Norntates PUBLISHED by the AMERICAN MUSEUM of NATURAL HISTORY CENTRAL PARK WEST at 79TH STREET, NEW YORK, N.Y

Home , Acrodus, Asteracanthus, Cobelodus, Hemipristis, Hybodontidae, Hybodontiformes

AMERICAN MUSEUM Norntates PUBLISHED BY THE AMERICAN MUSEUM OF NATURAL HISTORY CENTRAL PARK WEST AT 79TH STREET, NEW YORK, N.Y. 10024 Number 2724, pp. 1-48, figs. 1-17 April 14, 1982

The Anatomy and Interrelationships of Mesozoic Hybodont Sharks

JOHN G. MAISEY1

ABSTRACT

The anatomy of the head and jaws, the post- on the basis of tooth morphology. Geologically cranial and the dermal skeleton of Mesozoic hy- later hybodonts lack enlarged nutritive foramina bodont sharks (Hybodus, Acrodus, Lissodus, As- in their teeth, but some early members of the teracanthus, and allied genera) is reviewed. The group retain these foramina which probably rep- neurocranium and jaws are best known in Hy- resent a chondrichthyan synapomorphy. It seems bodus basanus, although little is known of its possible to use the presence of cephalic spines postcranial skeleton. Comparison of geologically (" Sphenonchus ") as a characteristic of the group. older hybodonts such as H. hauffianus and H. The morphology of these spines is discussed and fraasi reveals many similarities with H. basanus a simple terminology is proposed. Hybodont ce- in their cranial anatomy, which has been badly phalic spines are recognized from the late Paleo- misinterpreted in previous works. An attempt is zoic (Carboniferous and Permian) of North Amer- made to rectify the confusion surrounding jaw ica, where they are associated with characteristic suspension in hybodont sharks. Asteracanthus is teeth and finspines, and it is concluded that Me- shown to share several peculiarities with Hybo- sozoic hybodonts represent terminal members of dus and Acrodus in its palatoquadrate morphol- a selachian group which has a lengthy Paleozoic ogy, in spite of certain specializations toward a history. This is a far more precise statement than presumably durophagous habitus. The signifi- has hitherto been made regarding the early history cance of a pleural rib cage in hybodonts is dis- of hybodont sharks. The question of a relationship cussed; it supports the suggestion (based on de- between hybodonts and Heterodontus (Port Jack- velopmental studies) that the intermuscular ribs son sharks) is discussed. It is shown that Heter- of Recent sharks are homologous with pleural odontus shares numerous apparently synapomor- ("ventral") ribs of osteichthyans. Synapomor- phic characters with other Recent sharks and phies of hybodonts and Recent elasmobranchs in- rays, and possesses none of the hybodont syn- clude the presence of a cranial ectethmoid pro- apomorphies recognized in this work. Similarities cess, a continuous puboischiadic bar in the pelvic between Heterodontus and Hybodus, which have fins, a gap between the basihyal and basibranchi- been used in the past to suggest a relationship, als and posteriorly directed hypobranchials. are rejected on the grounds that they are either It is not possible to recognize hybodonts solely plesiomorphies, or convergent, or spurious.

