vertebra (WDC LA-188) from this quarry exhibits a unique morphology; WDC LA-188 is an abnormal mid-anterior caudal vertebra. To avoid confusion, the singular “vertebra” will be used when discussing this specimen as it is functionally, if not developmentally, one vertebra despite the anomalous neural and haemal arches. WDC LA-188 was prepared from the surrounding matrix by mechanical means; some matrix remained on the specimen such as within the neural canal and haemal arch.
Taphonomy and Depositional Setting
The LA quarry is located on the Warm Springs Ranch, Thermopolis, WY, within the Late Jurassic Morrison Formation. Vertebrate quarries on the Warm Springs Ranch have yielded numerous skeletal remains of sauropod dinosaurs (Bedell and Trexler, 2005; Upchurch et al., 2005; Ikejiri et al., 2006). LA quarry is stratigraphically higher than (<5 m) and lateral to (ca. 250 m) the well-documented Something Interesting (SI) quarry (see Jennings and Hasiotis, 2006; Jennings et al., 2011). It is slightly above (1 m) and lateral (35 m) to the Foot Site (FS) quarry from which an isolated partially articulated juvenile diplodocus was described (Bedell and Trexler, 2005).
The vertebra WDC LA-188 was found in a single-taxon quarry of limited lateral extent; further excavation is currently hindered by outcrop conditions. The specimen is represented by disarticulated but associated material including cranial elements, cervical, dorsal and caudal vertebrae, as well as fore and hind appendicular elements, and pelvic material (Fig. 1). The bone condition expresses little pre-depositional breakage, weathering, or rounding due to transport or exposure (e.g., Stages I–II sensu Behrensmeyer, 1978). In addition to vertebrate remains, large sections (1–2 m) of coalified logs and chonchostraca were also observed in the quarry.
Sedimentary structures in the very fine sandy- to silty-mudstone include millimeter scale planar laminated layers, small-scale asymmetric and symmetric ripples, and minor bioturbation in upper layers of the quarry. The depositional environment is thought to be marginal lacustrine. This is consistent with other environmental interpretations, which suggest rising and falling lake levels resulting in alternating lacustrine limestone and fine siliciclastics deposition (Jennings et al., 2011). WDC LA-188 was preserved in a paleoenvironment that facilitated the precipitation of barite (BaSO4) in the presence of decaying organic matter (Jennings and Hasiotis, 2006). Many elements excavated from the LA quarry were encrusted with concretionary barite and carbonate; barite nodules are only found adhering to or within a few centimeters of bone. This is similar to the barite-rich concretions and depositional conditions observed in the nearby SI quarry.
Computed tomography (CT) provides a nondestructive means of observing internal features and aids visualization of external characteristics of fossils that are too delicate to mechanically prepare (Sutton, 2008; Manning et al., 2009). Vertebrate fossils can be difficult to scan owing to the density of variable minerals associated with permineralization and preservation. Any remaining sediment on the specimen being scanned, both internally and externally, can have an adverse effect on scan quality—especially in cases where density differences between bone and sediment are minimal, or in some cases minerals that are too dense to allow X-ray transmission. Unfortunately, in this specimen, the dense mineral barite permeates cracks and natural openings such as the neural canal and haemal arch obfuscating the details of the internal structure, although some internal morphology can still be observed.
CT scans were obtained at the University of Wisconsin-Madison PET Imaging Center using a high-resolution multi-slice CT scanner (GE-Lightspeed VCT). A helical scan with a 0.625-mm slice thickness and settings of 120 kV and 400 mA were employed. Data were reconstructed via the Bone Plus algorithm on a GE Advantage Workstation and viewed in Escape Medical Viewer version 4.3.3 using the NIH contrast filter. Image data are stored on DVD and reposited with the specimen at the Wyoming Dinosaur Center, Thermopolis, WY.
