Embryonic Development of the Axial Column in the Little Skate, … · 2019-02-08 · Embryonic Development of the Axial Column in the Little Skate, Leucoraja erinacea Katharine E.

Embryonic Development of the Axial Columnin the Little Skate, Leucoraja erinacea

Katharine E. Criswell,1,2* Michael I. Coates,1 and J. Andrew Gillis2,3

1Department of Organismal Biology and Anatomy, University of Chicago, Chicago, Illinois2Marine Biological Laboratory, Woods Hole, Massachusetts3Department of Zoology, University of Cambridge, Cambridge, UK

ABSTRACT The morphological patterns and molecularmechanisms of vertebral column development are wellunderstood in bony fishes (osteichthyans). However, verte-bral column morphology in elasmobranch chondrichthyans(e.g., sharks and skates) differs from that of osteichthyans,and its development has not been extensively studied.Here, we characterize vertebral development in an elasmo-branch fish, the little skate, Leucoraja erinacea, usingmicroCT, paraffin histology, and whole-mount skeletalpreparations. Vertebral development begins with the con-densation of mesenchyme, first around the notochord, andsubsequently around the neural tube and caudal arteryand vein. Mesenchyme surrounding the notochord differ-entiates into a continuous sheath of spindle-shaped cells,which forms the precursor to the mineralized areolar calci-fication of the centrum. Mesenchyme around the neuraltube and caudal artery/vein becomes united by a popula-tion of mesenchymal cells that condenses lateral to thesheath of spindle-shaped cells, with this mesenchymalcomplex eventually differentiating into the hyaline carti-lage of the future neural arches, hemal arches, and outercentrum. The initially continuous layers of areolar tissueand outer hyaline cartilage eventually subdivide into dis-crete centra and arches, with the notochord constricted inthe center of each vertebra by a late-forming “inner layer”of hyaline cartilage, and by a ring of areolar calcificationlocated medial to the outer vertebral cartilage. The verte-brae of elasmobranchs are distinct among vertebrates,both in terms of their composition (i.e., with centra consist-ing of up to three tissues layers—an inner cartilage layer, acalcified areolar ring, and an outer layer of hyaline carti-lage), and their mode of development (i.e., the subdivisionof arch and outer centrum cartilage from an initially con-tinuous layer of hyaline cartilage). Given the evident varia-tion in patterns of vertebral construction, broad taxonsampling, and comparative developmental analyses arerequired to understand the diversity of mechanisms atwork in the developing axial skeleton of vertebrates. J.Morphol. 278:300–320, 2017. VC 2017 Wiley Periodicals, Inc.

KEY WORDS: elasmobranch; vertebral column; develop-ment; microCT; skate

RESEARCH HIGHLIGHTS

� Skate vertebrae form via a continuous condensa-tion of mesenchyme that differentiates into carti-lage and later subdivides.

� Vertebral centra consist of three layers: an innercartilage, a middle areolar calcification, and anouter hyaline cartilage.

INTRODUCTION

The vertebral column is a key component of thevertebrate skeleton, providing structural supportfor the skull and appendicular skeleton, as well asprotection for the spinal cord and axial blood ves-sels. A vertebral column can include a notochordand/or series of centra, neural, and hemal archesand spines, zygapophyses to articulate one vertebrato another, and transverse processes, perhaps artic-ulating with ribs (Fig. 1). Vertebrae are a definingfeature of the vertebrate clade, but different compo-nents of the vertebral skeleton, namely the archesand centra, have evolved independently of oneanother, are variable in their construction, andmight therefore be patterned by separate develop-mental mechanisms (Arratia et al., 2001). To gaininsight into the developmental and evolutionarybasis of axial column morphological variation, broadtaxon sampling of developmental and anatomical

Contract grant sponsors: Henry Hinds Fund for Graduate Stu-dent Research in Evolutionary Biology, a Company of BiologistsTravelling Fellowship, a Marine Biological Laboratory/University ofChicago Graduate Research Award, and a National ScienceFoundation Doctoral Dissertation Improvement (to K. E. C.); Con-tract grant number: DEB-1501749; Contract grant sponsor:National Science Foundation (to M. I. C.); Contract grant number:DEB-0917922; Contract grant sponsor: Royal Society UniversityResearch Fellowship (J. A. G.); Contract grant number:UF130182; Contract grant sponsors: Plum Foundation John E.Dowling and Laura and Arthur Colwin Endowed Summer ResearchFellowships at the Marine Biological Laboratory, and by a grantfrom the University of Cambridge Isaac Newton Trust [14.23z].

*Correspondence to: Katharine E. Criswell, Department ofOrganismal Biology and Anatomy, University of Chicago, Chicago,IL. E-mail:[email protected]

Received 3 August 2016; Revised 11 November 2016;Accepted 29 November 2016.

Published online 31 January 2017 inWiley Online Library (wileyonlinelibrary.com).DOI 10.1002/jmor.20637

VC 2017 WILEY PERIODICALS, INC.

JOURNAL OF MORPHOLOGY 278:300–320 (2017)

data within vertebrates, and particularly, withinjawed vertebrates (gnathostomes), is needed.However, at present, most data on vertebral columndevelopment have been collected from a handful ofosteichthyan taxa (i.e., bony fishes and tetrapods)(Fleming et al., 2015). There is a notable paucity ofdevelopmental data for the vertebral columns of chon-drichthyans (cartilaginous fishes—sharks, skates,rays, and holocephalans), and this results in a biasedand incomplete picture of vertebral column evolutionwithin gnathostomes and vertebrates as a whole.

The onset of vertebral development begins in asimilar fashion in all taxa studied to date, when asubpopulation of mesodermal cells undergoes anepithelial to mesenchymal transformation from theventral somites, and migrates medially as sclero-tome (Gadow and Abbott, 1895; Christ and Wilting,1992). These sclerotomal cells are thought to con-dense at dorsolateral and ventrolateral positionsrelative to the notochord, and form the basidorsal(neural arch) and basiventral (hemal arch) rudi-ments. After sclerotome migration and condensa-tion, vertebral development, and centrum formationin particular, is more varied. Depending on the tax-on, centra may arise via a combination of four gener-al processes: 1) Through arch primordia encirclingthe notochord (arcocentra); 2) through sclerotomalcells that invade and chondrify within the notochordsheath (chordacentra); 3) through chondrocytes

that differentiate around the notochord sheath (hol-ocentra); or 4) through direct ossification outsidethe notochord sheath (autocentra) (Arratia et al.,2001).

The developmental mechanisms and tissuesunderlying vertebral formation differ greatlybetween different vertebrate clades. In teleost fishes(actinopterygians) the vertebrae are composed ofacellular bone, in which bone-forming osteoblastsare not incorporated into the skeletal tissue afterbone matrix is secreted (Hall, 2005). Centra inzebrafish (Danio rerio) form before the arches, bynine days post fertilization, and vertebrae ossifydirectly without cartilaginous precursors (Fleminget al., 2004; Bensimon-Brito et al., 2012). The archesform from sclerotomal cells that condense both dor-sal to and ventral to the notochord, but the centraform from bone matrix that seems to be deposited bythe notochord itself (Fleming et al., 2004). In Atlan-tic salmon (Salmo salar), neural and hemal archcartilages form first and are then encased in bone,with centra developing later through four layers ofmineralization. These four layers include the chor-dacentrum, which mineralizes within the notochordsheath, a thin, continuous perinotochordal layerthat covers the notochord, an outer layer of lamininbone making up each centrum, and lastly, cancel-lous bone with longitudinal trabeculae (Nordviket al., 2005; Wang et al., 2013). As a third example,chordacentra in the Japanese medaka (Oryziaslatipes) develop within the notochord sheath prior tobeing surrounded by perichordal bone (Ekanayakeand Hall, 1987).

In tetrapods (salamander, chick, and mouse), ver-tebrae have been documented to form entirely frommigrating sclerotomal cells that condense around thenotochord to form a perichordal tube of tissue thatthen differentiates into individual vertebral units(Bagnall et al., 1988; Christ and Wilting, 1992; Pie-karski and Olsson, 2014). In chick (Gallus gallus),the ventral portion of the somite has been fatemapped to specific vertebral components, with themedial-most cell population contributing to the cen-tra, the dorsal sclerotomal cells giving rise to neuralspines, and the lateral portion of the sclerotome mak-ing up the remainder of the neural arches and ribs(Bagnall et al., 1988; Christ et al., 2000). In the axo-lotl (Ambystoma mexicanum), a urodele amphibian,fate mapping of somites three to five shows somiticcontributions to all vertebral parts, including thecentra, arches, and intervertebral discs (Piekarskiand Olsson, 2014). This variation in vertebral devel-opment across vertebrates highlights the need formore data from the sister group of osteichthyans, thecartilaginous fishes (chondrichthyans).

Extant chondrichthyans belong to one of two line-ages: the holocephalans and the elasmobranchs.Extant holocephalans comprise the chimaeroids, asmall group of largely deep-sea fishes, while extantelasmobranchs include skates and rays (collectively

Fig. 1. The ideal, typical vertebra redrawn from Owen (1848),showing the neural and hemal arches and spines, transverseprocesses, and centrum. n, neural canal; h, hemal canal.

