Anatomia, Histologia, Embryologia

ORIGINAL ARTICLE

Congenital Malformations of the Vertebral Column in Ancient Amphibians F. Witzmann1*, B. M. Rothschild2, O. Hampe1, G. Sobral1, Y. M. Gubin3 and P. Asbach4 €r Naturkunde, Leibniz-Institut fu €r Evolutions- und Biodiversit€atsforschung, Invalidenstraße 43, Berlin D-10115, Addresses of authors: 1 Museum fu Germany; 2 Biodiversity Center, University of Kansas, Lawrence, KS 66045, USA; 3 Paleontological Institute, Russian Academy of Sciences, ul. Profsoyuznaya 123, Moscow 117868, Russia; 4 Department of Radiology, Charit e – Universit€ atsmedizin Berlin, Chariteplatz 1, Berlin 10117, Germany

*Correspondence: Tel.: +49 30 2093 8820; fax: +49 30 2093 8565; e-mail: [email protected] With 8 figures Received November 2012; accepted for publication February 2013 doi: 10.1111/ahe.12050 €r This work was carried out at the Museum fu €r Evolutions- und Naturkunde, Leibniz-Institut fu Biodiversit€ atsforschung, Invalidenstraße 43, D-10115 Berlin, Germany.

Summary Temnospondyls, the largest group of Palaeozoic and Mesozoic amphibians, primitively possess rhachitomous vertebrae with multipartite centra (consisting of one horse-shoe-shaped inter- and paired pleurocentra). In a group of temnospondyls, the stereospondyls, the intercentra became pronounced and disc-like, whereas the pleurocentra were reduced. We report the presence of congenital vertebral malformations (hemi, wedge and block vertebrae) in Permian and Triassic temnospondyls, showing that defects of formation and segmentation in the tetrapod vertebral column represent a fundamental failure of somitogenesis that can be followed throughout tetrapod evolution. This is irrespective of the type of affected vertebra, that is, rhachitomous or stereospondylous, and all components of the vertebra can be involved (intercentrum, pleurocentrum and neural arch), either together or independently on their own. This is the oldest known occurrence of wedge vertebra and congenital block vertebra described in fossil tetrapods. The frequency of vertebral congenital malformations in amphibians appears unchanged from the Holocene.

Introduction Temnospondyl amphibians and their vertebrae Temnospondyl amphibians are the by far largest and most diverse group of basal tetrapods, ranging from the Early Carboniferous to the Early Cretaceous (Schoch, 2009). The group probably contains the ancestors of some (Anderson et al., 2008) or all (Ruta and Coates, 2007; Sigurdsen and Green, 2011) extant lissamphibians, although an alternative hypothesis exists (Marjanovic and Laurin, 2008). Temnospondyls were adapted to a large spectrum of habitats and are represented by aquatic, terrestrial and semi-terrestrial forms, spanning a wide size range from small, newt- or salamander-like forms like dissorophoids to the several-metre-long, crocodile-like stereospondylomorphs (Schoch, 2009; Witzmann et al., 2010). The vertebral morphology and ontogeny of temnospondyls differ from those of all extant vertebrates. © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

Temnospondyl vertebrae are plesiomorphically rhachitomous, that is, they are composed of the neural arch (including the processus spinosus) and a multipartite vertebral body, which consists of a large, unpaired intercentrum (or hypocentrum) and of paired, smaller pleurocentra (Moulton, 1974; Panchen, 1977; Shishkin, 1989; Warren and Snell, 1991) (Fig. 1a). The intercentrum is wedge-shaped in lateral and crescent in sagittal view, embracing the persistent notochord from ventral and lateral. The parapophysis for articulation with the capitulum of the ribs is located on the posterodorsal margin of the intercentrum. Posterodorsal to the intercentrum and posterior to the transverse processes of the neural arch are the diamond-shaped pleurocentra, embracing the notochord dorsolaterally. In the rhachitomous vertebra, the neural arch, intercentrum and pleurocentra are normally separated by cartilage rather than being co-ossified. In some stereospondyls (mainly Mesozoic temnospond-

1

Malformations in Ancient Amphibians

(a)

(b)

F. Witzmann et al.

(c)

(d)

(e)

(f)

(g)

yls), the intercentrum is often strongly ossified and has attained a disc-like or spool-shaped morphology, greatly reducing the space for the notochord (Warren and Snell, 1991). In contrast, the pleurocentra are often smaller than in the rhachitomous vertebra or are even non-ossified or reduced (Fig. 1b). This vertebral morphology is generally designated as ‘stereospondylous’ and can be found in large-growing stereospondyls, such as Mastodonsaurus (Schoch, 1999) and Cyclotosaurus hemprichi (Kuhn, 1942), and in metoposaurids (Warren and Snell, 1991; Sulej, 2007). Typical for metoposaurid intercentra of anterior trunk vertebrae is their opisthocoelous morphology, that is, the intercentrum is anteriorly convex and posteriorly concave, thus forming a kind of ball-andsocket joint. This may represent the origin of synovial intercentral joints. Among stereospondyls, the Triassic

2

Fig. 1. (a)–(c) Schematic drawings of rhachitomous, stereospondylous and plagiosaurid vertebrae. (a) Rhachitomous condition (redrawn from Shishkin, 1989). (b) Stereospondyl condition (redrawn from Warren and Snell, 1991). (c) Plagiosaurid condition, note the intervertebral position of neural arches and parapophyses (redrawn from Shishkin, 1989). (d)–(g) Skeletal reconstructions of some of the temnospondyl amphibians investigated in this study. (d) The Early Permian Sclerocephalus haeuseri (total body length approximately 1.5 m; redrawn after Schoch and Witzmann, 2009a). (e) The Late Permian Platyoposaurus stuckenbergi (total body length approximately 1.5 m; drawing based on a mounted skeleton at the Paleontological Institute and Museum of the Russian Academy of Sciences, Moscow, Russia). (f) A Triassic metoposaurid (Metoposaurus diagnosticus krasiejowensis, total body length approximately 2 m, redrawn after Sulej, 2007). (g) The Middle Triassic Gerrothorax pulcherrimus (total body length approximately 1 m; redrawn after Schoch, 2009). Drawings are not to scale. Abbreviations: c, centrum; dia, diapophysis; ic, intercentrum; na, neural arch; par, parapophysis; pc, pleurocentrum.

plagiosaurids have spool-shaped vertebral centra with intervertebral neural arches (Shishkin, 1987, 1989; Warren and Snell, 1991). Each parapophysis of plagiosaurid presacral vertebrae is formed by two successive vertebral centra, and thus, the ribs are intervertebral as are the neural arches (Fig. 1c). It is still a matter of debate as to which central elements form the plagiosaurid centrum. Panchen (1959) suggested that the centra are entirely formed by the pleurocentra, whereas the intercentra are lost. Shishkin (1987, 1989) interpreted the plagiosaurid centrum as fusion of the intercentrum with the pleurocentrum of the preceding vertebra, whereas Warren and Snell (1991) regarded the plagiosaurid centrum as an intercentrum and the pleurocentrum as reduced. Hellrung (2003) followed this view, but regarded the pleurocentrum as fused with the neural arch. © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

