Fig 1.
Schematic illustration of the state of fossil vertebrate material due to taphonomic processes.
The bone was fossilized into aragonite (grey-blue) and is enclosed in a thin layer of crystalline calcite (white). The fine-grained alkaline sediments (siltstone) around the bones usually turned grey-green in the process.
Fig 2.
Diagenetic distortion of the bone bearing strata at the Bromacker quarry result in plastic deformation of the fossils.
A: The original skull was bilateral symmetric with perpendicular rostro-caudal (not shown), latero-lateral (red), and dorso-ventral (blue) axes. B: Preservation is non-orthogonal to main directions of diagenetic distortion resulting in complex plastic deformation of a fossil. C: The main directions of diagenetic distortion were vertical compaction (large arrow) and, to a significantly lesser extent, horizontal tectonic deformations (smaller arrows). D: A volume render of the skull of MNG 10181 to illustrate the extent of plastic deformation, fractures of bones, and the high resolution of the CT scan data.
Fig 3.
CT scanning of the holotype specimen of Orobates pabsti (MNG 10181) and segmentation of bones and bone fragments from CT image stacks in Amira (here left manus shown as an example).
A: The main block of the specimen mounted within the v|tome|x L450, (GE phoenix|x-ray systems, Wunstorf, Germany) at the Institut für Leichtbau und Kunststofftechnik in Dresden, Germany. B: Because grey level differences between fossil bone and surrounding matrix were minimal and heterogeneous, outlines of bones had to be traced by hand on individual images of the stack. C: Bone outlines were traced on as many images as necessary to yield the correct 3D bone shape from interpolation. D, E: Voxels assigned to a bone were extracted from the matrix and volumes subsequently surface rendered.
Fig 4.
Workflow for the restauration of bone models here shown for the left humerus.
A: bone models segmented from the CT data were fragmented and suffered from plastic deformation. B: after importing bone models into Maya, fractures were repaired using available modelling tools. This step resulted in ‘watertight’ bone models. C: the surface mesh was re-mashed in ZBrush to more parsimonious quad-meshes to drastically reduce polygons and hence file size. D: surface models were un-distorted as explained in the text (also see Fig 5).
Fig 5.
Using simple lattice deformers to undistort MNG 10181 in Maya.
A, B: Direct evidence was used to restore bilateral symmetry. In (A) the original, distorted skull with rostro-caudal, latero-lateral, and dorso-ventral axes (bronze) is shown and distortion relative to the perfectly perpendicular green axes is obvious. By deforming a lattice containing the skull and the bronze axes so that the bronze axes match the green axes (B) bilateral symmetry is restored (rule 1). C, D, E: Circumstantial evidence used to correct for dorso-ventral flattening. In (D), the isolated vertebra MNG 8966 is not dorso-ventrally flattened due to preservation in flat orientation and has a round neural centrum. In (E), the fifth cervical vertebra of MNG 10181 displays dorso-ventrally flattened neural centrum (bronze) and was corrected until neural centrum had a round shape again (rule 2). Both rules were used to correct distortion for all CT derived bone models of the holotype specimen (C).
Fig 6.
Digital reconstruction of Orobates pabsti (MNG 10181).
A: Holotype specimen of Orobates pabsti (MNG 10181) housed at the Museum der Natur Gotha, Stiftung Schloß Friedenstein, Germany. B: Complete digital reconstruction in a non-physiological pose. C: Dorsal aspect, D: Ventral aspect. Grey bone models were derived from highest resolution scans (≤ 75μm voxel size); green and turquoise bone models derived from lower resolution scan (≤ 150μm voxel size). Turquoise bone models were slightly re-modelled to account for partially poor visibility of bone within the matrix. Blue bone models were modelled based on superficial visibility from photos and CT-scans and according to detailed description provided by Berman et al. [17]. Shape of thoracic vertebrae was modelled after highly detailed scan of an isolated vertebra (MNG 8966). Further explanation see text.
Fig 7.
The digital reconstruction of Orobates pabsti in a hypothesized naturalistic pose.
Stride length, stride width, and manus/pes orientation according to fossil trackways attributed to Orobates as the trackmaker [18]. Note that the naturalistic pose presented here requires a suite of assumptions, which should be explored with the help of modelling approaches in future studies. Assumptions include—but are not limited to—the amount of lateral bending of the trunk, the degree of adduction in the proximal limb joints, the relative contribution of pro- and retraction versus long axis rotation in the shoulder and hip to stride length. Within the constraints presented by the fossil trackways a range of poses is possible.
Fig 8.
Body mass estimation of MNG 10181.
A: Octagonal hugging hoops were used to estimate the body outline. B: Full body outline model. C: Cavities in the oropharyngeal area and lungs were accounted for. D: maximally caudal mass distribution model iteration, with CoM shown (heaviest plausible tail with lightest plausible trunk). E: maximally cranial mass distribution model iteration, with CoM shown (lightest plausible tail with heaviest plausible trunk). F, G: minimal and maximal overall body mass iterations. See text for further explanation.
Table 1.
Whole body and body segment mass estimation.
Minimal and maximal mass estimation and position of the respective CoM provided relative to the system origin (x, cranio-caudal; y, medio-lateral; z, dorso-ventral). Max. cranial and Max. caudal represent the extremes of the plausible envelope of the CoM position.
Fig 9.
Maximum range of motion in the shoulder (A) and hip (B) joint of Orobates pabsti (MNG 10181).
All values are relative to the reference pose and were determined from rotation of the stylopodium along axes of the anatomical joint coordinate system until collision of bones with an intermediate joint space of 1.5 mm (further explanation see text).
Table 2.
Maximum mobility of the shoulder (humerus) and hip (femur) relative to the reference pose until bone to bone contact in degrees.
Fig 10.
Details of the atlas-axis complex revealed by segmentation of CT image stacks.
Anterior directed to the left. A-C: In situ location of bone fragments in MNG 10181. D-E: idealized orientation and location of bone fragments (ribs and intercentrum of 3rd cervical not shown). A, D: dorsal aspect, B, E: lateral aspect, C, F: ventral aspect; rf: rib fragment, axi-atp: axial intercentrum plus atlantal pleurocentrum, atic: atlantal intercentrum, atr: atlantal rib, ax: axial neural arch plus axial pleurocentrum, atna: atlantal neural arch, pro: proatlas, axr: axial rib, 3cr: 3rd cervical rib, 3ic: 3rd cervical intercentrum. Please note that all bone models are distorted due to diagenetic processes. Also note the characteristic midventral furrow on the posterior surface of the atlantal intercentrum that receives an anteriorly projecting process of the axial intercentrum also described in Diadectes [36].