301EMBRYONIC DEVELOPMENT OF THE AXIAL COLUMN

Journal of Morphology

known as batoids) and sharks. Vertebral columnanatomy in adult elasmobranchs differs markedlyfrom that of osteichthyans. The elasmobranch ver-tebral column consists of a zipper-like series of neu-ral and intercalary arches dorsally, along withcentrum cartilages that support transverse process-es precaudally and hemal arches caudally (see Fig.1; Daniel, 1922). A series of mediolaterally flattenedneural spines sits in between vertebral boundarieson the dorsal-most surface of the vertebral columnthroughout the length of the body. The arches con-sist of a core of hyaline cartilage covered by amosaic of blocks of calcified tissue called tesserae inadults (Dean and Summers, 2006). Most elasmo-branchs have mineralized chordacentra (centraformed within the notochord sheath), in which cellsdifferentiate into dense rings within the fibrousnotochord sheath and are surrounded by apatite intheir intercellular spaces, forming a highly cellular,netlike mineralized tissue called areolar calcifica-tion (Gadow and Abbott, 1895; Applegate, 1967;Dean and Summers, 2006). These centra take onan hourglass shape, expanding anteriorly andposteriorly, and constrict the notochord consider-ably at their center, with remnants of the notochordpersisting at the intervertebral boundaries. Caudaldiplospondyly, a condition in which two vertebraecorrespond to each myomere and set of spinalnerves, is well known in many sharks (Ridewood,1899), but occurs in batoids as well.

Within elasmobranchs, vertebral morphology isvariable. In the broadnose seven-gill shark(Notorynchus cepedianus), the notochord is sur-rounded by an unsegmented cartilaginous tubeand supports thin, mineralized ring centra thatpersist into adulthood (Daniel, 1922). In the por-beagle shark (Lamna nasus), the hammerhead(Sphyrna blochii), the basking shark (Cetorhinusmaximus), and the guitarfish (Rhinobatos produc-tus), heavily mineralized centra form within thefibrous notochord sheath and expand throughoutontogeny to engulf the neural and hemal arches(Ridewood, 1921; Daniel, 1922; Goodrich, 1930). Inmany elasmobranchs, such as the horn shark(Heterodontus francisci), small-spotted catshark(Scyliorhinus canicula), little skate (Leucorajaerinacea), and some rays (e.g., Myliobatis aquila),the cartilage of the arches expands dorsally andventrally to envelop the calcified centra (Hasse,1879; Daniel, 1922).

Current understanding of elasmobranch vertebraldevelopment derives largely from historical studieson sharks, with only sparse data from skates andrays (for shark data see Gegenbaur, 1872; Hasse,1879; Klaatsch, 1893a,b, 1895; Gadow and Abbott,1895; Goodrich, 1958; and see Klaatsch, 1893b for adiscussion of vertebral development in the electricray Torpedo ocellata). These studies demonstratedthat the initiation of shark vertebral developmentproceeds, as in osteichthyans, with the migration of

sclerotomal cells toward the notochord sheath. Thenotochord sheath thickens into a ring of fibrous,spindle-shaped cells, surrounded by a thin mem-brane, the elastica externa. Sclerotomal mesenchymeis thought to both condense around the notochord,forming cartilaginous units that give rise to the neu-ral and hemal arches, and migrate through the perfo-rated elastica externa and into the fibrous sheathwhere it differentiates into the cartilage of thecentrum.

Historically, sharks have been used as represen-tatives of generalized, and by implication ancestral,gnathostome anatomical conditions (e.g., Goodrich,1930). As a result, the classic, and most used, modelof vertebral construction, based on vertebral devel-opment in sharks, was widely applied to gnathos-tomes as a whole (Gadow and Abbott, 1895; Gadow,1933). It suggests that for all gnathostomes eachvertebral unit is composed of four pairs of embryoniccartilages, or “arcualia”: two pairs of dorsal ele-ments (basidorsals and interdorsals) and two pairsof ventral elements (basiventrals and interventrals;Gadow and Abbott, 1895). The basidorsals form theadult neural arches and, when present, the dorsalintercalary arches, and the basiventrals form thehemal arches and ventral intercalary arches. Sever-al key works have pointed out that the “ArcualiaTheory” is not compatible with published evidencefor osteichthyan (Schaeffer, 1967) and acanthodianfishes (Miles, 1970) and the theory has been rejectedoutright for tetrapods (Williams, 1959). Despite thisapparent incongruity, the simplistic arcualia model,and its associated terminology, remains influentialand is often used to describe the vertebrae of non-tetrapod gnathostomes in the paleontological andcomparative anatomical literature (Jarvik, 1980a,b;Arratia et al., 2001). Additionally, it is unclearwhether all elasmobranchs develop according toGadow and Abbott’s (1895) model or whether someclades, such as batoids, deviate from this pattern.

It is clear from the bias of developmental studiestowards osteichthyans (zebrafish, chick, and mouse),the lack of studies on vertebral column developmentin cartilaginous fishes utilizing modern microtomo-graphic (microCT) techniques, and the confusionover current embryological terminology and models,that tests of historical hypotheses and a renewedfocus on vertebral column evolution are needed. Toreconstruct sequences of character evolution for thegnathostome vertebral column, that is, to determinewhich developmental processes are general for gna-thostomes, and to distinguish between homologousand homoplastic vertebral structures, it is necessaryto collect morphological and developmental datafrom numerous osteichthyan and chondrichthyantaxa. Additionally, an assessment of whether the pro-cesses of vertebral development that have beendescribed in sharks are representative for all elasmo-branchs requires data on axial column developmentin batoids. Here, we use computed tomography,

302 K.E. CRISWELL ET AL.


histology, and whole mount skeletal preparation tovisualize the embryonic development of the axial col-umn in a batoid elasmobranch, the little skate Leu-coraja erinacea. Crucially, the use of high-resolutionmicroCT scans allows for interpretations of detailedanatomy in three dimensions that have not previous-ly been possible. The little skate is an emerging mod-el for studies of gnathostome development andevolution, and as such, the methods employed herewill be broadly useful to researchers studying otheranatomical systems.

MATERIALS AND METHODSAnimal Collection, Husbandry, and Fixation

Developmental series of the little skate, Leucoraja erinacea(Mitchill, 1825), from Stages 27 to 34, were obtained from theMarine Resources Center (MRC) at the Marine Biological Labo-ratory in Woods Hole, MA. Embryonic stages correspond tothose described in the staging table of the winter skate, Leucor-aja ocellata (Maxwell et al., 2008). Little skate embryos reachStage 27 after approximately four weeks of development at188C and have developed small pectoral and pelvic fins. ByStage 34, after approximately 5 months of development, embry-os are close to hatching, with dark pigmentation and littleexternal yolk remaining. Embryos were maintained at approxi-mately 188C in seawater while at the MRC, until Stage 27, andwere then maintained at 158C in reconstituted Instant Ocean(Aquarium Systems) on at 12-h light-dark cycle. On reachingthe appropriate developmental stage, embryos were removedfrom their keratinous egg cases using a razor blade and eutha-nized with an overdose of MS-2222 (Ethyl 3-aminobenzoatemethanesulfonate—Sigma-Aldrich) (1 g/l bath). Embryos werethen fixed for 24–48 h in 4% paraformaldehyde, rinsed threetimes in 1X phosphate buffered saline (3 3 5 min), and gradedinto either 100% methanol (MeOH) or ethanol (EtOH) throughan ascending series (25%, 50%, 75%, and 100% in 1X PBS,5min/rinse) and stored at 2208C.

MicroCT Scanning

Embryos of all stages were stained with Iodine PotassiumIodide (IKI) (2% w/v iodine and 1% w/v potassium iodide inwater), diluted to 10% in water, for three to five days to providecontrast to soft tissues. Two Stage 31 embryos were stainedwith 5% w/v phosphomolybdic acid (PMA) in MeOH for fivedays. Both IKI and PMA provide contrast to soft tissues, butPMA is actually taken up by cartilage, and presents as aneven, mid-range gray tone in the CT slices.

Embryos at Stages 28–30 and 32–33 were microCT scannedat the University of Texas at Austin CT Laboratory (UTCT)using an XRadia microXCT 400 scanner. Embryos at Stages30–34 were microCT scanned in the Department of OrganismalBiology at the University of Chicago, using a General ElectricPhoenix v|tome|x 240 MicroCT scanner with two X-ray tubesand 160 3 160 detector panel of 2,048 3 2,048 pixel resolution(see Table 1 for scanning parameters). The 180 kV 15W high-power NanoFocus tube achieves a maximum voxel resolution of0.5 lm and can scan organisms in a size range of 15 cm inlength by 12 cm in diameter. Resulting scans were segmentedusing Mimics versions 17.0 and 18.0 (Materialise, Leuven,Belgium).

Skeletal Preparations

A second series of embryos, corresponding to the stages dis-cussed above, were cleared and stained with Alcian blue to visual-ize the proteoglycan-rich matrix present in vertebral cartilage.This protocol was based on a standard clearing and staining proto-col for fishes (Klymkowsky and Hanken, 1991) that was optimized

for use on elasmobranch fishes (Gillis et al., 2009). Alcian blue-stained embryos were imaged using a Leica MZFLIII microscopewith a Nikon D5000 camera, and the Camera Control Pro soft-ware. White balance was adjusted in Adobe Photoshop CS6 usingthe levels tool; no other editing was performed.

Paraffin Histology and Staining

A third series of embryos (two specimens per stage) was paraf-fin embedded and sectioned for histochemical staining withHematoxylin, Eosin, and Alcian blue (HEA). Hematoxylin is apositively charged, basic dye that stains acidic structures, such ascell nuclei containing DNA and RNA, with a dark violet to bluishcolor (Fischer et al., 2008). Eosin is a negatively charged, acidicdye that stains basic tissues that include proteins, such as cyto-plasm and sometimes extracellular matrix, pink. As mentionedpreviously, Alcian blue was used to identify proteoglycan-richextracellular matrix in section. Older embryos with heavily calci-fied vertebrae and dermal denticles (Stage 32 and above) weredemineralized in 10% ethylenediaminetetraacetic acid (EDTA)for 14 days prior to embedding. The embedding process consistedof three 20-min rinses in histosol (National Diagnostics), followedby two 30-min rinses in a 50:50 paraffin:histosol solution at 678C.The tissue was then placed in 100% paraffin overnight at 678C,and during the following day the paraffin was changed five timesbefore the tissue was placed in embedding blocks and positioned.The tissue was sectioned in transverse or sagittal orientation at athickness of 8 mm using a Microm HM 330 microtome withThermo Fisher HP 35 coated blades.