The ontogeny of temnospondyl vertebrae is well documented compared with vertebrae of other fossil tetrapods, as large growth series from small larvae to large adults do exist in several taxa (e.g. Boy, 1974; Schoch and Witzmann, 2009a,b). In general, vertebral ossification proceeded very slowly, starting with the initially paired neural arches, followed much later in ontogeny by ossification of the intercentrum (first laterally paired) and then of the laterally paired pleurocentra. Congenital vertebral malformations During somitogenesis, the paraxial mesoderm that is located lateral to the neural tube is segmented early in vertebrate embryogenesis and the bilaterally paired somites are formed. Somites contain sclerotomal cells that migrate from contralateral somite pairs in a medial and ventral direction and surround the notochord and neural tube, thus forming the mesenchymal anlagen of the vertebrae (Erol et al., 2002; Kaplan et al., 2005). Disruption of genes regulating embryonic somite formation (e.g. by environmental insults during early embryogenesis like oxygen deficiency, increased temperature and carbon monoxide) can cause abnormal segmentation and disruption of fusion of the paired mesenchymal vertebral anlagen, leading to congenital malformations like butterfly, block, wedge and hemivertebrae (Pourquie and Kusumi, 2001; Erol et al., 2002; Shawen et al., 2002; Kaplan et al., 2005). An interaction between genes and environment probably exists, that is, genetic defects cause the susceptibility of the embryo to disease-associated environmental factors (Erol et al., 2002, 2004). A hemivertebra may develop from complete failure of formation of one lateral vertebral anlage. Subsequent chondrification and ossification take place only on one lateral side. A special case of hemivertebra formation is the hemimetameric segmental shift, which is a defect of fusion of the paired vertebral anlagen (Shawen et al., 2002; Witzmann et al., 2008). Incarcerated and non-incarcerated types can be distinguished among hemivertebrae (McMaster, 2001). The non-incarcerated type acts like a wedge in the vertebral column and leads to a lateral curvature (scoliosis) of the column at the location of the hemivertebra. In the incarcerated type, which is less common in humans, the vertebrae anterior and posterior to the hemivertebra are shaped to compensate for the hemivertebra, such that no or only a slight curvature of the vertebral column occurs. Hemivertebrae were extensively studied in humans (e.g. McMaster and Ohtsuka, 1982; McMaster, 2001), but is also recognized among dogs, cats, horses and other domestic animals (Wong et al., 2005; Jeffery et al., 2007; Moura et al., 2010) as well as in snakes (Baur, 1891) and feline ectomorphs like Hoplophoneus (Rothschild et al., in © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

Malformations in Ancient Amphibians

press). A wedge vertebra has a similar shape, but in contrast to a hemivertebra, it extends to the contralateral side of the vertebral column. A failure of segmentation results in a block vertebra, in which the disc spaces between two or more vertebrae have become very narrow or fused (McMaster, 2001). Congenital vertebral pathologies are exceptional finds in fossil amphibians and reptiles (Rothschild et al., 2012). They were described in an Early Permian captorhinomorph reptile (Johnson, 1988), the Late Jurassic dinosaur Dysalotosaurus lettowvorbecki (Janensch, 1934; Witzmann et al., 2008), and briefly mentioned by Lydekker (1889) in the Late Jurassic cryptocleidid plesiosaur Muraenosaurus leedsi (designated as Cimoliasaurus plicatus by Lydekker). Among temnospondyl amphibians, congenital vertebral pathology has so far only been described in a Triassic capitosauroid that suffered from scoliosis caused by a hemivertebra (Witzmann, 2007). Pathologies of an extinct organism are important to document because they might give insights into the animal’s physiology and behaviour (Rothschild and Martin, 2006). Congenital vertebral malformations in temnospondyls provide additional information concerning the formation of the most primitive tetrapod vertebral pattern that does not have an analogue today. In this study, we describe different types of vertebral pathologies in different temnospondyls, including incarcerated and non-incarcerated hemivertebrae, wedge and block vertebrae, and discuss their aetiology and development.

Materials and Methods Sclerocephalus haeuseri (basal stereospondylomorph) Specimen MB.Am.1260.1, 2 from the Niederkirchen Bank (Meisenheim Formation: Jeckenbach Subformation) of Heimkirchen, Early Permian, Saar-Nahe Basin (Germany), consists of plate and counterplate. It is an almost complete postcranial skeleton of a larva. The trunk measures approximately 70 mm in length. Sclerocephalus haeuseri was a crocodile-like, semi-aquatic predator in the ancient lakes of the Saar-Nahe Basin and reached a total body length of more than 1.5 m (Fig. 1d) (Schoch and Witzmann, 2009a). ‘Cheliderpeton’ lellbachae (basal stereospondylomorph) Specimen SMNS 91279 is the cast of a complete skeleton, showing two succeeding trunk vertebrae whose neural spines are fused. The specimen is derived from the Kappeln Bank (Meisenheim Formation: Odernheim Subformation) of Klauswald/Odernheim, Early Permian, Saar-Nahe Basin (Germany). The taxon Cheliderpeton

3

Malformations in Ancient Amphibians

lellbachae was erected by Kr€atschmer (2006), but needs taxonomic revision as it does not share the autapomorphies of Cheliderpeton (Schoch and Witzmann, 2009b). ‘Cheliderpeton’ lellbachae had a similar mode of life as Sclerocephalus haeuseri and reached a total body length of 1–1.5 m. Platyoposaurus stuckenbergi (Archegosauridae, basal stereospondylomorph) Specimen PIN 164/71 is a single rhachitomous vertebra consisting of neural arch, inter- and paired pleurocentra. The intercentrum is posteriorly fused with the intercentrum of the succeeding vertebra, and the intercentrum of a hemivertebra is intercalated between the two on the right side. The specimen is derived from the Late Permian (Urzum stage, formerly Kazanian stage) of Belebey, Republic of Bashkortostan (Bashkiria). Platyoposaurus was an upto-1.5-m-long, aquatic piscivorous predator whose snout was extremely elongate and slender (Fig. 1e) (Gubin, 1991). Metoposauridae indet. (Trematosauria, Stereospondyli) Specimen MB.Am.1449 (formerly IPFUB Am.36) consists of two fused stereospondylous intercentra with intercalated hemivertebra on the left side. The neural arches and the pleurocentra were not co-ossified with the intercentra and are not preserved. It was found together with other remains of stereospondyls (vertebrae, fragments of pectoral girdle and skull of metoposaurids and mastodonsauroids) in the Gres de Silves Formation (Triassic/Jurassic boundary) of the Algarve Basin, south-western Portugal (Witzmann and Gassner, 2008). Metoposaurids were up-to-3-m-long aquatic predators that superficially resembled broad-skulled crocodiles or Giant Salamanders (Fig. 1f) (Sulej, 2007). Gerrothorax pulcherrimus (Plagiosauridae, Stereospondyli) SMNS 83498 is represented by two specimens: specimen A consists of two fused vertebral centra with a centrum of a wedge vertebra and two neural arches preserved; specimen B consists of two fused vertebral centra with the neural arches missing. The specimens are derived from the lower Keuper, Kupferzell, south-west Germany. Gerrothorax was a gill-breathing, flattened lurking predator that lived on the bottom of different types of water bodies (Fig. 1g) (Hellrung, 2003; Schoch and Witzmann, 2012). Micro-CT The vertebrae of Gerrothorax were scanned in the Museum f€ ur Naturkunde Berlin using a Phoenix|X-ray

