Scenes from the Past

Note: This copy is for your personal non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at www.rsna.org/rsnarights.

SPECIAL EXHIBITS

1553

Scenes from the Past Initial Investigation of Early Jurassic Vertebrate Fossils with Multidetector CT1 Stephan A. Bolliger, MD • Steffen Ross, MD • Michael J. Thali, MD Bernhard Hostettler, MSc • Ursula Menkveld-Gfeller, PhD The study of fossils permits the reconstruction of past life on our planet and enhances our understanding of evolutionary processes. However, many fossils are difficult to recognize, being encased in a lithified matrix whose tedious removal is required before examination is possible. The authors describe the use of multidetector computed tomography (CT) in locating, identifying, and examining fossil remains of crocodilians (Mesosuchia) embedded in hard shale, all without removing the matrix. In addition, they describe how threedimensional (3D) reformatted CT images provided details that were helpful for extraction and preparation. Multidetector CT can help experienced paleontologists localize and characterize fossils in the matrix of a promising rock specimen in a nondestructive manner. Moreover, with its capacity to generate highly accurate 3D images, multidetector CT can help determine whether the fossils warrant extraction and can assist in planning the extraction process. Thus, multidetector CT may well become an invaluable tool in the field of paleoradiology. ©

RSNA, 2012 • radiographics.rsna.org

Abbreviations: FPJ = Fondation Paléontologique Jurassienne, 3D = three-dimensional RadioGraphics 2012; 32:1553–1559 • Published online 10.1148/rg.325115742 • Content Codes: From the Institute of Forensic Medicine, Department of Forensic Medicine and Imaging, University of Bern, Buehlstrasse 20, CH-3012 Bern, Switzerland (S.A.B., S.R., M.J.T.); and Earth Sciences Department, Natural History Museum of the Burgher Community of Bern, Bern, Switzerland (B.H., U.M.G.). Received August 26, 2011; revision requested August 31 and received December 27; accepted January 10, 2012. All authors have no financial relationships to disclose. Address correspondence to S.A.B. (e-mail: [email protected]). 1

©

RSNA, 2012

1554 September-October 2012

Introduction

Fossils are silent witnesses to life on Earth. This life began as simple creatures over 3 billion years ago. The first life-forms with hard structures such as shells or skeletons—and therefore with a greater chance of being fossilized—appeared about 500 million years ago. Fossils provide insight, not only into the habitats, climates, and ecosystems in which they lived, but also—through their sequential appearance during the earth’s history— into evolutionary developments. However, the extraction of fossils from their surrounding matrix, often shale, is time consuming and painstaking. Moreover, if no part of a fossil is detectable at external inspection, the fossil may be missed altogether. To counter these problems, radiography of rocks with fossil content has been performed for many decades. However, this method can produce only two-dimensional images with limited resolution. Computed tomography (CT) can overcome this shortcoming by creating three-dimensional (3D) images, thus greatly facilitating the examination of a fossil. Furthermore, with use of different reconstruction windows, it becomes possible to easily distinguish between structures with higher or lower attenuation at CT. Indeed, CT has been used in the field of paleopathology for over 3 decades: Fossilized bones were examined as early as 1979 (1), and subsequent studies have confirmed the usefulness of CT (2–6). However, the real advantages of CT-based examinations became evident with the virtual removal of the surrounding matrix (7,8), especially with the advent of high-powered CT scanners (9). In this article, we discuss the multidetector CT findings of selected fossils, which provide background information on lifeforms and their habitats as they existed millions of years ago.

CT Protocol

Scanning was performed on a six–detector row CT scanner (Emotion 6; Siemens Medical Solutions, Erlangen, Germany). Scanning parameters for raw data acquisition were 130 kv, 350

radiographics.rsna.org

mAs, and 0.5-mm collimation; and for image reconstruction, 0.625-mm thickness, 0.3-mm increments, bone-weighted (B70) reconstruction kernel, and extended CT scale. Image analysis and secondary image reformation (with, for example, multiplanar reformation, maximum intensity projection, or volume rendering) were performed on a Leonardo CT workstation (Siemens Medical Solutions).

Materials

All examined specimens were discovered in the Posidonia Shale (Toarcian stage [early Jurassic period]) in Dormettingen in southwest Germany. They belong to the Fondation Paléontologique Jurassienne (FPJ) and are preserved in its collection at Glovelier (Canton of Jura, Switzerland), where they remain accessible for further research.