1 Assistant Curator, Department of Vertebrate Paleontology, American Museum of Natural History.

INTRODUCTION The dominant sharks during most of the these forms is planned. The present work is Mesozoic era were Hybodus, Acrodus, As- therefore a preliminary survey of Mesozoic teracanthus, and allied genera such as Lon- hybodont sharks, and is to be followed by chidion and Lissodus. These sharks have detailed morphological studies of various been collectively referred to as hybodonts, members of the group, beginning with Hyb- although there is no agreed definition of what odus basanus. the term hybodont means. Tooth morphology ought to be an important criterion, since ACKNOWLEDGMENTS hybodonts were first recognized on the basis I have received a great deal of help from of isolated teeth (Agassiz, 1837), but it is by the following people, who have lent speci- no means clear from the literature that sig- mens, sent casts and photographs, or dis- nificant differences between hybodont and cussed and constructively criticized this other shark teeth were (or can be) recog- work as it progressed: Drs. J. Chorn, L. nized. Agassiz (1837) included Cladodus in Compagno, P. Forey, N. Hotton, G. John- his "Hybodontes," and Paleozoic sharks son, C. Patterson, W. Reif, D. Rosen, B. with finspines and Cladodus-like teeth (e.g., Schaeffer, K. Thomson, R. Wild, M. Wil- "Ctenacanthus" costellatus) have been in- liams, R. Zangerl, J. Zidek, and Mr. G. cluded in a "hybodont" taxon by numerous Dingerkus. The illustrations for this work authors. These Paleozoic sharks resemble were produced by Ms. L. Meeker and Mr. Mesozoic hybodonts in general respects and C. Tarka. may be related at a higher taxonomic level ABBREVIATIONS (e.g., phalacanthous sharks sensu Zangerl, ANATOMICAL: 1979), but they do not share characters iden- acv, foramen for anterior cerebral vein tified here as synapomorphies of hybodonts, adbc, anterior dorsal basal cartilage and should not therefore be placed in that adsp, anterior dorsal spine taxon. It is now possible to refer some frag- af, adductor fossa mentary Carboniferous and Permian material artp, articular process of scapulocoracoid such as teeth, scales, cephalic spines, and at, accessory terminal cartilage finspines specifically to hybodont sharks like b, basal segment of pelvic metapterygium those from the Mesozoic, and to make a ba, barb on cephalic spine much more positive statement than was hith- bbr, basibranchial erto possible regarding the Paleozoic history bd, basidorsal of these sharks. bp, basal plate Recent discovery of com- by, basiventral plete hybodont-like sharks in the Pennsyl- ,B, beta cartilage vanian of Kansas (Zidek, personal commun.) c, crown of cephalic spine and Mississippian of Scotland (Dick, 1978; cbr, ceratobranchial Dick and Maisey, 1980) make a review of the cbrf, coracobranchial fossa better known Mesozoic hybodonts essential. ceph, cephalic spine Many inaccuracies and discrepancies have ch, ceratohyal come to light in preparing this review, par- cik, caudal internasal keel ticularly in important works on Mesozoic cl, lateral cusp (cephalic spine) hybodonts such as Brown (1900), Koken cor, coracoid region (1907), and Peyer (1946). Much of the pres- crd, dorsal crest (cephalic spine) crl, lateral crest (cephalic spine) ent comparison is based on as yet unpub- crm, mesial crest (cephalic spine) lished data for Hybodus basanus from the crp, posterior crest (cephalic spine) early Cretaceous of southern England (Mai- df, diazonal nerve foramen sey, in prep.). For the present my comments dm, dorsal marginal cartilage on forms other than H. basanus are intended dnc, dorsal nerve cord to be provisional, but a reinvestigation of dt, dorsal terminal cartilage 1982 MAISEY: HYBODONT SHARKS 3 eart, ethmoid articulation s, striae (cephalic spine) ebr, epibranchial sc, sapulocoracoid ect, ectethmoid process scap, scapular region eha, efferent hyoidean artery sof, spino-occipital foramen em, epaxial muscle som, somatic peritoneum endf, endolymphatic fossa sscap, suprascapula epif, epiphyseal foramen subs, suborbital shelf ethp, ethmopalatine process supc, supraorbital crest fepsa, foramen for efferent pseudobranchial ar- tfr, trigemino-facialis recess tery vt, ventral terminal cartilage fm, foramen magnum I-X, cranial nerve foramina gf, glenoid fossa INSTITUTIONAL: hm, hypaxial muscle hs, horizontal septum AMNH, American Museum of Natural History hym, hyomandibula BM(NH), British Museum (Natural History) hym art, hyomandibular articulation KU, Kansas University hyp, hypophyseal foramen USNM, United States National Museum, Smith- i, intermediate segment sonian Institution ica, internal carotid artery id, interdorsal HISTORICAL REVIEW OF ints, internasal septum HYBODONT SHARKS jc, jugular canal Agassiz (1837) placed Hybodus in a higher jJ, jaw joint taxon of "Hybodontes" along with Clado- lab, labial cartilage dus, Diplodus, and Sphenonchus. These latc, lateral commissure "Hybodontes" were divided into two groups lda, groove for lateral dorsal aorta on the basis of what were seen as differences 11, lateral lobe (cephalic spine) lm, mesial lobe (cephalic spine) in tooth morphology. One group, comprising lop, lateral otic process Hybodus and Cladodus, has teeth with acu- lp, posterior lobe (cephalic spine) minate cusps which lack a principal pulp cav- Mc, Meckel's cartilage (lower jaw) ity (i.e., osteodont teeth sensu Glikman, mes, mesopterygium 1964). The other group, comprising Sphe- met, metapterygium nonchus and Diplodus, has teeth with rather mil, lateral marginal indentation (cephalic spine) different cusp arrangements, and a principal mim, mesial marginal indentation (cephalic spine) pulp cavity is supposedly present (orthodont mpt, mixipterygium (clasper cartilage) teeth of Glikman, 1964). It was subsequently nc, nasal capsule established that Sphenonchus is founded not not, notochord o, orbit on a tooth but rather a cephalic spine of a oc, otic capsule shark with Hybodus teeth (Charlesworth, occ, occipital cotylus 1845; Day, 1864; Fraas, 1889). Moreover, olf, olfactory tract sharks (xenacanths) with Diplodus teeth are ora, foramen for orbital artery now also well known and differ in many re- palp, palatine process spects from Hybodus. Sharks with Cladodus pbr, pharyngobranchial teeth are known to have differed widely from pdbs, posterior dorsal basal cartilage one another and from Hybodus in many mor- pdsp, posterior dorsal spine phological features. Thus of Agassiz's (1837) pelvg, pelvic girdle taxa two and are pf, precerebral fontanelle (Hybodus Sphenonchus) pop, postorbital process synonymous and Sphenonchus is not a valid pq, palatoquadrate genus. The genus Acrodus was placed in the pro, propterygium "Cestraciontes" along with the living Het- psc, posterior vertical semicircular canal erodontus (Cestracion) and Ceratodus (sub- qf, quadrate flange sequently recognized as a dipnoan), Cte- r, radial noptychius, Orodus, Helodus, Chomatodus, rb, rostral bar Psammodus, Cochliodus, Poecilodus, Pleu- 4 AMERICAN MUSEUM NOVITATES NO. 2724 rodus, Ctenodus, Strophodus, and Ptycho- Jaekel (1889, 1898) placed great emphasis dus. All these forms possessed pavement on hybodont tooth histology and microstruc- teeth adapted for a durophagus habitus, and ture. In some respects his works seem more in hindsight comprise a group of "fishes that thorough than Woodward's (1889a), but crunch" rather than a monophyletic group. some systematic and taxonomic suggestions Although Agassiz (1837) separated Hybodus concerning hybodonts have been strongly and Acrodus into different higher taxa, he criticized (Koken, 1907; Stensio, 1921; was aware of Owen's work (published 1840) Kuhn, 1945). An early attempt to distinguish on tooth morphology which demonstrated Hybodus from Heterodontus was made by histological similarities between Hybodus Brown (1900), who concluded that the jaw and Acrodus teeth. The first work to draw suspension of Hybodus was more like that attention to similarities between Hybodus of hexanchoids than Heterodontus (for a de- and Acrodus in parts other than their teeth tailed discussion, see below). He effectively was by Day (1864, pp. 57-65). Teeth of Ac- regarded hexanchoids as a sister group to rodus anningiae Ag. were found associated Hybodus and other living sharks. Palaeo- with finspines very similar to those named spinax was considered to be more closely Hybodus incurvus by Agassiz (1837), who allied to squaloids and Heterodontus than had already suggested that these finspines Hybodus. While Brown's (1900) hypothesis were possessed by H. reticulatus, the type of relationships is questionable today, it was species of Hybodus. Day (1864) stressed the a more precise statement than any hitherto, close relationship between Hybodus and Ac- and can be expressed in the form of a clado- rodus. Instead of separating Hybodus and gram (fig. IC). Brown's (1900) views of hyb- Acrodus from Heterodontus, however, odont relationships were supported by many subsequent authors preferred to regard Goodrich (1909, 1930), except for the sys- Hybodus as a cestraciont. Woodward (1889a) tematic position of Heterodontus. placed Orodus, Sphenacanthus, Tristychius, The earliest suggestion that hybodonts Hybodus, Acrodus, Asteracanthus, Palaeo- should be ranked apart from all living sharks spinax, Synechodus, and Heterodontus seems to be in Zittel (1911). Hybodonts were (Cestracion) into the family Cestraciontidae, separated from cestracionts (which included on the grounds that "no distinctive charac- two important fossil genera, Palaeospinax teristics of value having been discovered, the and Synechodus). Brough (1935) also regard- so-called Orodontidae and Hybodontidae are ed hybodonts as an independent group and included in this family." This effectively listed the following characters by which he united hybodonts with many living sharks recognized them: (apart from squaloids, Squatina and batoids) "i. Body fusiform: of normal shape; fins of into the suborder Asterospondyli. The sub- moderate size. order was recognized on the basis of verte- ii. Anal fin very posteriorly placed. bral characters, whereas hybodonts lack cal- iii. Pectoral fin tribasal. cified vertebrae. However, asterospondylous iv. Teeth always separate, never fused. centra are present in Palaeospinax and Syn- v. The two dorsal fins dissimilar; the first dor- echodus, which Woodward (1889a) consid- sal having a spine lying at a low angle and ered to be hybodontids because of general being without radial cartilage; the second similarities in their teeth. Elsewhere he at- having the finspine upstanding at a higher angle and possessing a row of radial carti- tempted to derive hexanchoids and Chlam- lages. ydoselachus from Hybodus (Woodward, vi. Finspines ornamented by a series of lon- 1886b) and attempted to reinforce his view gitudinal furrows and bearing a series of of hybodont interrelationships by comparing denticles on their posterior surfaces. the jaws of Synechodus and hexanchoids vii. Jaws massive, jaw suspension probably (Woodward, 1886a, 1898; cf. Maisey, 1980). amphistylic or early hyostylic. 1982 MAISEY: HYBODONT SHARKS 5