4

F. Witzmann et al.

Nanotom (GE Sensing and Inspection Technologies GmbH, Wunstorf, Germany), which was especially designed for small samples and allows for higher resolution in the visualization of small structures. All 1440 slices were reconstructed with the software datos| x-reconstruction 1.5.0.22 (GE Sensing and Inspection Technologies GmbH, Phoenix|X-ray), and the threedimensional data were analysed in VG Studio Max 2.1 (Volume Graphics, Heidelberg, Germany). The scans were made with a tungsten target and a 0.1-mm-thick Cu filter in modus 0. The particular setting for Gerrothorax SMNS 83498 specimen A was 75kV, 350 lA, average 6, skip 3, exposure time of 250 ms and voxel size of 31.24 lm; for specimen B, the setting was 120 kV, 65 lA, average 3, skip2, exposure time of 250 ms and voxel size of 26.87 lm. Micro-CT scanning of the metoposaurid specimen MB.Am.1449 yielded no results. The scanning of the Platyoposaurus vertebrae PIN 164/71 could not be realized. The ‘Cheliderpeton’ lellbachae specimen SMNS 91279 is a cast of a lost original, and macroscopic investigation of the larval Sclerocephalus MB.f.1260 was sufficient because its very delicate, thin bones are all compressed two-dimensionally in a single layer. Abbreviations (institutional) MB, Museum f€ ur Naturkunde, Berlin, Germany; IPFUB, Institut f€ ur Pal€aontologie, Freie Universit€at Berlin, Germany; PIN, Paleontological Institute and Museum of the Russian Academy of Sciences, Moscow, Russia; SMNS Staatliches Museum f€ ur Naturkunde Stuttgart, Germany.

Results Hemivertebra in a larval rhachitomous vertebral column (Sclerocephalus) The vertebral column of the larval specimen of Sclerocephalus (MB.Am.1260.1, 2) is incompletely ossified, as common in temnospondyl larvae (Fig. 2a,b). The central elements, that is, inter- and pleurocentra, were completely cartilaginous in that growth stage and were thus not preserved. Only the neural arches are ossified, but are poorly differentiated with short zygapophyses and low neural spines. The neural arches are not fused in the midline (as in adult specimens), so each neural arch is represented by its paired, contralateral halves. However, the fourth preserved neural arch of the left side has two counterparts on the right side. Both of these right neural arches are anteroposteriorly shortened (they attain approximately 80% of the length of the left neural arch) but have approximately the same height. Ribs are poorly preserved in this specimen, but it appears that both of © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

Malformations in Ancient Amphibians

(a)

Fig. 2. Hemivertebra in a larval rhachitomous vertebral column, exemplified by a larval specimen of Sclerocephalus haeuseri (MB.Am.1260.1, 2) from Heimkirchen, Early Permian, Saar-Nahe Basin (Germany). (a) Complete specimen MB.Am.1260.1 in dorsal view. (b) Close-up of vertebral column with hemivertebra. Abbreviations: clei, cleithrum; fe, femur; fi, fibula; ha, haemal arch; hu, humerus; il, ilium; man, manus; na, neural arch; ra, radius; ri, rib; sna, smaller neural arches; sri, sacral rib; ti, tibia; ul, ulna.

(b)

the smaller neural arches are associated with ribs. The rib of the left counterpart is not preserved. This asymmetry can best be explained by the failure of formation of one left lateral vertebral anlage (i.e. left halves of neural arch, inter- and pleurocentrum did not develop), resulting in formation of a hemivertebra on the right side. The fact that the two right neural arches are anteroposteriorly shortened compensates partially for the presence of a hemivertebra so that the vertebral column is not curved in the region of this asymmetry (Fig. 2a). Hemivertebra in an adult rhachitomous vertebral column (Platyoposaurus) The investigated specimen (PIN 164/71) is derived from the presacral vertebral column and consists of two intercentra, a hemivertebral intercentrum, two pleurocentra and a neural arch, with all vertebral elements being connected by bone (Fig. 3a–e). The rhachitomous vertebrae in Platyoposaurus are interpreted as being ‘anteropleural’ sensu Shishkin (1989) [i.e. the pleurocentra of a respective vertebra are associated with the intercentrum posterior to it (see discussion below)] and the central elements are designated accordingly. The anterior, crescent intercentrum (named here intercentrum 1) is ventrally co-ossified with the posterior intercentrum (intercentrum 2) by unfinished bone (i.e. covered by cartilage in life), and thus, the boundary between both bones is well demarcated. The appertaining pleurocentra and neural arch of intercentrum 1 are not preserved, probably due to the lack of co-ossification with intercentrum 1. Both intercentra have slightly concave ventral and lateral sides. The periosteal bone surface of the right half of intercentrum 1 bears some large nutrient foramina on its ventrolateral part (Fig. 3b). The wedge-like left pleurocentrum 2 (belonging to intercentrum 2) is situated between © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

intercentra 1 and 2. It extends far ventrally, nearly reaching the ventral midline (Fig. 3a). It is co-ossified with intercentrum 2 by unfinished bone and is separated from intercentrum 1 by an unossified gap. On the right side, an intercentrum of a hemivertebra (hemivertebral intercentrum) is intercalated between intercentrum 1 and 2 and has a bony connection with both (Fig. 3b). Whereas the boundary between intercentrum 1 and the hemivertebral intercentrum is clearly traceable by a line of unfinished bone, the boundary with intercentrum 2 is evident only in its dorsalmost part (Fig. 3c). As on the right halves of intercentrum 1 and 2, an approximately circular parapophysis with unfinished surface is developed in the dorsolateral part of the hemivertebral intercentrum. Compared with the hemivertebral intercentrum, the parapophyseal facets on the left side of intercentra 1 and 2 are larger. The pleurocentrum of the hemivertebra apparently failed to develop completely. Posterodorsal to the hemivertebral intercentrum and anterodorsal to intercentrum 2 is the right pleurocentrum 2, developed only in its dorsalmost portion in contrast to its left counterpart. The neural arch belonging to intercentrum 2 and pleurocentra 2 is developed normally with the exception that the right transverse process is distinctly shorter than the left one. The neural arch of the hemivertebral intercentrum failed to develop. The hemivertebral intercentrum is accommodated in a niche on the right side between intercentra 1 and 2, thus representing a hemivertebra of the incarcerated type. It can be regarded a result of failure of formation, because there is no indication for hemimetameric segmental shift (see discussion in Witzmann et al., 2008). The complete, smooth fusion of the hemivertebral centrum with intercentrum 2 can be regarded as the result of embryonic failure of segmentation, which often accompanies the development of hemivertebra (McMaster, 2001). The remaining co-ossifications between the other