HEA-staining was performed on sections according to establishedprotocols (Gillis et al., 2009). Sections were imaged with a Zeiss Axi-oplan microscope, Nikon D5000 camera, and Camera Control Prosoftware. White balance was adjusted in Adobe Photoshop CS6using the levels tool; no other editing was performed.

RESULTS

Here, we describe the embryonic morphology ofthe little skate vertebral column through a seriesof embryos, beginning with the earliest appearanceof recognizable vertebral tissue, and ending with

TABLE 1. MicroCT scanning parameters for skate embryosscanned and used in this study

Stage # Scanned Stain kV lA Timing Voxel size

28 2 IKI 60 167 3 s 2.9070 143 2 s 3.08

29 1 IKI 60 167 3 s 2.4730 4 IKI 100 100 2 s 2.63

70 143 3 s 2.9370 143 3 s 2.9360 167 2.5 3.08

31 3 PMA 100 100 2 s 2.95PMA 40 250 4 s 2.79IKI 110 70 2 s 3.12

32 1 PC2 C

IKI 1006060

100167167

2 s2 s2 s

2.584.164.16

33 2 C IKI 60 167 1.5 sec 4.9934 1 PC

1 CIKI 80

10070

1004 sec2 sec

3.763.08

One scan sampling each separate morphological state wasfigured here. Images from additional scans were not includedbecause they did not differ significantly in their morphologyfrom other scans of similarly staged embryos. C 5 caudalregion, IKI 5 iodine potassium iodide, kV 5 kilovolts, PC 5 pre-caudal region, PMA 5 phosphomolybdic acid, lA 5 microamps,voxel size in micrometers.



the fully differentiated vertebral skeleton. Follow-ing this sequential description, we provide a sum-mary of the patterns observed in skate and aschematic that tracks the different layers of verte-bral tissue through development.

Stage 27

At the start of Stage 27, loose mesenchymeappears to surround the notochord as a single layerof cells just peripheral to the notochord epithelium(Fig. 2a,a0). As development progresses additionalmesenchymal cells condense around the notochord,up to five cells thick, to form a continuous layer thatextends the length of the body (Fig. 2b,b0).

Stages 28–29

Incipient vertebral structures (neural and hemalarch mesenchyme) appear at Stage 28 and remainlargely unchanged through Stage 29. The notochordsheath consists of an inner layer that is approxi-mately two cells thick, surrounded by an outer layerof elongate, spindle-shaped cells arranged concen-trically around the notochord and forming a thick-ened fibrous tube (Fig. 3a–c0). The spindle-shapedcells surrounding the notochord are embedded inAlcian blue-stained (and therefore presumably pro-teoglycan-rich) extracellular matrix, and microCTscans show that both neural and hemal arch mesen-chyme and this layer of spindle-shaped cells aredeposited as unsegmented tubes along the entireanterior–posterior axis (Fig. 3d–f). Mesenchymalcells condense dorsal and ventral to the notochord,to form the nascent neural and hemal arches, but atthis point the mesenchyme has not yet reached thedorsal-most portion of the neural tube (Fig. 3g–k).Sagittal and horizontal sections reveal that thenotochord is unconstricted (Fig. 3f,l).

Precaudally, small lateral condensations of mes-enchyme, also continuous along the anterior–poste-rior axis of the notochord, begin to form transverseprocesses (Fig. 3c,c0). The dorsal aorta is present asa wide canal just ventral to the notochord. Dorsalroot ganglia are large and teardrop-shaped, spacedclose to one another, and have corresponding nervesextending posteroventrally down the length of theincipient vertebrae (Fig. 3d,e). Nervous tissue stainsbrightly with IKI in young embryos (Fig. 3b,h,l),providing clear markers of each axial segment, evenwhen the vertebral tissue is continuous.

Caudally, development is largely contemporaneouswith the precaudal region. Mesenchymal cells extendventrally and condense around the caudal artery andvein, but these cells are not spindle-shaped in appear-ance and the matrix surrounding these cells does notstain with Alcian blue. The dorsal-most and ventral-most extensions of condensed mesenchyme arethicker than the rest of the tissue, giving it a bulgingsquare shape in cross section (Fig. 3g,h,i). The incipi-ent neural arch tissue extends dorsally from the base,near the notochord, but does not yet fully enclose theneural tube, as in the precaudal region. The dorsalroot ganglia and spinal nerves are spaced fartherapart caudally than in the precaudal region (Fig. 3k),foreshadowing that each set of nerves will correspondwith two vertebrae in the caudal region, as opposed toa single vertebral unit in the precaudal region. Thecaudal artery and vein are large, with small foraminaforming in the hemal arch cartilage to admit bloodvessels (Fig. 3j,k). In Stages 28 and 29 wholemountcleared and stained specimens, Alcian blue-stainedcartilage is not yet visible (not shown).

Stage 30

In the precaudal region in Stage 30 skateembryos, the neural tube, dorsal root ganglia, and

Fig. 2. Leucoraja erinacea, embryonic vertebral morphology in Stage 27 embryos. a) cross section through an early HEA-stainedStage 27 embryo; a0) 203 magnification of cells surrounding the notochord in a; b) cross section through a late HEA-stained Stage27 embryo; b0) 203 magnification of mesenchymal cells condensing around the notochord in b. Icons in upper right corner indicateplane of section. Scale bars represent 50 mm for a and b, and 25 mm for a0 and b0.



Fig. 3. Leucoraja erinacea, embryonic vertebral morphology in Stages 28–29 embryos. a) anterior view of a CT reconstruction of theprecaudal vertebrae; b) transverse section through the IKI-stained CT scan; c) transverse section stained with HEA, showing the thick-ened layer of spindle-shaped cells; c0) 203 magnification of the layer of spindle-shaped cells; d) anterolateral view of the precaudal CTreconstruction; e) lateral view of the precaudal CT reconstruction; f) sagittal section through the precaudal CT scan; g) anterior view ofthe caudal CT reconstruction; h) transverse section of the caudal CT scan; i) transverse section through the caudal vertebrae stainedwith HEA; i0) 203 magnification of the layer of spindle-shaped cells; j) anterolateral view of the caudal CT reconstruction; k) lateral viewof the caudal CT reconstruction; l) sagittal section of the caudal CT scan. a/v, caudal artery and vein, da, dorsal aorta, drg, dorsal root gan-glion, ha, hemal arch, hsp, hemal spine, me, mesenchymal sheath, mr, motor nerve root, na, neural arch, nc, notochord, ncs, notochordsheath, nt, neural tube, tp, transverse process. Icons in upper right corner indicate plane of section. Scale bars represent 100 lm for mostimages; scale bars represent 50 lm for c0 and i0.



spinal nerves remain unchanged from previousstages (Fig. 4a–e). Vertebral tissues at Stage 30are very similar to those at Stage 29, thoughseveral subtle differences are observed. The sheathof spindle-shaped cells remains continuous, buthas increased in thickness (Fig. 4f). Hints offuture notochord constriction (and, eventually, ver-tebral boundaries) can be seen in sagittal and hor-izontal sections in both the precaudal and caudalregions as the notochord begins to segmentallydecrease in size (Fig. 4f–g). The morphology of thecaudal vertebrae is similar to that of the precaudalregion, with a thickened layer of spindle-shaped cells(Fig. 4h–k), but with spinal nerves that are morewidely separated, indicating future diplospondyly, inwhich each myotomal segment will eventuallyinclude two vertebrae instead of one (Fig. 4k–l).

Mesenchymal condensations surrounding theneural tube dorsally and caudal artery and veinventrally now meet at the dorsal and ventral mid-line, respectively, to fully enclose the axial column(Fig. 4c,m–n). At this stage, the tissue of the trans-verse processes has condensed, but remains undif-ferentiated (Fig. 4b,c,c0). Compared to Stage 29, thetissue connecting the nascent hemal arches to thehemal spines has thinned slightly (Fig. 4h,k–l).

Stage 31

During Stage 31, the condensed mesenchyme ofthe neural and hemal arches differentiates intocartilage and is continuous with the cartilageenveloping the notochord and its surrounding lay-er of spindle-shaped cells. The notochord continuesto decrease in diameter relative to the vertebrae,and boundaries between vertebral elements (cen-trum cartilages, neural, and intercalary arches)become apparent. Late in Stage 31, the cartilagesurrounding the notochord becomes restricted toeach vertebral segment, and thins at presumptivevertebral boundaries. This cartilaginous tuberemains continuous, but variations in thickness canbe seen on the outer surface of the centrum carti-lage in CT reconstructions (Figs. 5g,i and 6d,e).

Precaudally, the mesenchyme of the neural archesand spines has differentiated into cartilage andstains strongly with Alcian blue (Fig. 5a–d). Bound-aries are now present separating the arches and car-tilage of the centra, and clear spinal nerve foraminahave developed (Fig. 5c,e–f). The neural arch tissuestill appears continuous throughout the anterior–posterior axis in microCT scans, but arches havebegun to form shallow grooves, corresponding withmyotomal segments (Fig. 5g–i). In sagittal section,boundaries separating the neural and intercalaryarches are visible (Fig. 5j). The transverse processesare well formed by late Stage 31, extending postero-laterally, and have begun to segment by subdividingmedially, although the lateral-most extensionsremain continuous (Fig. 5b,c,f,i).