Results and Discussion Specimen 1 Specimen 1 (FPJ 10215) measures 33 × 15 × 6 cm, with brown discolorations on two sides (Fig 1) and gray-white crystalline material on a third side (Fig 2). Multidetector CT revealed part of a thoracic spine with ribs and high-attenuation filigree structures (Fig 3). The shale, the fossilized bones, and the filigree structures had attenuation values of 1500–1800, 2300–2600, and 4000–14,000 HU, respectively. The nature of the filigree structures is unclear and therefore is open only to speculation. However, the distribution of this high-attenuation material strikingly resembles the multidetector CT appearance of adipocere in modern corpses found in wet, anaerobic conditions. Thus, we believe that the appearance of the filigree structures may reflect adipocere formation arising from the release of fatty acids during postmortem fat metabolism. Apart from having bactericide properties, adipocere can form an insoluble soap when hydrolyzed from the fat conjugate with bivalent ions such as Ca++ (calcium ions) (10–12). The formation of adipocere depends on the amount of body fat and the temperature, as well as the depth at which the body was submerged. Adipocere may have attenuation values in excess of 600 HU (13,14). Accord-

RG • Volume 32 Number 5

Bolliger et al 1555

Figures 1–4. (1) Photograph of specimen 1 shows a linear area of brown discoloration on one side (white arrow) and a brown spindle-shaped structure 19 cm in length on a different side (black arrows). The latter finding strongly suggests a vertebra and ribs. Scale is in millimeters. (2) Photograph of specimen 1 (side opposite the surface shown in Fig 1) shows the spindle-shaped structure (black arrows) with gray-white crystalline material in the center (white arrow). Scale is in millimeters. (3) On a multidetector CT scan of specimen 1, the left edge corresponds to the brown discoloration seen in Figure 1, and the bottom edge corresponds to the surface shown in Figure 2. Two rows of ribs are now clearly visible, thus proving the fossil to be of vertebrate origin. Some of the ribs display fractures (arrows). Furthermore, high-attenuation filigree structures corresponding to the gray-white crystalline material in Figure 2 are seen to meander paravertebrally. (4) Multidetector CT scan of specimen 1 (imaging plane shifted a few millimeters from that used in Figure 3) demonstrates rows of perforated rectangular plates (arrows) with an attenuation similar to that of the ribs. These plates correspond to the perforated bony armor plates seen in crocodilians. The morphologic characteristics of the plates prove that they belong to Steneosaurus bollensis.

ing to the literature, however, nothing is known about the changes that occur in adipocere during fossilization. It may well be that fossilized adipocere achieves the high attenuation values encountered in the filigree material. Validation of this hypothesis (based solely on morphologic appearance) has not been possible to date because the chemical changes in adipocere during fossilization will require further study. On the other hand, mineral deposits such as barite and pyrite are occasionally encountered in fossils. Pure barite and pyrite display attenuation

values of 11,000–16,000 and 9000–11,000 HU, respectively. The lower attenuation of the filigree structures may be influenced by the low attenuation of the adjacent fossilized bone, and it is at least possible that these structures are composed of such minerals (with their high attenuation) and therefore are not related to adipocere. Specimen 1 displays perforated bony plates lying on the rib cage (Figs 4–7), thus proving the specimen to be a crocodilomorph.



Figures 5, 6. (5) Three-dimensional reformatted multidetector CT image of specimen 1 shows rows of bony plates (arrowheads) alongside the spine (arrow). (6) Three-dimensional reformatted multidetector CT image (superoinferior view) of specimen 1 shows the distribution of the rows of bony plates (dashed lines) on either side of the spine.