C BROWN (1900)

BROUGH (1935) E MOY-THOMAS (1939a+b)

.so

REGAN (1906) WHITE (1937) G ROMER (1945)

FIG. 1. Some hypotheses of relationship between hybodont and Recent elasmobranchs, interpreted cladistically. Hybodonts are the sister group of Recent elasmobranchs in (E); in all other schemes some Recent elasmobranchs are the sister group of hybodonts and remaining Recent forms. 6 AMERICAN MUSEUM NOVITATES NO. 2724

viii. Sphenonchus head spines usually, if not Several different hypotheses of hybodont invariably, present in the males." relationships can be retrieved from the literature (fig. 1). According to Woodward's These characters are considered in the fol- (1889a) classification squaloids, Squatina, lowing comparative section. Their relative and batoids form a sister group to hybodonts importance as hybodont synapomorphies is and other Recent sharks, and hexanchoids discussed toward the end of this paper. form a sister group to hybodonts, Hetero- Brough (1935) went on to demonstrate the dontus, and galeomorphs. Glikman's (1964) equivocal nature of characters previously scheme produces similar results; hybodonts used to unite hybodonts with Heterodontus. and lamnoids are contained by a sister group Brough's (1935) hypothesis of relationships to other Recent sharks and rays. According can be expressed cladistically (fig. IE), as to Brown (1900), hexanchoids are a sister can that of Moy-Thomas (1939a, 1939b), who group to Hybodus and remaining Recent retained the Hybodontidae as a separate tax- sharks and rays. Goodrich (1909, 1930) re- on, closely following Zittel (1911) except that fined this hypothesis slightly by combining the order Protoselachii was erected to con- Hybodus and Heterodontus as a separate tain the Hybodontidae and Tristychiidae. group. According to Brough (1935), Moy- Young (1962) virtually followed Moy-Thom- Thomas (1939a, 1939b) hybodonts form the as (1939a, 1939b) but removed Heterodontus sister group of all Recent sharks and rays. from the Euselachii to the Protoselachii (fig. According to Regan (1906), White (1937), IF). Romer (1945) regarded Synechodus as Romer (1945), Berg (1955), and Patterson a hybodont, but placed Palaeospinax in the (1967) batoids are the sister group of hybo- Heterodontidae. No evidence to suggest sep- donts and Recent sharks. aration of these taxa other than at generic level has ever been presented, however, and there is certainly no justification to placing COMPARATIVE ANATOMY OF them in separate suborders. MESOZOIC HYBODONT SHARKS Patterson (1966) retained the Hybodonti- dae as a distinct taxon, but following Berg THE FORMS UNDER CONSIDERATION (1955) he subsequently relegated the group The best known Mesozoic hybodonts in- to the Heterodontiformes (Patterson, 1967). clude Hybodus hauffianus (Fraas, 1889, However, Berg's (1955) classification does 1896; Brown, 1900; Jaekel, 1906; Koken, little more than resurrect Woodward's 1907), H. fraasi (Brown, 1900), H. dela- (1889a) scheme in placing Tristychius, hybo- bechei (Charlesworth, 1839; Day, 1864; donts, Palaeospinax, and Heterodontus into Woodward, 1889a, 1889b), H. basanus a single order. (Egerton, 1845; Woodward, 1889a, 1916, Schaeffer's (1967b) morphological grade 1919), H. cassangensis (Teixeira, 1954, concept, intended as a tentative discussion 1956, 1978), and Lissodus africanus (Broom, of shark evolution without recourse to for- 1909; Brough, 1935). Partial skeletons and mal taxonomy, illustrated the mosaic of jaws of Acrodus, Asteracanthus, Palaeo- shared derived and primitive characters in bates, and other hybodonts have also been modern and fossil sharks. Formalization of described (e.g., Owen, 1869;- Woodward, this work by Blot (1969) resulted in a phe- 1889a; Vidal, 1915; Stensio, 1921; Kuhn, netic classification containing taxa defined 1945; Schaeffer and Mangus, 1976; Rieppel, largely or entirely on plesiomorphic charac- 1981). There are also partial skeletons and ters. The clade Hybodontiformes included isolated skeletal elements of Mesozoic hyb- hybodonts, ctenacanths, edestids, Heter- odonts in the British Museum (Natural His- odontus, Synechodus, Palaeospinax, Ortha- tory) collections which I examined in pre- codus, Anacorax, hexanchoids, and Chlam- paring this paper, and I have also examined ydoselachus. The clade Euselachiformes many of the previously described specimens contained remaining Recent sharks and rays. either directly or from peels and casts. 1982 MAISEY: HYBODONT SHARKS 7

pop pop

ora

FIG. 2. Neurocranium of Hybodus basanus, resstored in dorsal (A), ventral (B), and lateral (C) views. For abbreviations see page 2.

THE NEUROCRANIUM tremely large jugular canal. The maximum cranial width (between the processes) is The only Mesozoic hybodontid in which slightly less than its total length. The otico- the neurocranium is well known is Hybodus occipital region is short, although the deeply basanus (fig. 2; Woodward, 1916, 1919; concave articular cotylus of the occiput Maisey, in prep.). The dorsal surface of the forms a prominent posterior extension neurocranium inclines steeply toward the bounded laterally by triangular expansions. snout, and there are large, downturned post- There is a low occipital crest running from orbital processes, each penetrated by an ex- the foramen magnum forward to the poste- 8 AMERICAN MUSEUM NOVITATES NO. 2724

endf

pop

A af Mc

ect

af supc

FIG. 3. Hybodus delabechei head region; (A) from Woodward (1889b, pl. 1, fig. 1) reversed view to facilitate comparison; (B) from Woodward (1889a, pl. 8, fig. 1). 1982 MAISEY: HYBODONT SHARKS

psc

44 pq

B qf0 lab Mc palp af FIG. 4. (A) Hybodus fraasi, from Brown (1900, pl. XVL, fig. 1); (B) Hybodus hauffianus, from Koken (1907, pl. II) reversed view to facilitate comparison. 10 AMERICAN MUSEUM NOVITATES NO. 2724