5

Malformations in Ancient Amphibians

(a)

F. Witzmann et al.

(b)

(c)

(d)

(e)

vertebral elements are established by unfinished bone, and the sutures are well traceable (intercentrum 1 with intercentrum 2, left pleurocentrum 2 with intercentrum 2, right pleurocentrum 2 with neural arch, hemivertebral intercentrum and intercentrum 2), what can be interpreted as post-embryonic co-ossification. The fact that the neural arch of the hemivertebra failed completely to develop on both lateral sides of the column (whereas the hemivertebral intercentrum was formed) indicates that the development of the particular vertebral elements was affected independently. Hemivertebra in a stereospondylous vertebral column (Metoposauridae) Specimen MB.Am.1449 consists of two fused intercentra (named here intercentra 1 and 2) of the thoracic vertebral column (Fig. 4a–e). Neural arches and pleurocentra are not preserved. Each intercentrum is cylindric or spool-shaped, with slightly concave ventral and lateral sides. The intercentra show a clearly opisthocoelous

6

Fig. 3. Hemivertebra in a rhachitomous vertebral column. (a)–(e), Platyoposaurus stuckenbergi (PIN 164/71) from the Late Permian of Belebey, Republic of Bashkortostan. (a)–(c) Drawings in (a) left lateral, (b) right lateral and (c) ventral view. (d)–(e) Photographs of vertebral centra in left lateral and (e) right lateral view. Abbreviations: dia, diapophysis; hic, intercentrum of a hemivertebra; hpar, parapophysis of hemivertebra; ic, intercentrum; na, neural arch; par, parapophysis; pc, pleurocentrum.

morphology. The dorsal side of the centra, which was connected with the neural arches by cartilage, is unfinished and shows no anatomical details. The lateral and ventral surfaces of the centra consist of smooth, periosteal bone. In right lateral view, the boundary between intercentra 1 and 2 is indicated dorsolaterally and laterally by a dorsoventral indentation of unfinished bone and ventrolaterally by a broad, shallow ridge. This ridge ends abruptly on the ventral side of the specimen. Additionally, the right side of the specimen shows two anteroposteriorly elongate parapophyses with an unfinished surface. Each parapophysis is located in the dorsolateral part of the respective intercentrum. Contrasting with the right side, the left side of the specimen shows a third parapophysis between the anterior and the posterior parapophyses. It is located a short distance posterior to the anterior one and directly anterodorsal to the posteriormost parapophysis. The discrepancy between the left and the right sides of this specimen is caused by the intercalation of a hemivertebral intercentrum between intercentra 1 and 2 on the left side. The boundary © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

Malformations in Ancient Amphibians

(c)

(a)

(b)

Fig. 4. Hemivertebra in a stereospondylous vertebral column. (a)–(e) Metoposauridae indet. (MB.Am.1449) from the Triassic/Jurassic boundary of the Algarve, Portugal. (a)–(c) Drawings in (a) right lateral, (b) left lateral and (c) ventral view. (d)–(e) Photographs in (d) right lateral and (e) left lateral view. Abbreviations: hic, intercentrum of a hemivertebra; hpar, parapophysis of hemivertebra; ic, intercentrum; par, parapophysis.

(d)

between intercentrum 1 and the hemivertebral intercentrum is indicated laterally by a narrow dorsoventral indentation of unfinished bone; more ventrally, no suture is detectable. This region is marked by large nutrient foramina. The boundary between the hemivertebral intercentrum and intercentrum 2 is indicated by a short lateral indentation of unfinished bone. These observations indicate that fusion between the hemivertebral intercentrum and intercentra 1 and 2 was complete. The left parapophysis of intercentrum 1 corresponds in size to those on the right side. The parapophyses of the hemivertebral intercentrum and of intercentrum 2 are located more dorsally and that of intercentrum 2 is shorter. Both intercentra 1 and 2 differ from the ‘normal’ cylindric morphology of metoposaurid centra in that they are anteroposteriorly shortened on the left side. In this way, a recess is formed that accommodates the hemivertebral intercentrum. Thus, the hemivertebral intercentrum is of the incarcerated type as in Platyoposaurus. It can similarly be regarded a result of failure of formation; at least the right intercentrum did not develop, but nothing can be said about the neural arches. The metoposaurid hemivertebra is non-segmented (i.e. it is fused with the anteriorly and posteriorly neighbouring centra). As this fusion is complete and smooth, it can be interpreted as an embryonic defect of segmentation. © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

(e)

Wedge vertebra and block vertebrae in two plagiosaurid specimens (Gerrothorax) Because it is still not clear whether the plagiosaurid vertebral centrum represents an intercentrum, pleurocentrum or both, it will be designated in the following as ‘centrum’. In the isolated specimen A belonging to SMNS 83498, two spool-shaped vertebral centra (named here centra 1 and 2) are fused and form a block vertebra. A centrum of a wedge vertebra is fused to the posterior endplate of centrum 2 (Fig. 5a–c). The centrum of the wedge vertebra is anteroposteriorly much longer on the left than on the right lateral side of the column. On its shortened right side, its posterior parapophyseal facet forms a common, elongate parapophysis with centrum 1, but its anterior parapophyseal facet is not developed. No trace of a suture or boundary between centra 1 and 2 and between centrum 2 and the centrum of the wedge vertebra can be detected, even not on the parapophyses. The bone surface is entirely smooth. Compared with the length of ‘normal’ vertebral centra, the length of each segment is shortened. This is common in block vertebrae because longitudinal growth of the vertebrae is impaired by the fused disc spaces (McMaster, 2001). Two neural arches are preserved (referred to as neural archs 1 and 2) and are co-ossified. Neural arch 1 is co-ossified with

7

Malformations in Ancient Amphibians

F. Witzmann et al.

(a)

(b)

(c)

(d)

(e)

(f)

centra 1 and 2, and neural arch 2 is co-ossified with centrum 2 and the centrum of the wedge vertebra. The subsequent neural arch was connected with the posterior half of the centrum of the wedge vertebra by an unossified suture and is not preserved. Neural arches 1 and 2 are poorly preserved, but appear to be normally developed. It can be assumed that the missing subsequent neural arch was developed similar to the centrum of the wedge vertebra with an anteroposteriorly shortened right lateral half. The anterior endplate of centrum 1 and the posterior endplate of the centrum of the wedge vertebra form an angle of 18° with each other, thus causing a slight lateral flexure (scoliosis) of the vertebral column. In the isolated specimen B that belongs to SMNS 83498, two centra (called here centra 1 and 2) are fused (Fig. 5d–f). The neural arches were connected with the centra by unossified neurocentral sutures and are not