The vertebra WDC LA-188 described here is represented by what appears, upon gross inspection, to be a single mid-caudal vertebral body of an adult Apatosaurus with two neural arches and one haemal arch attached to it (Fig. 2A,B). The shape of the anterior and posterior articular surfaces, a distinct ridge along the mediolateral surface of the vertebral body, and the presence of a weak midventral longitudinal hollow support the taxonomic assignment to a diplodocoid (Curtice, 1996; Upchurch et al., 2005). The vertebral body is longer than predicted, although it is consistent with predicted width and height of a mid-caudal vertebra that would be associated with the axial and appendicular elements from the individual Apatosaurus reported from the LA quarry (Birkemeier and Swor, 2007).
A weakly developed diffuse vertical ridge divides the anterior and posterior halves of the vertebral body; it is almost indistinguishable along the lateral surfaces lending to the appearance of a “normal” vertebral body. On the mediolateral surface, at the midpoint, a small (8 mm) smooth-edged fenestra is visible (Fig. 2C). The midlateral longitudinal ridge typical of apatosaurine caudal vertebrae is present with only a minor disruption across the junction of the anterior and posterior halves of the vertebral body (Figs. 2 and 3). There is a midventral longitudinal hollow as seen in A. ajax (Upchurch et al., 2005); this feature is present, but likely exaggerated due to lateral compression during diagenesis.
Neural and haemal arches
WDC LA-188 displays an anomalous pair of neural arches and spinous processes dorsally. The prezygapophyseal lamina of the posterior neural arch and the postzygapophyseal lamina of the anterior neural arch, in conjunction with the fused zygapophyses, provide a smooth and continuous transition between the neural arches; the intervertebral joints between the zygapophyses are nonexistent externally. The bone is smooth with no evidence of fusion, thus making an even transition between the coupled pre- and postzygapophyses (Fig. 2D). Both the neural arches and spinous processes exhibit a “rounded” or smoothed appearance along the anterior spinal laminae and dorsal most aspect as well as an overall stunted growth compared with the more flattened spinous processes of other apatosaurines. A single haemal arch (chevron) is adhered to the midventral surface of the vertebral body (Fig. 2E). No articular facets are evident and the arch appears to be fused or “welded” to the vertebral body. The sediment-filled haemal canal extends nearly the 75% the length of the chevron (Fig. 2B). The distal end of the haemal spine exhibits a similar stunted and rounded appearance seen in the neural spines.
Internal Morphology (CT Data)
Three sagittal sections obtained from the CT-scan, one along the midline and two left and right of midline, were analyzed to determine internal morphology and the relationship between vertebral elements in WDC LA-188 (Fig. 3). The scans depict variable densities, where warmer colors correspond to increased density (i.e., white = high density barite). The vertebral body exhibits a slightly less-dense interior consistent with more dense compact bone surrounding less-dense cancellous bone. The interior of the vertebra is expressed by cooler colors with warm-colored “mottles” interpreted to be mineral filled voids in cancellous bone. The vertical ridge (external morphology) that separates the anterior and posterior halves of the vertebral body appears to have a slight increase in density across this boundary internally as well (Fig. 3C). The anterior half of the vertebral body exhibits an overall higher density than the posterior half. The medial margin that separates the two halves exhibits a narrow vertical zone with densities approaching or equal to the dense zone seen on the anterior and posterior articular surfaces of the vertebral body.
Neural and haemal arches
The CT scans demonstrate the continuity of bone density across the zygapophyses. No evidence of a zygopophyseal joint is observed. Similarly, the haemal arch appears to maintain continuity across the contact with the vertebral body in the lateral sagittal section. The medial sagittal section (Fig. 3B) displays a bright (white color) area dorsally and posterior to the haemal arch consistent with other areas such as the neural canal, which are filled with barite-rich sediment; this feature is isolated to the midline in association with the haemal canal.