In the thicker cartilaginous regions (centra), amiddle layer of eosin-stained tissue can be seen insection, which represents the first appearance of themineralized areolar tissue (Fig. 5c,c0). In the thin-ner, intervertebral regions, the layer of spindle-shaped cells has increased in thickness, formingdense, concentric rings around the notochord (Fig.5f,f0). In one CT-scan of a PMA-stained embryo,small bands of tissue seem to intrude into the noto-chord, forming on the inside of the notochordsheath, and at the widest part of the notochord ineach myotomal segment (Fig. 5b,d,h). These ringsare only present precaudally and are positioned justmedial to and in line with the grooves visible on thesurface of the outer cartilage, suggesting that theymay have a role in directing the subdivision of thecontinuous cartilage. However, these structureshave not yet been observed in any other specimens,indicating either that this specimen might be anom-alous, or that this tissue is ephemeral, being presentonly for a short time in development.

Embryonic morphology of the caudal vertebrae dif-fers slightly from the precaudal vertebrae by the endof Stage 31. The neural and intercalary arches havebegun to separate, with small slits forming in the car-tilage, although they remain continuous at the levelof the neural spines and at their bases (Fig. 6a–e).The centrum cartilages have also begun to thicken atintervals corresponding to each myotomal boundary,but the grooves representing incipient vertebralboundaries are not as deep as in the precaudal region(Fig. 6e). Both the fibrous intervertebral regions andthe widest parts of the centrum are similar to theprecaudal vertebrae, with a ring representing theearly areolar tissue medially and Alcian blue-stainedcartilage in the outer layer (Fig. 6f,g,g0,h,h0). Seg-mental constriction of the notochord has increasedfrom Stage 30, with clear decreases in notochorddiameter in each myotomal segment (Fig. 6i,j).Hemal spines remain as a continuous mesenchymalcondensation that appears discontinuous with thehemal arches (Fig. 6e–h). Dorsal root ganglia andspinal nerves remain broadly spaced, in contrast totheir close precaudal positioning (Fig. 6d,e).

Stages 32–33

All components of the vertebral column are pre-sent by Stage 32, and neural arches, intercalaryarches, and centra have all completely segmented(Figs. 7 and 8). The arch cartilages now consist ofclearly defined neural arches, which are long andthin in the precaudal region, and more substantialintercalary arches that contain spinal nerveforamina (Fig. 7a–d,g). The neural spines havealso begun to segment, but remain connected attheir bases (Fig. 7d,e). The transverse processeson the centrum cartilages are clearly pronounced,and extend both anteriorly at a slight ventralangle, and posterodorsally. The posterior portion of



Fig. 4. Leucoraja erinacea, embryonic vertebral morphology in Stage 30 embryos. a) anterior view of a CT reconstruction of theprecaudal vertebrae; b) transverse section through the IKI-stained CT scan; c) transverse section stained with HEA showing the lay-er of spindle-shaped cells and condensed mesenchyme of the transverse processes; c0) 203 zoom of the layer of spindle-shaped cells;d) anterolateral view of a CT reconstruction of the precaudal vertebrae; e) lateral view of a CT reconstruction of the precaudalvertebrae; f) horizontal section through a CT scan of the precaudal vertebrae; g) sagittal section stained with HEA; h) anterior viewof a CT reconstruction of the caudal vertebrae; i) transverse section through a CT scan of the caudal vertebrae; j) transverse sectionthrough the caudal vertebrae, stained with HEA; j0) 203 zoom of the layer of spindle-shaped cells shown in j; k) anterolateral viewof a CT reconstruction of caudal vertebrae; l) sagittal section through a CT scan of the caudal vertebrae; m) horizontal section of aCT scan through the caudal vertebrae; n) lateral view of a CT reconstruction of caudal vertebrae. a/v, caudal artery and vein, da, dor-sal aorta, drg, dorsal root ganglion, ha, hemal arch, hsp, hemal spine, me, mesenchymal sheath, mr, motor nerve root, na, neuralarch, nc, notochord, ncs, notochord sheath, nt, neural tube, tp, transverse process. Icons in upper right corner indicate plane of sec-tion. Scale bars are 100 lm for most images, and 50 lm for c0 and j0.

Fig. 5. Leucoraja erinacea, precaudal embryonic vertebral morphology in Stage 31 embryos. a) anterior view of a CT reconstructionof the precaudal vertebrae; b) transverse section through the PMA-stained CT scan showing the widest part of the vertebra, arrow-head refers to notochord band; c) transverse section stained with HEA, showing the thickened continuous cartilage; c0) 203 zoom ofthe centrum cartilage, dashed lines indicate first appearance of areolar tissue; d) sagittal section through a late Stage 31 skateembryo stained with HEA, dashed lines indicate areolar tissue; e) transverse section through the CT scan showing the morphology ofthe intervertebral region; f) transverse section stained with HEA, showing the narrow, fibrous nature of the intervertebral region; f0)203 zoom of the fibrous rings; g) lateral view of the CT reconstruction showing dorsal root ganglia and spinal nerves, arrows indicategrooves of incipient arch boundaries; h) lateral view of the CT reconstruction with centrum cartilage removed, arrowhead indicatesnotochord band; i) anterolateral view of the CT reconstruction showing the grooves of incipient vertebral boundaries, arrowheadsdepict location of grooves indicating incipient subdivision; j) horizontal section through the CT scan, arrowhead indicates notochordband. da, dorsal aorta, drg, dorsal root ganglion, c, cartilage, il, inner layer of cartilage, ml, middle layer of centrum/areolar calcifica-tion, mr, motor nerve root, na, neural arch, ncs, notochord sheath, nt, neural tube, ol, outer layer of hyaline cartilage, sr, sensorynerve root, tp, transverse process. Icons in upper right corner indicate plane of section. Scale bars are 200 lm for most images, scalebars are 100 lm for c0 and f0.



Fig. 6. Leucoraja erinacea, caudal embryonic vertebral morphology in Stage 31 embryos. a) anterior view of a CT reconstruction ofthe caudal vertebrae; b) transverse section through the PMA-stained CT scan showing the widest part of the vertebrae; c) transversesection through a CT scan showing the narrowest portion, where the continuous cartilage will eventually subdivide; d) lateral viewof the CT reconstruction; e) anterolateral view of the CT reconstruction; f) transverse section stained with HEA, showing the thick-ened continuous cartilage; g) transverse section showing the early formation of the middle layer of tissue that will eventually formthe areolar calcification; g0) 203 magnification showing the incipient centrum; h) transverse oblique section showing the fibrousnature of the intervertebral tissue; h0) 203 magnification of the fibrous intervertebral ring; i) sagittal section through the CT scanshowing segmental constriction of the notochord; and j) horizontal section through the CT scan showing segmental thickening andconstriction of the vertebral cartilage surrounding the notochord. a/v, caudal artery and vein, c, cartilage, drg, dorsal root ganglion,ha, hemal arch, hsp, hemal spine, il, inner layer of cartilage, ml, middle layer of centrum/areolar calcification, mr, motor nerve root,na, neural arch, nc, notochord, ncs, notochord sheath, nsp, neural spine, nt, neural tube, ol, outer layer of hyaline cartilage, sr, spinalnerve root. Scale bars represent 100 lm for most images; scale bars represent 50 lm for g0 and h0.



the transverse process overlaps the anteriorportion of the subsequent transverse process (Fig.7d–g).

The developing areolar tissue, stained with eosin insection, is clearly visible in CT scans, in HEA-stainedsections, and in whole mount skeletal preparations. In

Fig. 7. Leucoraja erinacea, precaudal embryonic vertebral morphology in Stage 32 embryos. a) anterior view of a CT reconstructionof the precaudal vertebrae; b) transverse section through the IKI-stained CT scan showing the widest part of the vertebrae; c) trans-verse oblique section stained with HEA; c0) 203 magnification of the fibrous ring shown in c; c00) 203 magnification of the mineraliz-ing centrum and outer layer of cartilage shown in c, dashed lines indicate middle layer/areolar calcification; d) lateral view of the CTscan of the precaudal vertebrae; e) transverse section through the intervertebral region of the CT scan; f) transverse section throughthe intervertebral region stained with HEA; f0) 203 magnification of the fibrous ring and transverse process shown in f; g) anterolat-eral view of the CT reconstruction; h) sagittal section of the CT scan; i) horizontal section of the CT scan; j) transverse sectionthrough a slightly older embryo showing the early areolar calcification of the centrum; j0) 203 magnification of the developing cen-trum, dashed lines indicate middle layer/areolar calcification; k) skeletal preparation of precaudal vertebrae in lateral view showingwell developed transverse processes. da, dorsal aorta, drg, dorsal root ganglion, il, inner layer of cartilage, iv, intervertebral region,ml, middle layer of centrum/areolar calcification, mr, motor nerve root, na, neural arch, nsp, neural spine, nt, neural tube, ol, outerlayer of hyaline cartilage, sr, spinal nerve root, tp, transverse process. Icons in upper right corner indicate plane of section. Scalebars represent 200 lm for most images; scale bars represent 100 lm for c0, c00, f0, and j0.