8

Fig. 5. Wedge and block vertebrae in plagiosaurid vertebrae (Gerrothorax pulcherrimus from the lower Keuper (Middle Triassic) of Kupferzell, south-west Germany). (a)–(c) SMNS 83498 (specimen A), block vertebra with fused wedge vertebra in (a) left lateral, (b) right lateral and (c) dorsal view. (d)–(f) SMNS 83498 (specimen B), block vertebra in lateral (d, e) and dorsal (f) view. Abbreviations: c, centrum; dia, diapophysis; fnc, floor of neural canal; na, neural arch; ncf, neurocentral facet; par1 + 2, parapophysis formed by centrum 1 and 2; par 2 + w, parapophysis formed by centrum 2 and centrum of wedge vertebra; su, suture in parapophysis between two fused centra; wc, centrum of wedge vertebra; wpar, parapophysis of wedge vertebra.

preserved. A suture in the middle of the parapophysis indicates that centra 1 and 2 are equal in size viewed from this side (Fig. 5d). The suture is still traceable on the neurocentral sutural facet dorsomedial to the parapophysis (Fig. 5f). Further traces of sutures cannot be detected because fusion is complete and the bone surface is entirely smooth. The parapophysis formed by centra 1 and 2 of the other lateral side is too poorly preserved to detect a suture on it (Fig. 5e). This parapophysis is distinctly anteroposteriorly shortened as compared to its counterpart (25% shorter) and is located not in the middle of the specimen. Observed from this side, centra 1 and 2 are of unequal length. In dorsal view, the floor of the neural canal and the neurocentral sutural facets are preserved (Fig. 5f). Because of the intervertebral position of both the neural arches and the ribs, the neurocentral sutural facets are exactly dorsomedial to the © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

parapophyses. The neurocentral sutural facet of the side with the shortened parapophysis is also anteroposteriorly shortened compared with its lateral counterpart. Thus, both centra have a slightly rhombic rather than square outline in dorsal view. Because the anterior endplate of centrum 1 and the posterior endplate of centrum 2 are parallel to each other, no curvature of the vertebral column took place. Specimen A shows a defect of segmentation (fused centra) and a partial failure of formation (wedge vertebra posterior to centrum 2), whereas specimen B shows only a defect of segmentation, that is, fusion of two complete centra. The fusions in both Gerrothorax specimens are entirely smooth and thus strongly suggest incomplete separation of somites or their associated mesenchyme during early embryogenesis. Micro-CT imaging, showing an entirely homogeneous aspect of the spongy bone with no evidence for distortion of the trabecular pattern and the absence of any cortical structures within the centrum (Fig. 6), clearly indicates a complete fusion of the centra. The centrum of the wedge vertebra in specimen A can be designated as a semi-segmented wedge vertebra, because it is fused to its anteriorly located neighbouring centrum, but not to the posteriorly following one. The wedge vertebra and congenital block vertebra described here in Gerrothorax (and ‘Cheliderpeton’, see below) are the oldest described occurrences of these malformations in fossil vertebrates.

Malformations in Ancient Amphibians

Fig. 6. Micro-CT scan of block vertebra with fused wedge vertebra in sagittal cross-section (Gerrothorax pulcherrimus from the lower Keuper (Middle Triassic) of Kupferzell, southwest Germany, SMNS 83498, specimen A). Abbreviations: c, centrum; na, neural arch.

(a)

(b)

Fused neural spines (‘Cheliderpeton’ lellbachae) The neural arches of the 11th and 12th presacral vertebrae of ‘Cheliderpeton’ lellbachae (SMNS 91279) are partially fused (Fig. 7a,b). This fusion affects the dorsal half of the spine and is so complete that even no trace of a suture or boundary is visible. This gives the dorsal half of the spine the appearance of a single, elongate bone. The bone surface shows no signs of ossified tendons or ligaments. Ventral to the fused portion, the two neural arches are clearly not co-ossified or sutured, but abut against each other and are thus much closer together than ‘normal’ adjacent neural arches. Post- and pre-zygapophyses on the 11th and 12th neural arches, respectively, are poorly developed. The transverse processes of both neural arches are normally developed and articulate with the corresponding ribs. Unfortunately, it cannot be ascertained how the centra of these rhachitomous vertebrae were affected. Apart from these partially fused neural spines, the vertebral column of this specimen shows no signs of pathologies and is straight. Fused neural spines superficially similar to those evident in ‘Cheliderpeton’ lellbachae are described in human and veterinary medicine as Baastrup’s phenomenon or disease © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

Fig. 7. ‘Cheliderpeton’ lellbachae SMNS 91279 from the Early Permian of Klauswald/Odernheim, Saar-Nahe Basin (Germany), with two neural arches that are completely fused dorsally. (a) Drawing of specimen. (b) Photograph of specimen. Abbreviations: dia, diapophysis; na, neural arch; ri, rib; tp, transverse process of neural arch.

(sometimes called ‘kissing spines’). This phenomenon is characterized by the approach and contact of adjacent neural spines, causing size increase, flattening and reactive sclerosis of apposing interspinous surfaces (Bywaters and Evans, 1982; Resnick, 1985; Kacki et al., 2011). However, the diagnosis of Baastrup’s phenomenon in ‘Cheliderpeton’ lellbachae can be rejected because the fusion is entirely smooth with no reactive bone surface and there is no size increase of the neural spines. The smooth, complete fusion and the poorly developed zygapophyses of the vertebrae in question indicate that the fused spines in ‘Cheliderpeton’ lellbachae can rather be attributed to failure of segmentation during early embryogenesis. Defects of segmentation do not only involve the complete vertebrae or the centra producing block vertebrae (as described in Gerrothorax

9

Malformations in Ancient Amphibians

F. Witzmann et al.

above), but might also affect the neural arches (Erol et al., 2004). To our knowledge, similar congenitally fused neural spines have not been described in a fossil vertebrate. Konishi et al. (2011), Fig. 8 reported and illustrated the presumably pathological fusion of the fourth to seventh presacral neural spines in the mosasaur Prognathodon overtoni. Generalized vertebral infection (e.g. in the mosasaur Platecarpus) is easily distinguished from congenitally derived spinous process fusion (Martin and Rothschild, 1989).