rior end of a large, anteroposteriorly extend- process. Overlying this process is a groove, ed endolymphatic (parietal) fossa, between an ectethmoid process, and the proximal the domelike dorsal surfaces of the otic cap- portion of the olfactory capsule. The ar- sules. Immediately behind the postorbital rangement is distinctly different from any liv- process and dorsal to the jugular canal is a ing shark, even Heterodontus, which has a peculiar lateral expansion which, with re- groove to accommodate the palatoquadrate, spect to the otic capsule and positions of the but which lacks an ethmopalatine process jugular vein and hyomandibula, corresponds (Luther, 1908; Haller, 1926; Holmgren, 1943; to the lateral otic process (terminology after Jollie, 1962; Moss, 1962, 1972; Nobiling, Schaeffer, 1981) of Chlamydoselachus and 1977). Tamiobatis. There is no evidence in H. bas- Figures 3 and 4 are compiled after illustra- anus of a persistent otico-occipital fissure tions of Woodward (1889a, 1889b), Brown (fissura metotica) either dorsally or ventral- (1900), and Koken (1907) and show that, in ly, but there is a large vagus-glossopharyn- general, the heads of H. basanus, H. dela- geal fossa lateral to the occiput. Thus H. bechei, H. hauffianus, and H. fraasi are an- basanus resembles living sharks in lacking a atomically similar, e.g., in the shape of the continuous adult otico-occipital fissure. Such neurocranial roof, elongate endolymphatic a continuous opening has been described fossa, rounded anterior fontanelle, and jaws. only in Recent shark embryos (see Goodrich, Now that H. basanus is known in greater 1930, and Schaeffer, 1981 for references), detail, comparable features can be provision- and adult xenacanth, Tamiobatis, and cten- ally identified on the heads of these other acanth neurocrania (Schaeffer, 1981), but it species, e.g., downturned postorbital pro- is apparently also present in other Paleozoic cesses and large jugular canal. In addition sharks (Zangerl, personal commun.). The some of Brown's (1900) and Koken's (1907) embryonic development of chimaeras is still material can be reinterpreted. poorly known, but adult neurocrania lack a Koken's (1907, fig. 1) illustration of H. fissura metotica and the vagus and glosso- hauffianus probably shows a ventral rather pharyngeal nerve passages do not pass be- than dorsal surface of a neurocranium. The neath the floor of the otic capsule. Although "Parietalgrube" is more probably the inter- Schaeffer (1981) provisionally regarded the nal carotid opening flanked by lateral aortic continuous otico-occipital fissure to be a syn- grooves. Another specimen (Koken's fig. 2) apomorphy of xenacanths and ctenacanths, has a median opening, corresponding to the it is possible that loss or reduction of the parietal fossa, bordered by flattened otic bul- fissure in hybodonts and Recent sharks is lae. It therefore seems to be the dorsal rather derived relative to the condition in some Pa- than the ventral surface. In any case, how- leozoic sharks. ever, it is incomplete since it lacks postor- Ventrally the neurocranium has broad sub- bital processes; articular facets for the hyo- orbital shelves. There is a single median in- mandibulae are present posterodorsally, but ternal carotid foramen immediately anterior the "palatoquadrate articulations" anterior to which is a slight swelling and a smaller to these are asymmetrical and may simply be hypophyseal opening. Shallow grooves in- where postorbital processes have broken dicate the course of the exposed internal ca- away. There is no indication of aortic rotids, orbital arteries, and paired lateral aor- grooves or of an internal carotid foramen. tae. The aortae were enclosed by cartilage Brown (1900, pl. 16, fig. 2) shows a frag- for a short distance posteriorly. ment of neurocranium with grooves resem- Anteriorly the suborbital shelves are ta- bling the aortic impressions of H. basanus, pered, and extend into an elongate rostral but in the text are unexplained "Gruben." region which separates the palatine ramus of A median depression at the focus of these one palatoquadrate from its antimere. Dorsal grooves, identified as "Hinterrand der vor- to this constriction the preorbital wall is ex- deren Fontanelle" is better explained as the panded laterally into a broad ethmopalatine occipital cotylus, flanked laterally by the for- 1982 MAISEY: HYBODONT SHARKS 11

B occ occ

FIG. 5. Occipital regions of (A) Hybodus basanus compared with (B) Hybodus hauffianus (after Brown, 1900, pl. 16, fig. 2). In (B) the postorbital processes have broken away to expose the floor of the otic capsules. amina where lateral aortae exit after a short the forms under discussion. Of H. fraasi, incursion into the basicranium, as in H. bas- Brown (1900, p. 152) wrote: "Das am Schad- anus (fig. 5). The structures labeled "Su- el befestigte Ende des Hyomandibulare liegt praorbitalleiste" would then be the floors of in einer in die Periotickapsel eingesenkten the otic capsule, after postorbital processes Vertiefung und der Knorpel der letzteren have broken away. There is a recurrent ten- setzt sich in einem stumpfen Pteroticfortsatz dency for these processes to become de- fort" (the end of the hyomandibula attached tached; Brown (1900, p. 160, and pl. 16, fig. to the cranium lies in a depression in the 3) notes this in another specimen of H. hauf- periotic capsule and the cartilage of the lat- fianus, and they broke away in H. basanus ter continues as a blunt pterotic process). In (BM[NH] 40718) during preparation (also H. hauffianus, the hyomandibular articula- noted above regarding Koken's specimen). tions on the braincase are shown in Koken's This suggests that the postorbital processes (1907) figure 2, but neither Brown (1900) nor were poorly attached to the braincase, with Koken (1907) were very definite about the only thin cartilage dorsal and ventral to the arrangement. Brown merely comments that large jugular canal. the hyomandibula borders on the spiracular Comparison of Brown's plate 16, figure 3 area posteriorly. with H. basanus suggests that the occipital region was even shorter in H. hauffianus (fig. 4). The median groove in the original THE JAWS figure may be where the floor of an uncalci- In Hybodus basanus the palatoquadrate is fied notochordal area has collapsed. A simi- long and shaped so as to fit against the brain- lar groove occurs in a fragment of hybodon- case for much of its length (fig. 6). There is tid basicranium from the Lias (BM[NH] a strong ethmopalatine articulation anterior- P3356), which also shows paired aortic ly (Maisey, 1980; cf. Woodward, 1889a, grooves. An uncalcified notochordal area is 1916), and the palate bears a low "ethmoi- visible in BM(NH) P50869, but in H. bas- dal" process, but the palatine rami do not anus (BM[NH] 40718), this region is intact meet symphysially. The dorsal margin of the and overlain by prismatic cartilage. The po- palatoquadrate rises into the orbit over the sition of efferent hyoidean arteries cannot be suborbital shelf and below the ethmopalatine determined in any specimen. process, passing posteriorly beneath the The hyomandibular articulations with the postorbital process. There is no articular sur- neurocranium are similarly positioned in all face on the quadrate moiety, however, and 12 AMERICAN MUSEUM NOVITATES NO. 2724

.qf

pq ch

lab lab FIG. 6. Hybodus basanus, original. Restoration of head and jaws in lateral view. For abbreviations see page 2.