Today, crocodilomorphs comprise eight genera of crocodiles, alligators, and gavials. Some 150 fossil crocodilians are known to date, the first of which appeared in the late Triassic period. Most of the Jurassic to Cretaceous crocodilians are considered to be Mesosuchia (as opposed to Eusuchia, which are modern crocodilians) and demonstrate great diversity of form. Among the Mesosuchia, the Thalattosuchia are thoroughly marine-adapted crocodilians such as the steneosaurids. Steneosaurids are divided into two groups on the basis of their skull shape: the longirostrine (long, slender jaws) and the brevirostrine (short, broad jaws). They hunted fish in the shallow seas and estuaries around Europe. In the Toarcian stage (early Jurassic period), two species of longirostrine steneosaurids are known: S bollensis (15), which lived in what is now England, France, and Germany; and S gracilirostris, which lived in modern-day England. The morphology of the bony armor plates, the age of the specimen (ie, Toarcian), and the location (Dormettingen, Germany) indicate that these were most likely the remains of an S bollensis, which closely resembled modern-day gavials and could grow to 5 m in length. In this case, multidetector CT sufficed to determine the species accurately in a nondestructive manner. The resulting images serve as a basis for an envisaged preparation, furnishing the preparer with information about the hidden fossil and the surrounding matrix. Moreover, multidetector CT was able to depict filigree structures of unknown origin, which may represent either fossilized adipocere or crystalline barite or pyrite deposits.

Figure 7. Three-dimen sional reformatted multidetector CT image of one of the bony plates shown in Figure 6 more clearly shows its perforated structure.

Specimen 2 Specimen 2 (no. 17 from FPJ 10254) measures 45 × 25 × 15 cm and is only a part of a whole set that probably contains crocodile fossils (Fig 8). Results of multidetector CT verified that this specimen contained a crocodile jaw (as had been assumed). In addition, multidetector CT showed that the jaw was broken at the front end and contained two tooth marks (Figs 9, 10), findings suggesting that the crocodile was bitten by a large predator. Later preparation confirmed the presence of bite marks (Fig 11). The shape of and distance between the teeth were consistent with a temnodontosaur (first described by Richard Lydekker in 1889), the largest predator of the Jurassic Sea (Fig 12).


Bolliger et al 1557

Figure 8. Photograph of specimen 2 (unprepared) shows structures resembling a long, slender snout (arrows). Scale is in centimeters.

Figures 9, 10. (9) Multidetector CT scan of specimen 2 depicts the jaw of a crocodile with a slender snout. The tip of the jaw (dashed green lines) has been bitten off and now lies at a right angle to the rest of the jaw (dashed yellow lines). Note the indentations at the fractured end of the jaw (arrows) representing bite marks from a very large predator. (10) On a multidetector CT scan (magnified view), the semicircular appearance of the indentation on the right (red arrow) cannot be explained as simply a postmortem fracture (although such a finding is frequently seen in fossil remains), but is evidence of a tooth mark. The semicircular shape indicates that the assailant was not a shark (a frequent and sometimes very large predator of the Jurassic seas), since shark teeth are essentially flat. Instead, the round shape clearly indicates that the assailant was an ichthyosaur (temnodontosaur). The other semicircular defect (yellow arrow) may be a broken-out dental alveolus belonging to the victim, or possibly another tooth mark from the bite of the predator. Figure 11. Photograph of specimen 2 obtained after 28 hours of removing the matrix from the bones clearly depicts the lower crocodile jaw. The end of the jaw (arrows) is fractured, and the tip of the jaw (dashed lines) is dislocated. Scale is in centimeters.



Temnodontosaurs were ichthyosaurs whose body shape somewhat resembled that of dolphins. They hunted in the early Jurassic seas in what is now England and Germany and could grow to over 12 m in length (16).

Specimen 3 Specimen 3 (FPJ 493) measures 35 × 27 × 1.5 cm, with discoloration and a discrete textural anomaly on one side (Figs 13, 14). CT and subsequent preparation showed a poorly preserved fish skeleton (Figs 15, 16). However, because of the disintegrated and incomplete state of the specimen, the species could not be determined, other than that it was most likely a predator. Multidetector CT also clearly depicted an ammonite siphuncle lying near the fish. A si phuncle is essentially a tube on the external side of the shell of an ammonite that passes longitudinally through the chambers and serves to exchange water. Siphuncles are believed to have served as a form of swim bladder and therefore probably allowed the mollusk to remain at a certain depth energy efficiently. Multidetector CT was able to depict other siphuncles from different specimens within the shale. However, it could not help determine the species of ammonites.