In all vertebrates, the vertebral column is derived entirely from paraxial mesoderm (Christ et al., 2000). Segmentation of the paraxial mesoderm leads to somitic formation and establishes an anterior–posterior polarity within the somites and their derivatives. The somitic mesoderm differentiates into a dorsal and ventral domain (dermomyotome and sclerotome, respectively) with undifferentiated somitocoele mesenchyme occupying the center of the somite (Fig. 4A). The dermomyotome (skeletal muscle, dermis, scapulae) migrates to the dorsolateral margin as the sclerotome differentiates into four distinct regions forming the dorsal, central, ventral, and lateral sclerotome (Fig. 4B,C). The dorsal sclerotome gives rise to the spinous process and the dorsal portion of the neural arch, while the pedicle and lamina of the neural arch and the proximal ribs are formed from central sclerotome with some contribution from somitocoele mesechyme. The ventral sclerotome gives rise to the vertebral body, whereas the distal ribs derive from the lateral sclerotome (Fig. 4D,E) (Huang et al., 1994; Christ et al., 2000; Stockdale et al., 2000; Mittapalli et al., 2005; for review see Christ and Scaal, 2008).
The somites undergo a further division called resegmentation, a process by which anteroposterior polarized somites divide into a cranial and caudal-half somite (Fig. 5A,B). The caudal-half somite then fuses with the following cranial-half somite to form a resegmented somite from which the metameric vertebrae and ribs arise (Fig. 5C) (Bagnall et al., 1988). It is during resegmentation that the somitocoele mesenchyme migrates to the caudal-half somite. The mesenchyme itself is segregated into dorsal and ventral compartments; these compartmentalized cells give rise to the intervertebral joints and intervertebral discs (Huang et al., 1994; Mittapalli et al., 2005; Christ and Scaal, 2008). This general model of vertebral development is dominantly derived from studies using the quail-chick chimeric system, although recent research has demonstrated similar results in an amphibian model (Piekarski and Olsson, 2013). The similarities seen in the amphibian and quail-chick chimeric system provide an extant phylogenetic bracket appropriate for the study of dinosaurian vertebral development (sensu Witmer, 1995).
Fusion of caudal vertebrae due to trauma or disease is not uncommon in sauropod dinosaurs. Rothschild and Berman (1991) reported osseous overgrowths spanning vertebral bodies in two of four specimens of Apatosaurus and three of six Diplodocus specimens studied with appropriate mid-anterior portions of the tail preserved. Diffuse idiopathic skeletal hyperostosis (DISH) has been suggested as the likely cause of these pathologies (Rothschild, 1987; Rothschild and Berman, 1991). The ligamentous, tendonous, or capsular ossification in specimens exhibiting DISH often display a “pasted-on” or “dripped candle-wax” appearance (Rothschild et al., 1994). Although WDC LA-188 does fall within the range of pathologically affected mid-anterior caudal vertebrae commonly observed in diplodocoid sauropods, its expressed morphology is not consistent with DISH. Specifically, there are no indications of ligamentous or tendonous fusions, no outgrowths that typify the physical manifestation of this disease, and no vacuity formerly occupied by the intervertebral disc (e.g., Rothschild and Berman, 1991).
Birkemeier and Swor (2007) suggested spondyloarthropathy (infectious arthritis) as a potential cause for the malformation seen in WDC LA-188. Spondyloarthropathy in vertebrae can lead to erosion or fusion of the zygapophyseal joints and vertebral body; zygapophyseal erosion has been reported in four consecutive caudal vertebrae of Camarasaurus (Rothschild et al., 2002). However, an aberrant vertebral body and reduced neural spines and lack of a space for an intervertebral disc (e.g., complete fusion) are not consistent with spondyloarthropathy (Rothschild et al., 1994).