the jaws were probably able to slide antero- tant anchorage for some superficial jaw mus- posteriorly beneath the braincase. The me- cles (Gegenbaur, 1898; Luther, 1908; Haller, sial surface of the palatoquadrate is smooth, 1926; Daniel, 1934; Moss, 1962; Nobiling, with a shallow dental groove anteriorly and 1977), they are not so well developed as in a faint thickening farther back. A pro- Hybodus. Labial cartilages are not known in nounced dorsal postorbital (otic) flange is ab- many fossil sharks, which suggests that gen- sent, but there is a lateral quadrate expan- erally these elements were not so well de- sion overlying a deep adductor fossa. In veloped as in H. basanus. The labial carti- lateral view a dorsoventral constriction of lages of other hybodonts are incompletely the palatoquadrate at the anterior end of the known, but H. hauffianus has large upper adductor fossa vaguely divides the element and lower cartilages (fig. 4B) which seem to into palatine and quadrate components. have been arranged much as in H. basanus. The lower jaws meet at a narrow sym- A well-developed labial cartilage complex of physis. They do not seem to be significantly this type was therefore probably widespread different from other elasmobranch lower among Mesozoic hybodonts and may be a jaws, except for the presence of lateral synapomorphy for the group. anteroposterior grooves which house the Where known, the jaws of Hybodus and lower anterior labial cartilages (fig. 6). The Acrodus sp. seem to be similar (fig. 7). Ko- labial cartilages are massive, and there are ken (1907) described a specimen of H. hauf- five cartilages per side in H. basanus. Labial fianus (p. 267, fig. 1) with palatoquadrates cartilages seem to be as large in H. hauffi- supposedly in visceral view and showing an anus and H. fraasi as in H. basanus. oval mark locating ligamentous connections Among Recent sharks there are usually three between the palatoquadrates (fig. 7B). In H. or fewer labial cartilages on each side of the basanus there is no trace whatever of such mouth. These cartilages are usually small a prominence (fig. 7A). The "oval mark" ap- and, while they sometimes provide impor- parently represents the prominent ethmoid 1982 MAISEY: HYBODONT SHARKS 13

polp palp e art

A ~~~~~~~~~B palp

~e.art

C D -

FIG. 7. Jaws of various hybodonts in lateral view, right side; (A) Hybodus basanus, original; (B) H. hauffianus, after Koken (1907); (C) Acrodus sp., after Kuhn (1945, fig. 1); (D) Acrodus ?nobilis (BM [NH] P50809). Jaws of H. basanus drawn as if flattened out (cf. fig. 6) to facilitate comparison with other forms which have been compressed slightly in preservation.

process, but this would not be visible in vis- constricted between its quadrate and pala- ceral view. Also the quadrate groove can be tine regions. These similarities in the pala- seen in Koken's (1907) figure. This would be toquadrate support the view that Hybodus hidden in visceral view. Thirdly, there is no and Acrodus sp. are closely related. trace of a dental groove for the teeth, al- A well-preserved pair of small Asteracan- though this ought to be seen in visceral view. thus palatoquadrates (BM[NH] P12614) are The palatoquadrates of the Tubingen speci- similar to those of Hybodus except that each men may therefore actually lie in lateral one expands anteriorly into a semicircular view. Acrodus sp. from the Triassic of Tes- buttress to support the upper toothplate and sin (Kuhn, 1945; Rieppel, 1981) has remark- is traversed diagonally by grooves marking ably similarjaws to H. basanus and H. hauf- the position of the tooth files (fig. 8). There fianus (fig. 7C). In all these forms, the is a deep adductor fossa which is broader quadrate region forms a prominent lateral than in Hybodus and which could have flange overhanging a deep adductor fossa, housed correspondingly larger mandibular and the palatoquadrate is dorsoventrally adductors. (fig. 8C). 14 AMERICAN MUSEUM NOVITATES NO. 2724

C D \J FIG. 8. Asteracanthus sp. palatoquadrates; (A) anterior view, after Peyer (1946); (B-D) from BM[NH] P12614 (slightly restored); (B) anterior view; (C) lateral view, left side; (D) dorsal view.

In Asteracanthus, there is flat surface me- teracanthus dentition clearly show that the sial to the semicircular buttress. If this principal replacement tooth files diverge an- formed a symphysis as Peyer (1946) thought, teriorly as they pass labio-lingually, as in the caudal intemasal wall would have been Heterodontus. In Asteracanthus these tooth excluded from the roof of the mouth, and the files are paralleled by grooves on the jaw ele- tooth buttress would be angled upward lat- ments both in Peyer's specimens and in erally. If the jaws are restored as in Hybo- BM(NH) P12614. The grooves on the Swiss dus, however, the buttress would be flatter material are convergent anteriorly, however, and the symphysis would be small (fig. 8B, i.e., not parallel but in opposition to the D) but jaw protrusion could occur. tooth files. This condition can be duplicated The latter interpretation of the Asteracan- in BM(NH) P12614, if the mandibles are thus jaw arrangement differs profoundly oriented as Peyer shows. Peyer's (1946) fig- from that of Peyer (1946, figure shown here ures do not show the occlusal surfaces of as fig. 8A), in which the very long symphysis Meckel's cartilage or the palatoquadrate, leaves nowhere for the neurocranium to ar- and consequently the discrepancy between ticulate, and in which the basicranium would the jaws, the orientation of tooth files, and not form the roof of the mouth. Peyer's dental grooves is not immediately obvious. (1946, figs. 9, 11, and 13) illustrations of As- However, enough of the groove arrangement 1982 MAISEY: HYBODONT SHARKS 15 can be seen in his oblique view of the re- teracanthus and passed dorsal to the quad- stored jaws (his fig. 12) to show that grooves rate part of the palatoquadrates. on the palatoquadrates converge anteriorly and curve around the labial margin in a direction quite contrary to the orientation of BRANCHIAL ARCHES tooth files. A similar effect results when There are five gill arches in H. basanus, BM(NH) P12614 is misoriented, but it is then but only the cerato- and epibranchials are immediately obvious that the mandibular clearly calcified. There is no sign of a basi- joint is vertical rather than horizontal, and branchial series, nor of hypobranchials al- could not possibly function in this way. Re- though they presumably were present. orientation of the jaws so that the grooves Pharyngobranchials are not known in H. diverge anteriorly also brings the jaw joint basanus but are noted by Brown (1900) in back into a more normal orientation, as well H. fraasi. It is not possible to determine as providing a space for the neurocranium their precise number in H. fraasi, but I have which Peyer's (1946) model lacks. Put sim- identified five separate posteriorly directed ply, Peyer (1946) has mistaken labial and lin- pharyngobranchials in H. cassangensis (fig. gual, and left and right surfaces of his As- 9A, D). In many living sharks pharyngobran- teracanthus jaw material. chials IV and V are fused (usually also including the fifth epibranchial). The only THE HYoID forms in which this is not the case are those ARCH with consistently more than five gills (hex- The hyoid arch of H. basanus is unusual anchoids, Chlamydoselachus). Even here, because the hyomandibula passes dorsal to the hindmost pharyngobranchials are not the quadrate region (fig. 6). Its proximal end normally discrete elements; in Notorhyn- articulates just anterior to the vagus-glosso- chus and Hexanchus pharyngobranchials V pharyngeal fossa in typical selachian fashion and VI are united, in Heptranchias VI and (see Schaeffer, 1981), however, and its distal VII, and in Chlamydoselachus the sixth end meets the mandible mesial to the jaw epi- and pharyngobranchials are united. joint. Therefore, the peculiar relationship be- Therefore, living sharks seem to differ from tween the hyomandibula and palatoquadrate Hybodus in having their posterior pharyngo- seems to result from the presence of the lat- branchials modified to some extent, which eral quadrate flange on the latter (Maisey, may be a synapomorphy of Recent sharks. 1980). The distinctively shaped hyomandib- Unfortunately, knowledge of other fossil ula is similar in H. basanus, H. hauffianus, elasmobranch pharyngobranchials is too H. fraasi, H. delabechei, and Acrodus. scanty to strengthen or refine this statement. In H. fraasi this element is slightly curved, Nelson (1969) noted that the presence of with broad, flat lateral surfaces and expand- a gap between the basihyal and basibranchi- ed upper and lower ends. The upper end lay als, and the orientation of hypobranchials in an otic articular facet and the lower round- (directed posteriorly toward the midline) are ed end was closely connected with the pos- important differences between Recent sharks terior surface of the jaw arch. As Brown and other gnathostomes. Although hypo- (1900, p. 154) notes, the hyomandibula branchials and basibranchials are unknown makes much greater contact with the neu- in H. basanus, the shape and arrangement rocranium than in hexanchoids. of the basihyal and first ceratobranchial sug- In Asteracanthus the hyomandibula is un- gests that hypobranchials and basibranchials known, but the shape of the posterior part of were arranged as in Recent elasmobranchs. the palatoquadrate closely resembles that of The hypobranchials of H. hauffianus are di- H. basanus (cf.- figs. 5-7). Therefore, it rected posteriorly toward the midline (see seems probable that the hyomandibula was Brown, 1900, pl. 16, fig. 1). As far as is positioned similarly in H. basanus and As- known, a comparable arrangement does not 0)~~~~~~~~~~~~~~~~~ L .0~~~~~~~