Fig. 8. Leucoraja erinacea, caudal embryonic vertebral morphology in Stage 32 embryos. a) anterior view of a CT reconstruction ofthe caudal vertebrae; b) transverse section through the IKI-stained CT scan, arrow indicates fibrous ring and arrowhead indicatesmineralizing centrum; c) transverse section stained with HEA; c0) 203 magnification of the fibrous ring shown in c; d) transversesection stained with HEA and showing the mineralizing centrum; d0) 203 magnification of the incipient areolar calcification, indicat-ed by the dashed lines; e) anterolateral view of the CT reconstruction; f) horizontal section through the CT scan; g) sagittal sectionstained with HEA and showing the early areolar calcifications and inner layer of cartilage; h) skeletal preparation of caudal verte-brae; i) lateral view of a CT scan of the caudal vertebrae; j) lateral view of a CT scan of the caudal vertebrae with outer cartilageremoved to show the inner layer (arrowhead) and intervertebral region (arrow); k) transverse oblique section of a slightly olderembryo showing both the mineralizing centrum and fibrous ring (arrow). a/v, caudal artery and vein, ca, cartilage, ha, hemal arch,hsp, hemal spine, ia, intercalary arch, il, inner layer of cartilage, iv, intervertebral region, ml, middle layer of the centrum/areolarcalcification, mrf, motor nerve root foramen, na, neural arch, nsp, neural spine, nt, neural tube, ol, outer layer of hyaline cartilage,srf, sensory nerve root foramen. Icons in upper right corner indicate plane of section. Scale bars represent 200 lm for most images;scale bars represent 100 lm for c0 and d0.



sagittal and transverse CT sections the areolar tissuebegins to substantially constrict the notochord (Fig.7h,i). The middle layer of areolar tissue forms withinthe layer of spindle-shaped cells and is separate fromthe outer hyaline cartilage that envelops the noto-chord (Fig. 7c,c00,j,j0,k). Interestingly, this areolar calci-fication first forms just external to the notochordsheath in Stage 32 (Figs. 5c,c0 and 7c,c00), but in slight-ly older, Stage 33, embryos, a layer of chondrocytessparsely embedded in Alcian blue-stained matrix (the“inner zone of vertebral development” of Ridewood,1921) is present medial to the areolar calcification,suggesting that cartilage is differentiating on theinside of this mineralized tissue as well as on the out-side (Fig. 7j,j0). These segmented centra, along withthe inner layer of cartilage, have increased in sizefrom Stage 31 and now substantially constrict thenotochord in the center of each vertebral unit (Fig.7h,i,j,j0). In the intervertebral regions, the fibrousrings have enlarged dramatically and surround thepersistent notochord (Fig. 7a,c,c0,d,e,f,f0,g). The outercartilage is no longer continuous, but is now presentin the form of rectangle-shaped boxes that surroundthe centrum and notochord. The dorsolateral cornersslant slightly posteriorly and contain spinal nerveforamina (Fig. 7d).

During Stage 32 several changes in vertebralmorphology through the precaudal to caudal tran-sition become more striking than in youngerembryos. The same population of mesenchymethat condenses to form the transverse processesprecaudally forms the hemal arches in the caudalregion. Throughout the precaudal-caudal transi-tion, in each successive vertebra, the transverseprocesses extend more and more ventrally untilthey are no longer transverse processes, but hemalarches (Fig. 8a; see Daniel (1922) for a clear figureof this transition in Rhinobatis). The caudal arteryand vein are surrounded by the hemal arches lat-erally and dorsally, and are protected ventrally byhemal spines (Fig. 8a–e). The first hemal spinepossesses two anterior prongs that cup the caudalartery and vein, as well as a median keel projec-ting ventrally (Fig. 8e). More posterior hemalspines are not as well developed and extend as asingle rod of ventral cartilage throughout the tail.At this point, the hemal arches and spines are notcontinuous with each other as they were in previ-ous stages. The fibrous rings, mineralized cen-trum, and matrix deposited medial to thecentrum, are visible in both CT and histologicalsection (Fig. 8b–d0,f,g).

Another visible difference between the caudaland precaudal regions includes a transition to thediplospondylic condition seen across elasmo-branchs, in which two vertebrae are present ineach myotomal segment. This change is visible inthe spacing out of spinal nerve foramina, with oneset of spinal nerves corresponding to each verte-bral unit precaudally, but corresponding with two

vertebral units in the caudal region (Fig. 8h–j).Neural and intercalary arches also change shapein relation to the re-spaced spinal nerves. Precau-dal neural arches are slender and articulate to thedorsal margin of the vertebral cartilage, while theintercalary arches are much wider and supportspinal nerve foramina ventrally. In the caudalregion, however, both neural and intercalaryarches adopt mostly the same shape, with the onlydifferences in morphology being a slight expansionof the apex in the intercalary arches and the basein the neural arches (Figs. 7k and 8h). One neuralspine is present for each set of neural and interca-lary arches. The CT-reconstructions in Figure8 (a,e,i,j) show that the anterior–posterior transi-tion to caudal diplospondyly does not occur at thefirst appearance of hemal arches, but rather sever-al vertebral segments posterior. The first severalcaudal vertebrae have hemal arches and spines,but both the relationship of spinal nerve foraminato vertebral cartilages and the neural arch mor-phology show the same pattern as the precaudalregion. The whole mount, cleared and stainedembryo (Fig. 8h) shows the more typical morpholo-gy of the arches in caudal vertebrae.

In transverse oblique HEA-stained sections ofslightly older (Stage 33) embryos, all layers of thedeveloping centrum are visible, including theinner layer, middle layer of areolar calcification,and outer layer of hyaline cartilage (Fig. 8k). Thelayer of fibrous, spindle-shaped cells that persistin the space between vertebral cartilages can alsobe seen in the intervertebral regions in obliquesections.

Stage 341

By Stage 34 the vertebrae have taken on anadult-like morphology, with more sharply definedvertebral boundaries, mineralized centra, and agreatly reduced notochord (Figs. 9 and 10). Theouter edges of the cartilage have begun to calcifyas dense blocks of tesserae, which are easily visi-ble in the slices of the microCT scans (Figs. 9b,e,hand 10b,d,e,g), and the mineralized centrum hasdramatically increased in both length and width.

The arches in the precaudal region are welldeveloped, and the large intercalary arches envel-op most of the spinal cord (Fig. 9a–g). The sensorynerve roots exit the vertebral column through theforamina in the intercalary arches, while motornerve roots exit through the dorsal portion of theouter hyaline cartilage (Fig. 9d,g). The neuralarches are present as small slivers of cartilagethat extend dorsally around the neural tube andexpand at their bases to provide foramina for theexiting motor nerve roots. Neural spines are fullyseparated now, with expanded bases that articu-late to the intercalary arches and wide dorsal mar-gins, forming an elongate hourglass in cross



section (Fig. 9e). Histological sections show afibrous perichondrium surrounding the hyalinecartilage of the arches and neural spines

(Fig. 9c,c0,f,f0). Both sensory and motor nerve rootsare large by this stage, to provide innervation tothe anteroposteriorly expanded pectoral fins.

Fig. 9. Leucoraja erinacea, precaudal embryonic vertebral morphology in Stage 34. a) anterior view of a CT reconstruction of thevertebrae; b) transverse section through the IKI-stained CT scan showing the mineralizing centrum; c) transverse section stainedwith HEA; c0) 103 magnification of the neural spine cartilage; c00) 103 magnification of the areolar calcification of the centrum; d)lateral view of a CT reconstruction; e) transverse section through the CT scan showing the narrowest part of the centrum; f) trans-verse section stained with HEA and showing the narrowest part of the centrum; f0) 103 magnification of the neural spine cartilage,arrow indicates developing tesserae; f00) 103 magnification of the areolar calcification of the centrum; g) anterolateral view of the CTreconstruction; h) sagittal section of the CT scan; i) skeletal preparation of precaudal vertebrae showing transverse processes, neuralarches, and intercalary arches. ca, outer cartilage, ce, centrum, ia, intercalary arch, il, inner layer of cartilage, iv, intervertebralregion, ml, middle layer of cartilage/areolar calcification, mr, motor nerve root, na, neural arch, nab, neural arch base, nc, notochord,nsp, neural spine, nt, neural tube, ol, outer layer of hyaline cartilage, sr, sensory nerve root, tp, transverse process. Icons in upperright corner indicate plane of section. Scale bars represent 200 lm for most images; scale bars represent 100 lm for c0, c00, f0, and c00.



Fig. 10. Leucoraja erinacea, caudal embryonic vertebral morphology in Stage 34 embryos. a) anterior view of a CT reconstruction ofthe caudal vertebrae; b) transverse section through the IKI-stained CT scan showing the mineralizing centrum; c) transverse sectionstained with HEA; c0) 103 magnification of the neural spine cartilage; c00) 103 magnification of the areolar calcification of the centrum;d) sagittal section of the CT scan; e) transverse section of the CT scan showing the narrowest part of the centrum; f) transverse sectionstained with HEA and showing the narrowest part of the centrum; f0) 103 magnification of the neural spine cartilage; f00 103 magnifica-tion of f, dashed lines indicate middle layer/areolar calcification of the centrum; g) horizontal section through the CT scan; h) lateral viewof a CT reconstruction; i) lateral view of the CT scan with outer cartilage removed; j) sagittal section stained with HEA, dashed lines indi-cate middle layer/areolar calcification. a/v, caudal artery and vein, ca, outer cartilage, ce, centrum, ha, hemal arch, il, inner layer of carti-lage, iv, intervertebral region, ml, middle layer of the centrum/areolar calcification, mr, motor nerve root, na, neural arch, nc, notochord,nsp, neural spine, nt, neural tube, ol, outer layer of hyaline cartilage, sr, sensory nerve root. Icons in upper right corner indicate plane ofsection. Scale bars represent 200 lm for most images; scale bars represent 100 lm for c0, c00, f0, and c00.