It is suggested that a developmental error, rather than trauma or disease, led to the unique morphology exhibited by WDC LA-188, a morphology that resembles previously reported incidences of congenital block vertebrae (Kaplan et al., 2005; Rothschild and Tanke, 2005). Failures of segmentation in the spinal column lead to defects such as hemivertebae, scoliosis, wedge vertebrae, and block vertebrae (Kaplan et al., 2005; Witzmann, 2007; Witzmann et al., 2008). Block vertebrae involve the entire vertebra and there is no growth plate or somitocoele mesenchyme to form the intervertebral disc or intervertebral joints (Huang et al., 1994; Mittapalli, et al., 2005; Christ and Scaal, 2008). The lack of a discrete separation, both internally and externally, between the anterior and posterior halves of the vertebral body suggests an absence of an intervertebral disc. This is consistent with the hypothesis that WDC LA-188 is a congenital block vertebra (Fig. 5D) and not the result of spondyloarthropathy or DISH.
Externally, it is possible to distinguish between the anterior and posterior halves of WDC LA-188 by the weakly developed vertical ridge externally. Internally, there is an evident increase in density in the anterior half, in part due to differential permineralization; this disparity is thought to highlight differences in internal morphology rather than a strict result of taphonomic circumstance (Fig. 3). Furthermore, there is a thin (1–2 cm) vertical zone of markedly increased density along the anterior and posterior margins of the specimen, as well as medially along the same plane as the weakly developed ridge seen externally (Fig. 3C). The location and density contrast are consisted with known regions of cortical bone development indicating that WDC LA-188 is composed of two vertebral bodies.
Although it is difficult to diagnose the exact nature of this malformation, it is suggested that the most likely cause is the loss of the somitocoele mesenchyme during somitogenesis. A failure of the mesenchyme to migrate or proliferate in the caudal-half somite during resegmentation (sensu Bagnall et al., 1988) would result in the absence of intervertebral joints and an intervertebral disc (Huang et al., 1994; Mittapalli et al., 2005). The differences in density between the anterior and posterior halves of WDC LA-188 could also be explained by the lack of somitocoele mesenchyme. In addition to contributing to the proximal portion of dorsal ribs, intervertebral joints, and intervertebral discs, somitocoele mesenchyme is also know to have angiogenic potency (Huang et al., 1994). Loss of mesenchyme would have affected vascularization of the posterior vertebra, which might account for the differential permineralization of pore spaces seen in CT images.
What differentiates WDC LA-188 from other reported block vertebrae is the expected length for two fused vertebrae from this anatomical position. The morphology is consistent with a diplodocoid caudal and its association with the skeletal remains of an isolated Apatosaurus is compelling. However, if this vertebra does belong to the excavated Apatosaurus skeleton, the fused vertebral bodies are shorter than expected for two “normal” vertebrae. Although the shortened vertebrae may be the result of a misdiagnosed taxonomic assignment, it is suggested the shortening is a result of aberrant vertebral development due to the lack of a growth plate in addition to the absence of an intervertebral disc.
In conclusion, the developmental abnormality expressed in WDC LA-188 is consistent with a somitic resegmentation error. This error occurred during early embryonic development and led to the formation of a block vertebra. The absence of somitocoele mesenchyme would have inhibited the formation of intervertebral joints, including the intervertebral disc leading to the fusion of two vertebral elements. It is suggested that the loss of the growth plate at this margin prevented the anterior and posterior vertebral bodies that comprise the block vertebra to reach the lengths typical for its hypothesized anatomical position. WDC LA-188 displays morphological features that are strongly associated with failures of segmentation and represents the first congenital malformation observed in a sauropod dinosaur.
I thank the University of Wisconsin-Madison PET Imaging Center for imaging the specimen, the Wyoming Dinosaur Center for access to the specimen, and Escape Medical for software donation. Thanks also to Scott Hartman and Ben Linzmeier for comments, as well as two anonymous reviewers whose direction greatly improved the manuscript. Thanks also to Tristan Birkemeier for helpful discussions and excavation of the specimen.