Conclusions

Multidetector CT can be of assistance in locating, identifying, and examining fossils embedded in a rocky matrix in two ways. First, it can help experienced paleontologists determine, in a nondestructive manner, the kinds and locations of fossils in the matrix of a promising rock specimen. Second, with its capacity to generate highly accurate 3D images, multidetector CT can help determine whether these fossils warrant extraction and can assist in planning the extraction process. Acknowledgments.—The authors wish to thank San-

dra Mathier, radiology technician, for assistance with multidetector CT scanning; and Michael J. Bolliger for the artwork in Figure 12.

Figure 12. Artist’s rendering of the attack, during which the large temnodontosaur bites the snout of the steneosaur. (Courtesy of Michael J. Bolliger, Lengnau, Switzerland.)

References 1. Jungers WL, Minns RJ. Computed tomography and biomechanical analysis of fossil long bones. Am J Phys Anthropol 1979;50(2):285–290. 2. Tate JR, Cann CE. High-resolution computed tomography for the comparative study of fossil and extant bone. Am J Phys Anthropol 1982;58(1):67–73. 3. Senut B. Computerized tomography of a Neanderthal humerus from leRegourdou (Dordogne, France): comparison with modern man. J Hum Evol 1985;14(8):717–723. 4. Sumner DR, Mockbee B, Morse K, Cram T, Pitt M. Computed tomography and automated image analysis of prehistoric femora. Am J Phys Anthropol 1985;68(2):225–232. 5. Golder W, Christian A. Quantitative CT of dinosaur bones. J Comput Assist Tomogr 2002;26(5): 821–824. 6. Ruf I, Luo Z, Martin T, Wible JR. CT scanning investigation of the petrosal bone and the inner ear of Henkelotherium guimarotae (Paurodontidae, Dryolestoidea). J Vertebr Paleontol 2007;27(3 suppl):137A. 7. Conroy GC, Vannier MW. Noninvasive threedimensional computer imaging of matrix-filled fossil skulls by high-resolution computed tomography. Science 1984;226(4673):456–458. 8. Prossinger H, Seidler H, Wicke L, et al. Electronic removal of encrustations inside the Steinheim cranium reveals paranasal sinus features and deformations, and provides a revised endocranial volume estimate. Anat Rec B New Anat 2003;273(1): 132–142. 9. Vock P, Soucek M, Daepp M, Kalender WA. Lung: spiral volumetric CT with single-breath-hold technique. Radiology 1990;176(3):864–867.


Bolliger et al 1559

Figures 15, 16. (15) Multidetector CT scan of specimen 3 reveals the spine (white arrow), ribs (black arrow), and tail fragments (white arrowhead) of a bony fish. A spiraled ammonite siphuncle (black arrowheads) is also seen. (16) Photo graph of specimen 3 (prepared) confirms the presence of the spine (white arrow), ribs (black arrow), and tail fragments (white arrowhead). The ammonite siphuncle is not visible in the matrix at this level.

Figures 13, 14. (13) Photograph of specimen 3 shows no indication of the presence of fossils within the slab. Scale is in centimeters. (14) Photo graph of one edge of specimen 3 shows a discrete textural anomaly (arrows), a finding that indicates the presence of a fossil within the slab.

10. Cotton GE, Aufderheide AC, Goldschmidt VG. Preservation of human tissue immersed for five years in fresh water of known temperature. J Forensic Sci 1987;32(4):1125–1130. 11. Takatori T. Investigations on the mechanism of adipocere formation and its relation to other biochemical reactions. Forensic Sci Int 1996;80(1-2): 49–61. 12. Takatori T. The mechanism of human adipocere formation. Leg Med (Tokyo) 2001;3(4):193–204.

13. Jackowski C, Thali M, Sonnenschein M, Aghayev E, Yen K, Dirnhofer R. Adipocere in postmortem imaging using multislice computed tomography (MSCT) and magnetic resonance imaging (MRI). Am J Forensic Med Pathol 2005;26(4):360–364. 14. Bohnert M, Schmialch A, Faller-Marquardt M, et al. Umwandlung des Gehirns und der Gesichts weichteile in Leichenlipid: morphologische und radiologische Befunde. Rechtsmedizin 1998;8(4): 135–140. 15. Jaeger GF. Über die fossile Reptilien, welche in Würtemberg aufgefunden worden sind. Stuttgart, Germany: Verlag der Metzler’schen Buchhandlung, 1828. 16. Maisch MW, Matzke AT. The ichthyosauria. Stuttg Beitr Naturkd B 2000;298:1–159.