Centra consist of a layer of outer, hyaline cartilagesurrounding a middle layer of densely mineralizedrings of areolar calcification (Fig. 9b,c,c00,e,f,f00,h,i).The inner layer of the centrum, situated betweenthe notochord and the areolar calcification, hasalmost completely constricted the notochord in thecenter of each vertebra, leaving only a small noto-chordal canal in the center of these centra (Fig.9f,f00). In section, these three layers stain differential-ly, with the outer cartilage staining light pink, themineralized centrum staining a bright pink, and thethird, inner layer remaining white or staining lightblue (Fig. 9f00).

Similar to Stages 32 and 33 embryos, there areseveral differences in morphology between precau-dal and caudal vertebrae in Stage 34 embryos.The neural spines are anteroposteriorly longercaudally than in the precaudal region, and have amore rectangular and block-like appearance withlaterally expanding bases that articulate to boththe neural and intercalary arches (Fig. 10a–e,h–j).The arches and outer cartilage of the centrum con-sist of hyaline cartilage with a peripheral layer oftesserae (Fig. 10c,c0,f–g).

The hemal arches have extended ventrally to fullyenclose the caudal artery and vein (Fig. 9a), andindividual vertebrae are completely separated fromone another in the caudal region (Fig. 10h). Hemalarches and spines are fused together to form a singlestructure, but boundaries in the cartilages can stillbe seen in section (Fig. 10b,c,e,f). The hemal spinesare offset slightly posteriorly from the hemal arches,and the most posterior portion of each hemal spinerests ventral to the anterior portion of the next set ofhemal arches (Fig. 10h). The caudal mineralized cen-tra are very similar to those in the precaudal region,

showing a thin ring of areolar calcification in thecenter of each vertebral segment that thickens con-siderably near the anterior and posterior margins ofeach vertebra (Fig. 10c,c00,d,g,f,f00,i,j). The duplicationof caudal vertebrae results in the presence of twovertebrae for each myotomal segment and in theincreased spacing of spinal nerves (Fig. 10h). Theneural and intercalary arches are similar in thick-ness in the caudal region, with the base of the neuralarches articulating to the outer centrum cartilage toform spinal nerve foramina (Fig. 10h,i).

Summary of Vertebral Developmentin Leucoraja erinacea

The computed tomographic and histological analy-ses in this study provide a comprehensive view ofvertebral development in the little skate. We showthat a continuous layer of mesenchyme condensesaround the notochord at Stage 27 and persiststhrough vertebral development as a layer of spindle-shaped cells until it differentiates into the fibrousintervertebral rings and the areolar tissue in Stage31 (Fig. 11). Beginning at Stage 28, mesenchymalcells condense around the neural tube and caudalartery and vein to form continuous neural arch andhemal arch tissues that extend the length of the axialcolumn. During Stage 31, the neural and hemal archmesenchyme differentiates into hyaline cartilageand surrounds the areolar tissue, where it willremain into adulthood. In Stage 32 arch cartilagessubdivide into intercalary and neural arches andspines dorsally, and hemal arches and spines ventral-ly. An inner layer of cartilage forms medial to the are-olar calcification, and substantially constricts thenotochord. By Stage 34, the areolar calcification ofthe centrum is large and hourglass-shaped, and the

Fig. 11. Schematic showing the stages of vertebral development in the little skate, from Stages27–34. The two rows on the right side of the schematic show cross sections of the intervertebralregion (top) and the middle of a vertebra (bottom). Green 5 neural tube/spinal cord andred 5 caudal artery and vein.



inner layer of cartilage almost completely constrictsthe notochord in the middle of each vertebral seg-ment. In the intervertebral regions, the notochordpersists into adulthood and is surrounded by thethickened ring of fibrous cells.

DISCUSSIONThe Fibrous Sheath and Perichordal Tube

Several notable features were identified in thelittle skate in this study, including the initial for-mation of a continuous layer of spindle-shapedcells surrounding the notochord. This layer of cellssubsequently differentiates into the areolar calcifi-cations of the centra and the fibrous rings of theintervertebral regions. The spindle-shaped cellswere first described as the thickened “fibroussheath” of the notochord in several shark species,including Squalus acanthias, Mustelus vulgaris,and Scyliorhinus canicula, and this sheath waswidely accepted as a derivative of the notochordepithelium (Gegenbaur, 1862; Gadow and Abbott,1895). However, some researchers believed it to besclerotome-derived, along with the rest of the ver-tebral components (Hasse, 1879; Klaatsch, 1893).In sharks, the fibrous sheath is bounded external-ly by the elastica externa, a thin membrane ofintercrossing elastic fibers. The fibrous sheath isthought to be invaded in large numbers by sclero-tomal cells that perforate the elastica externa atthe bases of the neural and hemal arches and thendivide and differentiate into cartilage within thesheath (Gadow and Abbott, 1895; Goodrich, 1930).Cross sections of Stage 27 skate embryos showmesenchymal cells condensing around the noto-chord prior to taking on an elongate spindle shape(Fig. 1), suggesting that the cells of the fibroussheath are mesenchymal in origin. In sections ofStages 29 and 30 skate embryos, we were not ableto easily identify the elastica externa or any cellsthat were in the process of migrating into thefibrous layer of spindle-shaped cells. Additionally,this layer of spindle-shaped cells is morphological-ly different than those described previously insharks, with the cells having more irregularshapes and the layer including more extracellularmatrix (Figs. 2c0,i0 and 3c0,j0; compare to Eameset al. (2007) Fig. 6). These differences imply thatthe fibrous sheath, previously identified to be aderivative of the notochord epithelium in sharks,in batoids develops from mesenchymal cells thathave condensed around the notochord. If this isthe case, then the areolar calcifications in skatesdo not form through sclerotomal cells invading anotochord-derived fibrous sheath (the processdescribed in sharks by Gadow and Abbott, 1895),but rather through mesenchymal cells first differ-entiating into concentric layers of elongate cellsand then calcifying.

Similar continuous condensations of tissue sur-rounding the notochord have been described inseveral other vertebrate taxa, such as salamander,chick, and mouse, in which they have been termedthe perichordal tube (Williams, 1942; Williams,1959; Wake and Lawson, 1973; Christ and Wilting,1992). However, despite superficial similarities,there are important differences between the peri-chordal tube of tetrapods and the fibrous sheath ofelasmobranchs. For example, in skate, the cells ofthe fibrous sheath become elongate and thinthroughout the length of the axial column, run-ning parallel to the notochord epithelium. Thecells of the perichordal tube in chick, however,have the characteristic appearance of mesenchy-mal cells, and flatten into fibrous rings only at theintervertebral boundaries, early in vertebral devel-opment (Williams, 1942; Christ et al., 2000). Thissegmental thickening in chick occurs before theneural arches are fully developed. Similarly, in aplethodontid salamander, Eurycea bislineata, theperichordal tube is not consistent in diameterthroughout the axial column, but rather formsthickened rings of elongate cells at incipient verte-bral boundaries and thins to just one or two cellsin the middle of each vertebral segment (Wakeand Lawson, 1973). This unevenness contrastswith the fibrous sheath and the continuous carti-lage of skates, which remains a constant thicknessfor several weeks of development, and does notsubdivide until the neural and hemal arch mesen-chyme has condensed and differentiated intocartilage.

It is unclear how widespread the perichordaltube is among vertebrates. Recent studies of theJapanese medaka (Oryzias latipes), a teleost fish,do not mention any layer of continuous tissue, butinstead report that mesenchyme invades the noto-chord sheath at segmental intervals (e.g., Inohayaet al., 2007) and studies on vertebral developmentin other teleosts, such as zebrafish, do not discussthe condensation of sclerotomal mesenchyme ingreat detail (Morin-Kensicki et al., 2002; Fleminget al., 2004). However, there is some evidence of athin tube of continuous mesenchyme in some tele-ost fishes. A thin layer of undifferentiated mesen-chyme surrounding the notochord prior to centrumformation has been documented in Atlantic salm-on, but, by the time centra begin to mineralize,neural and hemal arch cartilages are already wellformed (Grotmol et al., 2003).

In addition to the notochord, mesenchyme alsosurrounds the neural tube in the form of unseg-mented, nascent neural and intercalary arches inskate that differentiate into cartilage and laterseparate. We have found no evidence in the litera-ture describing this continuous arch tissue in ver-tebrates outside sharks. Van Wijhe (1922) and DeBeer (1924) documented the continuous early con-densation of the neural arch mesenchyme in



Squalus acanthias and Heterodontus francisci,respectively, but these observations were later dis-missed by Goodrich (1930) and “better ascribed toa temporary fusion of rudiments set very closetogether than to the arcualia having been evolvedfrom an originally continuous cartilaginous band”(p. 16). Our observations, in the present work tak-en from histological sections and microCT scans ofthe little skate and in previous works observed inseveral species of sharks (van Wijhe, 1922; deBeer, 1924), are inconsistent with Gadow andAbbott’s (1895) model of vertebral development inwhich pairs of basidorsal and interdorsal carti-lages form the building blocks of the neural andintercalary arches. Using microCT scans and his-tology, we have shown here that in skate the ear-liest condensations of mesenchyme around theneural tube and caudal artery and vein areindeed continuous throughout the vertebral col-umn length.