Discussion Development of congenital vertebral malformations in temnospondyls Witzmann (2007) suggested that hemivertebrae in the rhachitomous vertebral column might be the result of failure of ossification of one cartilaginous lateral half of an intercentrum, because ossification usually proceeded (a)

(b)

(c)

(d) (e)

Fig. 8. (a)–(d) Schematic reconstruction of parts of the vertebral columns of some of the temnospondyls described in this manuscript in ventral view. The preserved pathologies are held in light grey. Sutures or boundaries that are not visible in the specimens due to complete fusion are dashed. (a) Rhachitomous vertebrae of Platyoposaurus stuckenbergi with non-segmented hemivertebra. (b) Stereospondylous vertebrae of an undetermined metoposaurid with non-segmented hemivertebra. Note that the hemivertebrae are incarcerated in (a) and (b); thus, no scoliosis is produced. (c)–(d) Plagiosaurid vertebrae of Gerrothorax pulcherrimus. (c) Block vertebra fused with wedge vertebra anteriorly; the wedge vertebra is semi-segmented and not incarcerated and produces a scoliosis. (d) Block vertebra producing no scoliosis. Note that the centra of the block vertebrae in (c) and (d) are anteroposteriorly shortened due to impairment of longitudinal growth by fusion of the disc spaces. (e) ‘Anteropleural’ rhachitomous vertebra; the intercentrum is associated with the anteriorly neighboured pleurocentra and neural arch (redrawn from Shishkin, 1989), compared with ‘normal’ rhachitomous vertebra in Fig. 1a. Abbreviations: hic, intercentrum of a hemivertebra; ic, intercentrum; na, neural arch; par, parapophysis; pc, pleurocentrum; wc, centrum of wedge vertebra.

10

very slowly in centra and long after the ossification of the neural arches in temnospondyl ontogeny (Boy, 1974; Schoch and Witzmann, 2009a,b). This assumption was based on a capitosauroid vertebral fragment consisting of fused intercentra with an intercalated hemivertebral intercentrum, but the neural arches were not preserved. However, failure of the neural arch (which ossifies early in temnospondyl ontogeny) to form in the hemivertebra of Platyoposaurus and in the larval Sclerocephalus specimen shows that the defect must have occurred early in embryogenesis of early tetrapods, before chondrification and is not a defect of ossification. The congenital vertebral pathologies described here thus show that defects of formation and segmentation in the tetrapod vertebral column represent a fundamental failure of somitogenesis before chondrification and ossification of the vertebral anlagen and can be followed throughout tetrapod evolution. This is irrespective of the type of vertebra that is affected, that is, stereospondylous, rhachitomous or plagiosaurid. All components of the vertebra can be involved (intercentrum, pleurocentrum, neural arch), either together or independently. Consequences of the described vertebral malformations for the living animals According to Kaplan et al. (2005), a mixture of defects of formation and of segmentation is often evident in humans and may produce quite complex malformations in one individual. This is also the case for the described hemivertebrae of Platyoposaurus, for the Algarve metoposaurid and for the wedge vertebra in specimen A of Gerrothorax. In contrast, the malformations in Gerrothorax specimen B and ‘Cheliderpeton’ lellbachae are solely defects of segmentation, and the malformation in the larval Sclerocephalus is solely a defect of formation. Congenital abnormalities of the spine are frequently associated with defects in the urogenital, pulmonary and cardiac systems (Kaplan et al., 2005). However, the individuals described here were probably not severely affected by their vertebral malformations, because the sizes of their vertebrae suggest that they were quite large-grown adults, and only the Sclerocephalus specimen is a small larva. However, the lake sediments of the Saar-Nahe Basin have yielded hundreds of specimens of larval Sclerocephalus and other temnospondyl amphibians (Schoch and Witzmann, 2009a), and the hemivertebra in this specimen did not cause scoliosis. Thus, it can be assumed that this malformation did not cause the death of this individual. Similarly, the hemivertebrae in Platyoposaurus and the Algarve metoposaurid did not cause scoliosis, and this was also the case in Gerrothorax specimen B showing a block vertebra, whereas the lateral curvature of © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

the column in Gerrothorax specimen A caused by a wedge vertebra was only slight (Fig. 8a–d). (As we do not know the complete vertebral columns of these specimens except for Sclerocephalus, we cannot say whether further wedge or hemivertebrae were actually present in each of these individuals.) In all described specimens except for Sclerocephalus, this apparently did not represent a severe disadvantage for these individuals, as these temnospondyls were no axial swimmers. Gerrothorax is interpreted as a bottomdwelling ambush predator, with a dorsoventrally flattened trunk that was stiffened by heavy dorsal and ventral osteoderms (Hellrung, 2003). Basal stereospondylomorphs like Platyoposaurus or ‘Cheliderpeton’ have a long, powerful swimming tail, whereas the trunk was stabilized by heavy ribs with large flanges and processes (Fig. 1d,e). Sulej (2007) reported a rather stiff presacral vertebral column but a flexible tail in metoposaurids. Thus, movement of the tail rather than of the trunk was responsible for drive during swimming in these forms. Of course, fusion of vertebrae might be disadvantageous in taxa that rely to a large extent on lateral undulations of the body for locomotion, the more so if several vertebrae are involved in fusion. On the other hand, fusion of certain parts of the vertebral column might also be of benefit and is characteristic for many tetrapod taxa. Thus, it is not always easy to decide whether the phenomenon of fused vertebrae is pathological or is an adaptation, for example, for mechanical strength. A number of extant and fossil tetrapods show fusion of neural spines and/or centra for stiffening of the trunk or to stabilize the pectoral and sacral regions. Among temnospondyls, this is evident in the dissorophid Astreptorhachis. The distal portions of trunk neural spines are fused together to stiffen the trunk, probably an adaptation for terrestrial locomotion (Vaughn, 1971). It is well known that pterosaurs have a notarium in which the spines of the anterior dorsal vertebrae are fused to support the shoulder girdle and to serve as attachment site of muscles of the foreleg (Wellnofer, 1983). In both ornithischians and saurischians, the sacral region may be stabilized by fused neural } and F€ spines and centra (e.g. Osi ozy, 2007; Sullivan et al., 2011). A mechanical function can be ruled out for the fused vertebrae described here, as this fusion is not known from any other individual of these taxa or their close relatives, and there is no obvious mechanical necessity to strengthen or stiffen the column in this region. Indication for resegmentation of rhachitomous vertebrae sensu Shiskin The pathological Platyoposaurus specimen described here also sheds light on Shishkin’s (1987, 1989) hypothesis of © 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