LL.

.0~~~~~~~~~%

0.~~~~~~~

ScnZ

.0~~~~~~~~~~~~~~0C

OQo 1982 MAISEY: HYBODONT SHARKS occur in Cladoselache (Dean, 1909) or Co- is verbunden gewesen sein" (the palatoquad- belodus (Zangerl and Case, 1976). Data for rate rises into two processes, of which the xenacanths is ambiguous (Koken, 1889; wider one has a direct articulation with the Fritsch, 1895; Jaekel, 1895; Reis, 1897; Nel- proc. postorbitalis on the anterior part of the son, 1969); for a discussion of the visceral process. The anterior process is not as wide skeleton of chimaeras and sharks see Nelson and was probably ligamentously connected (1969). with the corresponding proc. preorbitalis [by means of an ethmopalatine ligament]). Later MANDIBULAR (p. 268) he says: "Der Processus orbitalis ist SUSPENSION sehr stark und ubernimmt die kraniale Ver- There are fundamental differences of opin- bindung, welche aber ligamentos bleibt, ion in Brown (1900), Jaekel (1906), Koken nicht gelenkartig wird. Die Verbindung des (1907), and Woodward (1889a, 1916) regard- quadratalen Abschnittes wird nicht gelost, ing mandibular suspension in Mesozoic hyb- aber das Hyomandibulare erhiilt doch schon odonts. Taking these in chronological order, im wesentlichen die Funktion, die es z.B. by Woodward (1889a, pl. 12, fig. 1) figured well- Scymnus ausiibt; es gelenkt mit dem Crani- preserved Hybodus basanus jaws, (BM[NH] um proximal, mit dem Kieferbogen distal" P2082) and commented: "There is no evi- (the processus orbitalis is very strong and dence of an articulation of the pterygo-quad- takes over the cranial connection, which, rate with the cranium, either in advance of however, remains ligamentous and does not or behind the orbit." Brown (1900, pl. 168, form an articulation. The connection of the fig. 5) thought there was a strong postorbital quadrate region does not disappear, but the articulation, as in living hexanchoids, and hyomandibula more or less functions as in also a strong orbital process. "Auf der an- Scymnus, for instance, articulating proxi- deren Seite treten am Schiidel von Hybodus mally with the cranium and distally with the (fig. 5.B) vor allen hervor die bedeutende mandibular arch). postorbitale Articulation des Palatoquadra- Thus he considered there to be a good tum, die nach vome liegenden Augenhohlen postorbital articulation and a weaker (liga- und das lange kraftige Hyomandibulare" (on mentous) anterior one, but the material on ... the skull of Hybodus [fig. 5.B] one main- which this interpretation was largely based ly notices the significant postorbital articu- is also fragmentary. Woodward's (1889a) de- lation of the palatoquadrate, the anteriorly nial of a postorbital articulation was moder- positioned orbits, and the long, strong hyo- ated in 1916, page 7: "It can scarcely have mandibula). As already noted, some of the articulated with the postorbital prominence material used to formulate this restoration of the cranium." He maintained his views on was damaged. His restoration consequently the preorbital articulation, however (p. 3): shows the postorbital process incorrectly "pterygoquadrate cartilage not articulated and attempts to make a postorbital articula- with the preorbital region of the skull." Re- tion out of what may be broken surfaces; gan (1906) followed Brown (1900) and Jaekel Jaekel's (1906, fig. 2) reconstruction is simi- (1906) in considering that the palatoquadrate lar in this respect. of Hybodus has a postorbital articulation as Koken's (1907) account differs from all of in hexanchoids but Regan (1906) and Smith the above. On page 266 he remarks that: (1942) followed Woodward (1889a) in deny- "Das Palatoquadratum erhebt sich zu zwei ing Hybodus a preorbital (ethmoidal) artic- Fortsiitzen, von denen der breitere hinten ulation between the neurocranium and pala- eine direkte Gelenkung mit dem Proc. post- toquadrate. In summarizing Hybodus and orbitalis vermittelt; sie liegt in der vorderen Heterodontus jaw suspension, Smith (1942) Hiilfte dieses Fortsatzes. Der vordere Fort- wrote: "In view of the well-known difficul- satz ist nicht so breit und diirfte ligamentos ties attending the restoration of the fossil (durch ein Ethmopalatinligament) mit dem vertebrate remains to life-like attitudes, one genau korrespondierenden Proc. praeorbital- suspects there is a flaw in the data some- 18 AMERICAN MUSEUM NOVITATES NO. 2724 where . . ." Actually, there were two im- "typically amphistylic" suspension. Good- portant issues causing confusion at the time rich's (1909) statement is falsified on these Smith wrote. One resulted from prior mis- two counts and much of Smith's (1942) sub- interpretation of Hybodus from Germany, in sequent confusion is clarified once this is re- which the large, downturned postorbital pro- alized. cesses were mistaken for palatoquadrate otic We can now see that the ethmoidal artic- processes (preceding discussion; see Fraas, ulation in Hybodus basanus is much better 1896; Brown, 1900; Jaekel, 1906; Koken, than anyone has suggested and that while the 1907). The other problem stems from Smith's postorbital part of the braincase does make own confusion over Synechodus dubrisien- contact with the palatoquadrate, the union is sis. This was originally referred to Hybodus at best a sliding articulation lacking any well- (e.g., Woodward, 1886a). It appears from defined facets or joints. Smith's (1942) discussion that he mistakenly The length of the snout is uncertain in H. thought Synechodus dubrisiensis and Hyb- basanus and this region may have been pre- odus dubrisiensis were different fishes. On pared away in Koken's (1907) more complete page 697 he wrote: "Similar differences (of specimen of H. hauffianus. In H. fraasi and the teeth) occur in the three genera of the Lissodus africanus there is evidence for a Cestraciontidae. The teeth of Synechodus moderately long snout making the jaws less (text-figure 31) are much like those of Hyb- "terminal" (Brown, 1900; Brough, 1935). odus (text-figures 29 and 30) except that the anterior teeth of Synechodus are larger than THE AXIAL SKELETON the posterior ones. The teeth of Palaeospi- nax show progress in the direction taken by Although Hybodus basanus offers us ex- Heterodontus . . ." Smith illustrates his ex- cellent cranial material, its postcranial skele- ample with Woodward's (1886a, fig. 12) fig- ton is poorly known. However, some scraps of ure of Synechodus dubrisiensis. However, postcranial skeleton have been referred to this he later (p. 700) stated that: "In Hybodus species (Woodward, 1891, 1916). Relatively hauffianus, according to Jaekel (1906), the complete postcranial skeletons are known suspension of the jaws is amphistylic .... for H. hauffianus, H. fraasi, H. cassangen- The skull of Hybodus dubrisiensis, as de- sis, and Lissodus africanus. Few characters scribed by Woodward (1886) is even more are known in H. basanus. The axial noto- typically amphistylic, resembling that of chord is "persistent," i.e., not constricted, Heptanchus." This was reaffirmed on page septate or calcified in any way that can be 700: "This view accords with Woodward's determined, but there are calcified cartilagi- observation (1886) that the skull of Hybodus nous dorsal (neural) and ventral (haemal) ele- dubrisiensis is typically amphistylic, and ments (Woodward, 1891, 1916; Brown, 1900; with Jaekel's interpretation of the skull in Jaekel, 1906; Koken, 1907). Hybodus bas- Hybodus hauffianus ..., but it does not anus is excluded from discussion of remain- harmonize with Woodward's later statement ing characters since they are not known in (1916) that the pterygoquadrate (palatoquad- this species. However, I think it extremely rate) of Hybodus basanus can scarcely have unlikely that this species differed significant- articulated with the postorbital prominence ly from the others in its postcranial anatomy. of the cranium." Perhaps the most interesting peculiarity of The "view" of which Smith wrote was hybodont axial skeletons is the well-devel- that of Goodrich (1909): "it is well estab- oped ribcage (fig. 10). There are 11 paired lished that Hybodus and Synechodus had ribs in H. cassangensis, 11 or 12 in H. hauf- typically amphistylic skulls, with the pala- fianus, and possibly 12 in H. fraasi, although toquadrate and hyomandibula as in the No- the latter species seems to have more ribs tidanidae and other primitive Elasmo- because some are exposed from both sides branchs." Not only is Synechodus not a of the ribcage in Brown's (1900) specimen. hybodont, but Hybodus does not have a Above the chordal space there are numerous 1982 MAISEY: HYBODONT SHARKS 19