Subdivision of Initially ContinuousVertebral Cartilage

The mode of vertebral segmentation described inthis study, in which segmented paraxial mesodermmigrates to condense around the notochord andneural tube, forms a continuous sleeve of cartilagethat extends the entire length of the body, and thensecondarily subdivides into discrete, repeating ver-tebral elements, has received minimal attention inthe literature. Studies describing vertebral develop-ment in chick, mouse, and human mention a peri-chordal tube of condensed mesenchyme, but nomechanism to explain the secondary subdivision ofthis structure is proposed (Christ and Wilting, 1992;Wallin et al., 1994; Christ et al., 2000). It is possiblethat other segmented structures, such as spinalnerves and intersegmental blood vessels, driveperichordal tube subdivision, but little functionalevidence exists to support this.

Segmentation of the perichordal tube in chickseems to take place when the incipient interverte-bral discs begin to develop within the perichordalmesenchyme and replace the notochord (Christ andWilting, 1992). These discs provide boundaries fordeveloping vertebral centra, which later fuse to neu-ral arch rudiments. In mammals, a similar processoccurs, though the notochord is not completelyreplaced, but instead becomes the nucleus pulposusat the center of each intervertebral disc (Lefebvreand Bhattaram, 2010). In skate, cells within thefibrous sheath first differentiate into hyaline carti-lage, which thickens substantially, before secondari-ly subdividing into vertebrae late in Stage 31.

The small bands of tissue that surround the noto-chord and correspond to nascent vertebral bound-aries, as observed in one late Stage 31 skate embryo(Fig. 4b,j), might relate to the segmentation of thecontinuous vertebral cartilage in skate. The location

of the bands at the widest part of the notochord, andjust medial to the grooves in the dividing cartilage,is intriguing, and we speculate that these bandscould represent iterative signaling centers thatdirect the subdivision of the tube of hyaline carti-lage. Notochord segmentation has been proposed asa driver of centrum development in chick and insome teleost fishes (Stern, 1990; Grotmol et al.,2003). For example, in Atlantic salmon, calcificationof the centrum seems to be initiated by the segmen-tal organization of a layer of “chordoblast” cellslocated in the notochord sheath (Grotmol et al.,2005; Nordvik et al., 2005). In each vertebral seg-ment these cells transition from an anteroposteriororientation to circular bands perpendicular to theAP axis. Subsequently, chordacentra begin to calcifyin the notochord sheath at the position of these chor-doblast bands, and sclerotomally derived osteoblastsbegin to deposit bone on the surface of the chorda-centra (Grotmol et al., 2003). However, beyond thesemorphological observations, the phenomenon ofnotochord segmentation is controversial: it has notbeen documented outside of a small number ofteleost species (Grotmol et al., 2003, 2005;Haga et al., 2009), and an underlying molecularmechanism remains elusive.

The Composite Nature of the AxialSkeleton in Gnathostomes

Previous studies of skeletogenesis in sharks haveidentified three tissue layers early in the formationof the centrum, which is similar to what is docu-mented in skate in this study (Daniel, 1922; Eameset al., 2007; Enault et al., 2015). These three layers,or zones of deposition, include an inner layer consist-ing of mostly extracellular matrix located superficialto the notochord, a middle layer of areolar calcifica-tion that forms the double cone of the centrum, andan outer layer of hyaline cartilage that is continuouswith the arches and is jacketed with tesserae inadults. These layers stain differentially in later stageswell shark embryos (9 cm total length) (Eameset al., 2007) with the inner and outer layers stainingwith Safranin O and Alcian blue, indicating the pres-ence of proteoglycans in the cartilage matrix, and themiddle layer staining with Aniline blue, denoting thepresence of collagen fibrils. Both the inner and mid-dle layer, as well as the inner-most margin of the out-er layer, showed alkaline phosphatase activity, whichis indicative of mineralization. We recover the samethree layers in the vertebrae of the little skate (Fig.8f,f00), suggesting that centrum construction is simi-lar across many elasmobranchs. As noted previously,however, in some sharks and batoids (the porbeagleshark Lamna nasus, the hammerhead Sphyrna blo-chii, the basking shark Cetorhinus maximus, and theguitarfish Rhinobatos productus) there is no outerlayer of hyaline cartilage around the centrum;



hyaline cartilage is restricted to the arches (Ride-wood, 1921; Daniel, 1922; Goodrich, 1930).

Multiple-part centra are known in many kindsof fishes. In most of these, the dual centrum com-plex consists of a chordacentral layer formed with-in the notochord sheath and an autocentrumconsisting of a layer of bone deposited on the out-side of the sheath and independent of the neuraland hemal arches (Arratia et al., 2001). For exam-ple, the centrum of the fossil holostean Ophiopsisis formed from an autocentrous deposition ofsmooth and compact bone on the outside of thenotochord sheath, as well as an inner chordacen-trum containing one or two rings of small tubes(Bartram, 1975). The Jurassic teleost Leptolepisalso has an inner chordacentrum with a smoothouter autocentrum (Arratia and Hikuroa, 2010).Similarly, centra in Atlantic salmon have beendocumented to form from an inner layer of cellslocated within the notochord sheath and an outerlayer of bone that is deposited by mesenchymalcells (Nordvik et al., 2005). In the little skate,however, the outer centrum cartilage does notdevelop independently of the neural and hemalarches, but is continuous with them and forms viathe same mesenchymal cell population. These vari-ous examples provide evidence that centrum devel-opment is often complex and difficult to document.Centra are not merely formed through a singleprocess, but can develop through a combination ofmineralization within the notochord sheath, aswell as condensation of sclerotomal cells aroundthe notochord and neural tube.

This variation in centrum construction raisesquestions about the embryonic origins of these dif-ferent centrum components. As mentioned previ-ously, fate mapping experiments in axolotl andchick have shown that tetrapod centra, along withthe remainder of the vertebral skeleton, arederived from cells of the ventral somites (Bagnallet al., 1988; Piekarski and Olsson, 2014). However,the same does not seem to be entirely true for tele-ost fishes. The notochord appears to play a signifi-cant role in centrum development in Japanesemedaka and Atlantic salmon, with a separate lay-er of centrum bone developing within the noto-chord sheath, independent of the sclerotome-derived skeletal tissue (Ekanayake and Hall, 1987;Nordvik et al., 2005). Even more compelling is theevidence from zebrafish, in which notochords havebeen shown to secrete bone matrix, and laser abla-tions of notochord cells result in the loss of centra(Fleming et al., 2004).

CONCLUSIONS

This study has comprehensively described theembryonic morphology and histology of the verte-bral column in a skate for the first time. Histologi-cal and, in particular, 3D-microtomographic data

have helped to elucidate the process of early mes-enchymal condensation and differentiation. Theanteroposteriorly continuous layer of cells thatsurrounds the notochord early in vertebral devel-opment appears to be mesenchymal in origin (like-ly derived from the sclerotome), and not derivedfrom the notochord epithelium, as has been previ-ously described in shark. It is possible that thenotochord epithelium does play some role in induc-ing the formation of the layer of spindle-shapedcells, but additional experiments are necessary totest this. Additionally, rather than being separatestructures at their genesis, neural arch mesen-chyme initially condenses continuously throughoutthe anterior–posterior axis before separating intoneural and intercalary arches. Centra haveevolved independently in several clades of verte-brates (Arratia et al., 2001), and those separateorigins are evident when comparing centrumdevelopment across taxa. Skate centrum construc-tion is markedly differently from that ofosteichthyans (e.g., zebrafish and chick), in thatthe centrum consists of three layers of tissue: theinner layer of cartilage that constricts the noto-chord, the areolar calcification that makes up themiddle layer of the centrum, and an outer layer ofhyaline cartilage. Zebrafish and chick centra bothconsist of one type of bone: acellular bone in theformer and cellular bone in the latter (Hall, 2005).

We have provided critical information on themorphology and tissues present throughout skatevertebral development, but many broader ques-tions remain unresolved, including the phylogenet-ic distributions of structures like the perichordaltube and composite centra, the embryonic originsof the centrum, and the mechanism driving thesubdivision of the cartilaginous tube in elasmo-branchs. These can be remedied with contrast-enhanced microCT, scans documenting axial col-umn development across a broader range of taxa.Fate-mapping experiments to track the derivativesof both somites and the notochord in elasmo-branchs will also be crucial in determining if thenotochord makes a substantial and general (per-haps ancestral) contribution to centrum develop-ment in gnathostomes, or whether the notochord-derived centra of teleosts are unique. Additionalmorphological studies aimed at identifying seg-mental features of the notochord (such as thenotochord bands or variations in notochord thick-ness) in other embryos, and experiments to iden-tify and investigate the function of genesexpressed segmentally along the notochord, willfurther test the role of the notochord in axial col-umn segmentation.

When comparing vertebral development acrossgnathostomes, it becomes clear that this process iscomplex and variable within major taxa, and thatGadow and Abbott’s (1895) “Arcualia Theory” ofvertebral development does not apply to all



elasmobranchs, let alone all vertebrates. Featuresthat were clearly defined in several different spe-cies of shark, including the elastica externa, seemto be absent in the little skate. These differencesagain underscore the fact that shark anatomy anddevelopment are not general for vertebrates. Rath-er, as the sister group to osteichthyans, develop-mental data from elasmobranchs can help topolarize character transformations within sub-groups of the bony fishes. As vertebrae representsuch a fundamental feature of the vertebrate bodyplan, their complex construction and independentevolutionary histories underpinning vertebral skel-etal diversity provide an opportunity to study vari-ation in the mechanisms of axial segmentationand the developmental basis of convergentevolution.

AUTHOR CONTRIBUTIONS

This project was conceived of and designed byKEC with advice from MIC and JAG. CT-scananalysis, histology, and skeletal preparations werecompleted by KEC. The manuscript was writtenby KEC with input from MIC and JAG.