Malformations in Ancient Amphibians

resegmentation of rhachitomous vertebrae in temnospondyls. The intercentrum was topographically associated (and sometimes co-ossified) with the preceding (anterior) pleurocentra in the rhachitomous vertebrae of numerous Palaeozoic temnospondyls. Shiskin referred this mechanical association between posterior inter- and anterior pleurocentra as ‘anteropleural’ (Fig. 8e). Shishkin (1989), however, did not deny the presence of the ‘normal’ association of inter- and pleurocentrum in rhachitomous vertebrae of certain temnospondyls, with the pleurocentra being associated with the anterior intercentrum, as shown in Fig. 1a. He thus stated the presence of two alternative conditions of central element association in adult temnospondyls: the ‘normal’ and the ‘anteropleural’ conditions. According to him, the ‘anteropleural’ situation is the plesiomorphic condition in temnospondyls and can also be demonstrated in tetrapodomorph fishes. As indicated by the intersegmental position of the ribs, the ‘anteropleural’ centra are intrasegmental, whereas the ‘normal’ centra are intersegmental and thus resegmented (Shishkin, 1987, 1989). In the pathological specimen of Platyoposaurus described here, the left, normally developed pleurocentrum is co-ossified with the intercentrum posterior to it, whereas it is separated from the anterior intercentrum by a broad gap (Fig. 3a). This condition, which would have been not preserved in a ‘healthy’ rhachitomous vertebra (because of the generally cartilaginous connections of the vertebral components), is clearly ‘anteropleural’ sensu Shishkin (1989) and might support his hypothesis of the occurrence of two different conditions in rhachitomous vertebrae. Comparison with the Holocene record of vertebral anomalies Examination of frog vertebrae from the Hiscock site (Rothschild and Laub, in press), a Paleoindian archaeological excavation in western New York (United States) dated at 9000 years before present, revealed only six congenital vertebral anomalies. This represented examined 0.2% of bones, a frequency indistinguishable from that noted in temnospondyls. Conclusions 1. Defects of formation (hemi- and wedge vertebra) and segmentation (block vertebra) can be found in the vertebral column of Palaeozoic and Mesozoic amphibians. The wedge vertebra and congenital block vertebra described here are the oldest known occurrences of these malformations in the fossil record. 2. The vertebral malformations of ancient amphibians occur in rhachitomous, stereospondylous and plagiosaurid

11

Malformations in Ancient Amphibians

vertebrae and can affect all components of the vertebrae, either together or independently on their own. 3. Although vertebral ontogenies and morphologies of ancient amphibians have no extant analogue among tetrapods, the malformations found here can be attributed to the same underlying factors as in extant tetrapods including humans, that is, fundamental failure of somitogenesis caused by genes or environmental factors. 4. Given the quite rare prevalence of congenital vertebral malformations in humans (e.g. the occurrence of hemivertebrae is estimated at 5–10 in 10 000 births, WynneDavies, 1975), the congenital malformations of the vertebral column in ancient amphibians could be more frequent if one considers the small sample size of investigated specimens. One may speculate that this might be an indication that ancient amphibians were more susceptible to the underlying genetic or environmental factors resulting in disrupted somitogenesis. Large-sample-size studies of fossil amphibians have to be carried out to confirm or reject this hypothesis. 5. The close topographical association of the intercentrum with the anteriorly neighboured pleurocentrum in Platyoposaurus, as shown by the pathological co-ossification in specimen PIN 164/71 examined here, suggests the presence of ‘anteropleural’ rhachitomous vertebrae as outlined in Shishkin’s (1987, 1989) hypothesis of resegmentation. 6. The frequency of vertebral congenital malformations in amphibians appears unchanged from the Holocene. Acknowledgements We thank Rainer Schoch (Staatliches Museum f€ ur Naturkunde Stuttgart) for access to the collection under his care and the two anonymous reviewers for their thorough work. References Anderson, J. S., R. R. Reisz, D. Scott, N. B. Fr€ obisch, and S. S. Sumida, 2008: A stem batrachian from the Early Permian of Texas and the origin of frogs and salamanders. Nature 453, 515–518. Baur, G., 1891: On intercalation of vertebrae. J. Morphol. 4, 329–336. Boy, J. A., 1974: Die Larven der rhachitomen Amphibien (Amphibia: Temnospondyli; Karbon – Trias). Pal€aont. Z. 48, 236–282. Bywaters, E. G. L., and S. Evans, 1982: The Lumbar Interspinous Bursae and Baastrup’s Syndrome. Rheumatol. Int. 2, 87–96. Erol, B., K. Kusumi, J. Lou, and J. P. Dormans, 2002: Etiology of congenital scoliosis. Univ. Penn. Orthop. J. 15, 37–42. Erol, B., M. R. Tracy, J. P. Dormans, E. H. Zackai, M. K. Maisenbacher, M. L. O’Brien, P. D. Turnpenny, and

12

F. Witzmann et al.

K. Kusumi, 2004: Congenital scoliosis and vertebral malformations: characterization of segmental defects for genetic analysis. J. Pediatr. Orthop. 24, 674–682. Gubin, Y. M., 1991: Permian Archegosauroid Amphibians of the USSR. Moscow: Nauka. [In Russian]. Hellrung, H., 2003: Gerrothorax pustuloglomeratus, ein Temnospondyle (Amphibia) mit kn€ ocherner Branchialkammer aus dem Unteren Keuper von Kupferzell (S€ uddeutschland). Stuttg Beitr Natkd Ser B. 330, 1–130. Janensch, W., 1934: Eine halbseitige u¨berza¨hlige Wirbelbildung bei einem Dinosaurier. Sitzungsber. Ges. Naturf. Freunde. Berlin. 1933, 458–462. Jeffery, N. D., P. M. Smith, and P. E. Talbot, 2007: Imaging findings and surgical treatment of hemivertebrae in three dogs. J. Am. Vet. Med. Assoc. 230, 532–536. Johnson, G. D., 1988: An abnormal captorhinomorph vertebra from the Lower Permian of North-Central Texas. J. Vert. Paleontol. 8(Suppl), 19A. Kacki, S., S. Villotte, and C. J. Kn€ usel, 2011: Baastrup’s sign (kissing spines): a neglected condition in paleopathology. Int. J. Paleopathol. 1, 104–110. Kaplan, K. M., J. M. Spivak, and J. A. Bendo, 2005: Embryology of the spine and associated congenital abnormalities. Spine J. 5, 564–576. Konishi, T., D. Brinkman, J. A. Massare, and M. W. Caldwell, 2011: New exceptional specimens of Prognathodon overtoni (Squamata, Mosasauridae) from the upper Campanian of Alberta, Canada, and the systematics and ecology of the genus. J. Vert. Paleontol. 31, 1026–1046. Kr€atschmer, K., 2006: Neue temnospondyle Amphibien aus dem Rotliegend des s€ udwestdeutschen Saar-Nahe-Beckens, Teil 1. Geowiss. Beitr. Saarpf. Rotl. 4, 3–46. € Kuhn, O., 1942: Uber Cyclotosaurus hemprichi Kuhn und einige weitere Tetrapodenreste aus dem Keuper von Halberstadt. Beitr Geol Th€ ur. 6, 181–202. Lydekker, R., 1889: Catalogue of the Fossil Reptilia and Amphibia in the British Museum. Part II. Containing the Orders Ichthyopterygia and Sauropterygia. London: British Museum of Natural History. Marjanovic, D., and M. Laurin, 2008: A reevaluation of the evidence supporting an unorthodox hypothesis on the origin of extant amphibians. Contrib. Zool. 77, 149–199. Martin, L. D., and B. M. Rothschild, 1989: Paleopathology and diving mosasaurs. Am. Sci. 77, 460–467. McMaster, M. 2001: Congenital scoliosis. In: The Pediatric Spine. Principles and Practice (S. L. Weinstein, ed). Philadelphia: Lippincott Williams and Wilkins. pp. 161–177. McMaster, M., and K. Ohtsuka, 1982: The natural history of congenital scoliosis. A study of two hundred and fifty-one patients. J. Bone Joint Surg. Am. 64, 1128–1147. Moulton, J. M., 1974: A description of the vertebral column of Eryops based on the notes and drawings of A. S. Romer. Breviora 428, 1–44. Moura, E., S. M. Cirio, and J. A. Villanova, Jr, 2010: VACTERL association in a cat. Am. J. Med. Genet. 152A, 777–780.