I.. a3 aJ .0,; )DJO scap j ... .. bv p /; 0 .-

pdsp - adsp

-rib C

FIG. 10. Hybodont ribcages; (A) Hybodus cassangensis, from peel; (B) H. fraasi (after Brown, 1900, reversed); (C) H. hauffianus (after Koken, 1907). Not to scale. neural spines, usually in a double series sug- supraneurals. Ribs are borne on the basiven- gesting that left and right halves were not tral elements although some rib support may fused into a complete arch. In neoselachians also be given by the interventral immediately the basidorsals of the vertebral column are anterior to it. Neural spines occupy an inter- squat elements which occupy a vertebral po- vertebral position, while the ribs occupy a sition and are pierced by a ventral spinal vertebral (=intermyotomal) position. nerve root (Shute, 1972). By contrast the in- It has not been possible to locate spinal terdorsals are taller, are pierced by a dorsal nerve foramina in the dorsal arcualia of hyb- spinal nerve root, and give rise to the vault odonts, and thus it is not possible to state of the neural arch, which may be overlain by conclusively whether these dorsal elements 20 AMERICAN MUSEUM NOVITATES NO. 2724

represent basidorsals, interdorsals, or both. However, the bases of these elements are notched as though they curved around other structures including the spinal nerves. From their number and arrangement, the dorsal arcualia seem to correspond to the interdorsals of living elasmobranchs, but do not extend so far ventrally as to enclose a spinal nerve root (figs. 10, 11). Since there appears to be a one-to-one arrangement of interdorsals (or pairs of interdorsals) and ribs, it is probable that all the interdorsals are primary (neural) in this region. Further caudally a one-to-one arrangement of dorsal and ventral arcualia persists, but it is quite possible that A secondary diplospondyly was developed, since in living elasmobranchs the number of secondary basidorsals, interdorsals, and basiventrals is increased uniformly. Without knowledge of their innervation, however, we cannot tell primary from secondary dorsal arcualia in Hybodus spp. Despite generally good preservation of the axial skeleton in Hybodus, supraneural (and infrahaemal) elements have not been found. Although ribs have been noted in hybodonts on numerous occasions (e.g., Fraas, 1896; Brown, 1900; Jaekel, 1906; Koken, 1907; Zittel, 1911; Woodward, 1916; Zangerl, 1979; Maisey, 1975) they have been paid little attention. Hybodont ribs differ from the intermuscular ribs of Recent sharks and rays in two respects; hybodont ribs are much longer than those of modern elasmobranchs, and are expanded proximally against the no- em -em tochordal space to include the basiventral cartilages, which are generally separate in Recent sharks and rays (see below). Impres- sions of myotomal muscles have not been IVhs 'hs described in any hybodonts, and there is consequently no direct evidence for the relationship between their ribs and the my- hm ocommata, horizontal septum, and somatic peritoneum. C N D MtV Ribs of Recent sharks and rays (fig. 1 lC) FIG. 11. (A, B) Hybodus cassangensis axial skeleton, restored in (B); (C, D) diagrammatic develop centrifugally from cartilaginous an- sections through trunk of modern shark (C) and lagen adjacent to the vertebra, and ultimately Hybodus (D) showing different rib positions. occupy an intermuscular position at the intersections of myocommata and the horizontal septum (Balfour, 1878; Goppert, 1895; Devillers, 1954; Shute, 1972). The view that Schauinsland, 1906; Goodrich, 1909, 1930; ribs fall into two categories, dorsal and ven- 1982 MAISEY: HYBODONT SHARKS 21 tral (e.g., Regan, 1906; Goodrich, 1909, p. termuscular position would be if the horizon- 68; 1930, p. 20 et seq.) has been revised rad- tal septum was extended ventrolaterally, ically by subsequent authors (e.g., Devillers, which would also involve extension of epax- 1954; Rosen et al., 1981), who have shown ial muscles ventrally so as to envelop the that simple topographic criteria are inade- hypaxial muscles over the flank region. Such quate for establishing homology between the fundamental modification of the trunk mus- ribs of different gnathostome groups. Rosen cles is unlikely and would undoubtedly im- et al. (1981, p. 242) concluded that "In fos- pair their mechanical efficiency. Instead it is sils, unless more than one series of ribs is concluded that the ribs of hybodonts and Pal- present, the ribs can ... be identified as dor- ezoic sharks such as Tristychius, Onycho- sal or ventral only by a comparative argu- selache, and xenacanths were topographi- ment (showing that the fossil is a member of cally ventral (pleural), unlike ribs of living a group characterized by one type of rib), or sharks. Elsewhere it has been implied that by evidence of mode of growth (centripetal pleural ribs are primitively present in gnatho- or centrifugal)." Comparison with Recent stomes (Devillers, 1954; Rosen et al., 1981). sharks suggests that the ribs of hybodonts This suggests that non-pleural, but develop- are probably ventral. In view of the length mentally ventral (sensu Emelianov, 1935), of these ribs in Hybodus hauffianus, H. fraa- ribs were secondarily acquired by apodans si, H. cassangensis, and Paleozoic sharks and tetrapods (Rosen et al., 1981). When such as Tristychius, Onychoselache, and shark ribs are considered in conjunction with xenacanths (see below for discussion and other structures (e.g., the appendicular skel- references), it is possible that all these sharks eton, see below), the intermuscular ribs of had pleural rather than intermuscular ventral Recent sharks and rays seem to be derived ribs. Although the intersections between my- relative to pleural ribs. ocommata and the somatic peritoneum of In Recent sharks and rays the ribs usually fishes are folded into a zigzag pattern, the articulate with a basiventral cartilage be- first fold of the hypaxial trunk musculature neath the notochord (fig. 1 iC). The basiven- is usually