ACKNOWLEDGMENTS

We thank Robert Ho, Melina Hale, and HaleyStinnett for valuable discussions and feedback onthe manuscript. Scott Bennett and David Remsenat the Marine Biological Laboratory in WoodsHole, MA, provided Leucoraja erinacea embryosfor this study. Jessie Maisano at the University ofTexas High-Resolution X-Ray Computed Tomogra-phy Facility scanned some Stages 28–32 embryos.April Neander provided technical assistance withthe GE CT scanner. We also thank Richard Beh-ringer and Alejandro S�anchez Alvarado for sup-port with lab space at the MBL. This project wassupported by NSF DDIG DEB-1501749 and theMBL/UChicago Graduate Student ResearchAward.

LITERATURE CITED

Applegate SP. 1967. A survey of shark hard parts. In: Gilbert PW,Mathewson RF, Rall DP, editors. Sharks, Skates and Rays. Bal-timore, Maryland: The Johns Hopkins Press. pp 37–67.

Arratia G, Hikuroa D. 2010. Jurassic fishes from the LatadyGroup, Antarctic Peninsula, and the oldest teleosts from Ant-arctica. J Vert Paleontol 30:1331–1342.

Arratia G, Schultze H-P, Casciotta J. 2001. Vertebral columnand associated elements in dipnoans and comparison with oth-er fishes: Development and homology. J Morphol 250:101–172.

Bagnall KM, Higgins SJ, Sanders EJ. 1988. The contributionmade by a single somite to the vertebral column: Experimen-tal evidence in support of resegmentation using the chick-quail chimaera model. Development 103:69–85.

Bartram AWH. 1975. The holostean fish genus Ophiopsis Agas-siz. Zool J Linn Soc 56:183–205.

Bensimon-Brito A, Cardeira J, Cancela ML, Huysseune A,Witten PE. 2012. Distinct patterns of notochord

mineralization in zebrafish coincide with the localization ofOsteocalcin isoform 1 during early vertebral centra formation.BMC Dev Biol 12:28.

Christ B, Wilting J. 1992. From somites to vertebral column.Ann Anat 174:23–32.

Christ B, Huang R, Wilting J. 2000. The development of theavian vertebral column. Anat Embryol 202:179–194.

Daniel JF. 1922. The Elasmobranch Fishes. Berkeley, Califor-nia: University of California Press. 334 p.

de Beer GR. 1924. Memoirs: Contributions to the study of thedevelopment of the head in Heterodontus. J Cell Sci s2–68:39–65.

Dean MN, Summers AP. 2006. Mineralized cartilage in theskeleton of chondrichthyan fishes. Zoology 109:164–168.

Eames BF, Allen N, Young J, Kaplan A, Helms JA, SchneiderRA. 2007. Skeletogenesis in the swell shark Cephaloscylliumventriosum. J Anat 210:542–554.

Ekanayake S, Hall BK. 1987. The development of acellularityof the vertebral bone of the Japanese medaka, Oryzias latipes(Teleostei; Cyprinidontidae). J Morphol 193:253–261.

Enault S, Mu~noz DN, Silva WTAF, Borday-Birraux V, BonadeM, Oulion S, Vent�eo S, Marcellini S, Debiais-Thibaud M.2015. Molecular footprinting of skeletal tissues in the cat-shark Scyliorhinus canicula and the clawed frog Xenopus tro-picalis identifies conserved and derived features of vertebratecalcification. Front Genet 6. doi: 10.3389/fgene.2015.00283.

Fischer AH, Jacobson KA, Rose J, Zeller R. 2008. Hematoxylinand eosin staining of tissue and cell sections. CSH Protoc2008:pdb.prot4986.

Fleming A, Keynes R, Tannahill D. 2004. A central role for thenotochord in vertebral patterning. Development 131:873–880.

Fleming A, Kishida MG, Kimmel CB, Keynes RJ. 2015. Build-ing the backbone: The development and evolution of vertebralpatterning. Development 142:1733–1744.

Gadow H. 1933. The Evolution of the Vertebral Column. Cam-bridge: Cambridge University Press. 356 p.

Gadow H, Abbott EC. 1895. On the evolution of the vertebralcolumn of fishes. Philos Trans R Soc Lond B 186:163–221.

Gegenbaur C. 1862. Untersuchungen zur vergleichenden Anat-omie der Wirbelsaule bei Amphibien and Reptilien. Leipzig:Wilhelmn Engelmann. 71 p.

Gillis JA, Dahn RD, Shubin NH. 2009. Chondrogenesis andhomology of the visceral skeleton in the little skate, Leucor-aja erinacea (Chondrichthyes: Batoidea). J Morphol 270:628–643.

Goodrich E. 1930. Studies on the Structure and Development ofVertebrates. London: Dover Publications. 837 p.

Grotmol S, Kryvi H, Nordvik K, Totland GK. 2003. Notochordsegmentation may lay down the pathway for the developmentof the vertebral bodies in the Atlantic salmon. Anat Embryol207:263–272.

Grotmol S, Nordvik K, Kryvi H, Totland GK. 2005. A segmentalpattern of alkaline phosphatase activity within the notochordcoincides with the initial formation of the vertebral bodies.J Anat 206:427–436.

Haga Y, Dominique VJ, Du SJ. 2009. Analyzing notochord seg-mentation and intervertebral disc formation using the twhh:gfp transgenic zebrafish model. Transgenic Res 18:669–683.

Hall BK. 2005. Bones and Cartilage: Developmental and Evolu-tionary Skeletal Biology. San Diego, California: Elsevier Aca-demic Press. 760 p.

Hasse K. 1879. Das nat€urliche System der Elasmobranchier aufGrundlage des Baues und der Entwicklung ihrer Wirbels€aule:eine morphologische und pal€aontologische Studie. Jena: Gus-tav Fischer. 284 p.

Inohaya K, Takano Y, Kudo A. 2007. The teleost intervertebralregion acts as a growth center of the centrum: In vivo visuali-zation of osteoblasts and their progenitors in transgenic fish.Dev Dyn 236:3031–3046.

Jarvik E. 1980a. Basic Structure and Evolution of Vertebrates.London: Academic Press. 575 p.

Jarvik E. 1980b. Basic Structure and Evolution of Vertebrates.London: Academic Press. 337 p.



info:doi/10.3389/fgene.2015.00283

Klaatsch H. 1893. Beitrage zur vergleichenden Anatomie derWirbelsaule. I. Ueber den Urzustand der Fischwirbelsaule.Morphologisches Jahrbuch 19:649–680.

Klymkowsky MW, Hanken J. 1991. Whole-mount staining ofXenopus and other vertebrates. Methods Cell Biol 36:419–441.

Lefebvre V, Bhattaram P. 2010. Chapter Eight—Vertebrate Skele-togenesis. In: Koopman P, editor. Current Topics in Developmen-tal Biology, Vol. 90. San Diego, CA: Academic Press. pp 291–317.

Maxwell EE, Fr€obisch NB, Heppleston AC. 2008. Variability andconservation in late chondrichthyan development: Ontogeny ofthe winter skate (Leucoraja ocellata). Anat Rec 291:1079–1087.

Miles RS. 1970. Remarks on the vertebral column and caudalfin of acanthodian fishes. Lethaia 3:343–362.

Morin-Kensicki EM, Melancon E, Eisen JS. 2002. Segmentalrelationship between somites and vertebral column in zebra-fish. Development 129:3851–3860.

Nordvik K, Kryvi H, Totland GK, Grotmol S. 2005. The salmonvertebral body develops through mineralization of two pre-formed tissues that are encompassed by two layers of bone.J Anat 206:103–114.

Owen R. 1848. On the Archetype and Homologies of the Verte-brate Skeleton. London: Richard and John E. Taylor. 203 p.

Piekarski N, Olsson L. 2014. Resegmentation in the mexicanaxolotl, Ambystoma mexicanum. J Morphol 275:141–152.

Ridewood W. 1921. On the calcification of the vertebral centrain sharks and rays. Philos Trans R Soc Lond B 210:311–407.

Ridewood WG. 1899. Some observations on the caudal diplo-spondyly of sharks. Zool J Linn Soc Lond 27:46–59.

Schaeffer B. 1967. Osteichthyan vertebrae. Zool J Linn SocLond 47:185–195.

Stern CD. 1990. Two distinct mechanisms for segmentation?Semin Dev Biol 1:109–116.

van Wijhe JW. 1922. Fr€uhe Entwicklungsstadien des Kopf- undRumpfskeletts von Acanthias vulgaris. Bijdragen tot de Dier-kunde 1922:271–298.

Wake DB, Lawson R. 1973. Developmental and adult morpholo-gy of the vertebral column in the plethodontid salamanderEurycea bislineata, with comments on vertebral evolution inthe Amphibia. J Morphol 139:251–299.

Wallin J, Wilting J, Koseki H, Fritsch R, Christ B, Balling R.1994. The role of Pax-1 in axial skeleton development. Devel-opment 120:1109–1121.

Wang S, Kryvi H, Grotmol S, Wargelius A, Krossøy C, Epple M,Neues F, Furmanek T, Totland GK. 2013. Mineralization ofthe vertebral bodies in Atlantic salmon (Salmo salar L.) isinitiated segmentally in the form of hydroxyapatite crystalaccretions in the notochord sheath. J Anat 223:159–170.

Williams EE. 1959. Gadow’s arcualia and the development oftetrapod vertebrae. Q Rev Biol 34:1–32.

Williams JL. 1942. The development of cervical vertebrae inthe chick under normal and experimental conditions. Am JAnat 71:153–179.



Embryonic Development of the Axial Column in the Little Skate, … · 2019-02-08 · Embryonic Development of the Axial Column in the Little Skate, Leucoraja erinacea Katharine E.

Documents