© 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

F. Witzmann et al.

} A., and I. F€ Osi, ozy, 2007: A maniraptoran (Theropoda, Dinosauria) sacrum from the Upper Cretaceous of the Hateg Basin (Romania) in search of the lost pterosaurs of Baron Franz Nopcsa. Neues. Jahrb. Geol. Palaontol. Abh. 246, 173–181. Panchen, A. L., 1959: A new armoured amphibian from the Upper Permian of East Africa. Philos. Trans. R. Soc. Lond. B Biol. Sci. 242, 207–281. Panchen, A. L., 1977: The origin and early evolution of tetrapod vertebrae. In: Problems in Vertebrate Evolution. (S. M. Andrews, R. S. Miles and A. D. Walker, eds). London: Linnean Society of London. pp 289–318. Pourquie, O., and K. Kusumi, 2001: When body segmentation goes wrong. Clin. Genet. 60, 409–416. Resnick, D., 1985: Degenerative diseases of the vertebral column. Radiology 156, 3–14. Rothschild, B. M., and R. S. Laub: Epidemiology of anuran pathology in the holocene component of the Hiscock site: rare or not survived. J. Herpetol (in press). Rothschild, B. M., and L. R. Martin, 2006: Skeletal impact of disease. Bull. New Mexico Mus. Natl Hist. 33, 1–226. Rothschild, B. M., H.-P. Schultze, and R. Pellegrini, 2012: Herpetological Osteopathology. Annotated Bibliography of Amphibians and Reptiles. Heidelberg: Springer. Rothschild, B. M., L. D. Martin, J. Babiarz, and D. A. Berman, in press: Hemivertebrae as pathology and as a window to Oligocene behavior. Internet J. Vert. Paleontol. Ruta, M., and M. I. Coates, 2007: Dates, nodes and character conflict: addressing the lissamphibian origin problem. J. Syst. Palaeontol. 5, 69–122. Schoch, R. R., 1999: Comparative osteology of Mastodonsaurus giganteus (Jaeger 1828) from the Middle Triassic (Lettenkeuper: Longobardian) of Germany (Baden-W€ urttemberg, Bayern, Th€ uringen). Stuttg. Beitr. Natkd. Ser B 278, 1–175. Schoch, R. R., 2009: Evolution of life cycles in early amphibians. Annu. Rev. Earth Planet. Sci. 37, 135–162. Schoch, R. R., and F. Witzmann, 2009a: Osteology and relationships of the temnospondyl genus Sclerocephalus. Zool. J. Linn. Soc. 157, 135–168. Schoch, R. R., and F. Witzmann, 2009b: The temnospondyl Glanochthon from the Lower Permian Meisenheim Formation of Germany. Special Papers Palaeontol. 81, 121–136. Schoch, R. R., and F. Witzmann, 2012: Cranial morphology of the plagiosaurid Gerrothorax pulcherrimus as an extreme example of evolutionary stasis. Lethaia 45, 371–385. Shawen, S. B., P. J. Belmont, T. R. Kuklo, B. D. Owens, K. F. Taylor, R. Kruse, and D. W. Polly, 2002: Hemimetameric

© 2013 Blackwell Verlag GmbH Anat. Histol. Embryol.

Malformations in Ancient Amphibians

segmental shift: a case series and review. Spine 27, 539–544. Shishkin, M. A., 1987: The Evolution of Ancient Amphibians. Moscow: Nauka [in Russian with English summary]. Shishkin, M. A., 1989: The axial skeleton of early amphibians and the origin of resegmentation in tetrapod vertebrae. Fortschr. Zool. 35, 180–195. Sigurdsen, T., and D. M. Green, 2011: The origin of modern amphibians: a re-evaluation. Zool. J. Linn. Soc. 162, 457–469. Sulej, T., 2007: Osteology, variability, and evolution of Metoposaurus, a temnospondyl from the Late Triassic of Poland. Palaeontologia Polonica. 64, 29–139. Sullivan, R. M., S. G. Lucas, S. E. Jasinski, and D. H. Tanke, 2011: An unusual sacral neural spine osteopathy of a chasmosaurine (Dinosauria: Ceratopsidae) from the Upper Cretaceous Kirtland Formation (Hunter Wash member), San Juan Basin, New Mexico. Bull New Mexico Mus Natl Hist. 53, 484–488. Vaughn, P. P., 1971: A Platyhystrix-like amphibian with fused vertebrae, from the Upper Pennsylvanian of Ohio. J. Paleontol. 45, 464–469. Warren, A. A., and N. Snell, 1991: The postcranial skeleton of Mesozoic temnospondyl amphibians: a review. Alcheringa. 15, 43–64. Wellnofer, P., 1983: A pterosaurian notarium from the Lower Cretaceous of Brazil. Pal€aont. Z. 57, 147–157. Witzmann, F., 2007: A hemivertebra in a temnospondyl amphibian: the oldest record of scoliosis. J. Vert. Paleontol. 27, 1043–1046. Witzmann, F., and T. Gassner, 2008: Metoposaurid and mastodonsaurid stereospondyls from the Triassic-Jurassic boundary of Portugal. Alcheringa. 32, 37–51. Witzmann, F., P. Asbach, K. Remes, O. Hampe, A. Hilger, and A. Paulke, 2008: Vertebral pathology in an ornithopod dinosaur: a hemivertebra in Dysalotosaurus lettowvorbecki from the Jurassic of Tanzania. Anat. Rec. 291, 1149–1155. Witzmann, F., H. Scholz, J. M€ uller, and N. Kardjilov, 2010: Sculpture and vascularization of dermal bones, and the implications for the physiology of basal tetrapods. Zool. J. Linn. Soc. 160, 302–340. Wong, D. M., W. K. Scarratt, and J. Rohleder, 2005: Hindlimb paresis associated with kyphosis, hemivertebrae and multiple thoracic vertebral malformations in a Quarter horse gelding. Equine. Vet. Educ. 17, 187–194. Wynne-Davies, R., 1975: Congenital vertebral anomalies: aetiology and relationship to spina bifida cystica. J. Med. Genet.. 12, 280–288.

13