extensive, and an elongate rib tral cartilage, which arises in the perichordal could therefore develop in the lining of the sheath, partly cups the notochord ventrolat- body cavity without being abruptly bent. erally. Hybodontid ribs rested directly Among Recent sharks and rays the proximal against the notochord without a separate part of the rib (or its basiventral element) basiventral articulation. In H. cassangensis effectively occupies a pleural position mesial and H. hauffianus, the proximal end of the to the myotomal muscles (Emelianov, 1935, rib is expanded into a cup which partly en- figs. 49, 52; Rosen et al., 1981, fig. 55D, E). closed the notochord (fig. lIA, B). These Since the proximal part is ontogenetically ribs give the impression of being a much earliest it is not difficult to envisage subse- elongated basiventral, and may not therefore quent development of the rib taking one of be homologous with the combined basiven- two courses, either laterally along the hori- trals and ribs of living sharks; or it may be zontal septum (to produce an intermuscular that a joint has simply not formed in hybo- rib, as in modern sharks, fig. iC) or ventro- dontid ribs, as sometimes occurs in Recent laterally around the somatic peritoneum so elasmobranchs, e.g., Squatina (Gegenbaur, that the rib remains in a pleural position (fig. 1898, fig. 156), Chlamydoselachus (Goodey, ID). A rib as long as those found in hybo- 1910, pl. 44, fig. 11), Torpedo (Emelianov, donts could not have been accommodated 1935, fig. 52; Rosen et al., 1981, fig. 55E). between the epaxial and hypaxial muscles if Elongated ribs have been described in two these were distributed as in Recent sharks Paleozoic sharks, Tristychius and Onycho- and rays, even if the fish was flattened dor- selache. Dick (1978, p. 91) comments that soventrally (e.g., Torpedo; Emelianov, 1935, anteriorly in Tristychius the chordal space fig. 52; Rosen et al., 1981, fig. 55E). The only "is restricted ventrally by a small basiventral way that hybodont ribs could occupy an in- cartilage (fig. 15A, BV). Most of these car- 22 AMERICAN MUSEUM NOVITATES NO. 2724 tilages have been lost or disturbed but a from a hybodontid. The "Pleuracanthus" short, broad rib is attached to one of those material of Fritsch (1895) should be therefore that still remain." In the trunk region he reexamined to determine the arrangement of comments that the ribs "are flared near their ribs in better specimens more accurately. articular surfaces" and "were not fused to In other gnathostomes, ribs are found in the basiventrals," but there is no evidence sarcopterygians (including tetrapods) and ac- of basiventrals here apart from the ribs them- tinopterygians. However, among their fossil selves, as in Hybodus. Caudally the haemal representatives, ribs are somewhat variably arches are "firmly attached to the bases of distributed. Pleural ribs are well developed the well calcified basiventrals," so that dis- in dipnoans (Goodrich, 1909; Rosen et al., crete basiventrals cannot be found, as in 1981), Diplurus and Chinlea have elongate Hybodus. Koken (1907, p. 13) observed that pleural ribs (Schaeffer, 1952, 1967a), Coel- the haemal arches of Hybodus "sind nicht acanthus has a series of short ribs (Moy- halb so lang wie die Rippen, aber ihnen of- Thomas and Westoll, 1935) but Latimeria fenbar homolog" (are not half as long as the does not (Andrews, 1977). Palaeoniscoids ribs, but are obviously homologous with typically lack ribs or even calcified abdomi- them). Dick (1978) identifies possible internal basiventrals although some evidence for ventral cartilages in Tristychius, but I have basiventrals is found in Tarrasius (Moy- not found them in Hybodus spp. Essentially Thomas, 1934) and Dorypterus (Westoll, similar ribs occur in Onychoselache (Dick 1941). Ribs are typically well developed in and Maisey, 1980). "higher" actinopterygians, e.g., caturids, Ribs also occur in xenacanths such as pachycormids, and semionotids (all of which Xenacanthus and Orthacanthus (Doderlein, lack complete vertebral centra) and amiids, 1889; Fritsch, 1889, 1895; Koken, 1889; Jae- "leptolepids" and other neopterygians kel, 1906). Fritsch (1895, p. 39, fig. 235) (where vertebral centra are primitively pres- shows Xenacanthus decheni with separate ent). Among dipnoans, Griphognathus, ribs and basiventrals, and in reconstructions Rhynchodipterus, and perhaps Dipterus (fig. 236 and pl. 101) indicates both basidor- have ossified centra (Jarvik, 1952; Rosen et sal and interdorsal elements but no interven- al., 1981). Ribs are absent in chimaeras trals. Orthacanthus senckenbergianus (his (Shute, 1972). fig. 234) has basidorsals, interdorsals, basi- Paired ventral elements (basiventrals?) are ventrals, and interventrals. However, some known in some acanthodians (Dean, 1907; specimens, which were referred to "Pleu- Watson, 1937; Miles, 1970) and placoderms racanthus" parallelus (e.g., nos. 84 and 98; (0rvig, 1960; Miles and Westoll, 1968), but Fritsch, 1895, pl. 94, figs. 6, 7, 8, and text ribs are unknown. The absence of interca- fig. 237), have no separation between rib and laries in acanthodians separates them not basiventral, and instead the rib is expanded only from primitive crossopterygians and around the notochordal space as in hybodon- dipnoans (Miles, 1970, p. 351) in which such tids. No basidorsal or interventral elements elements occur (Jarvik, 1952; Denison, 1968; seem to have been present in these speci- Andrews and Westoll, 1970), but also from mens. Fritsch's (op. cit.) illustrations of oth- primitive actinopterygians such as Acipen- er (more complete) specimens referred to ser, Polyodon, and Caturus (Rosen et al., this species are unhelpful; it is possible that 1981, figs. 56C, 58B, 59), Recent sharks and the small fragments on which Fritsch based rays (ibid., fig. 58A), and some xenacanths his restoration of "Pleuracanthus" are from (Fritsch, 1895, figs. 234-236). Small, poorly a hybodont rather than a xenacanth. Hybo- preserved "interdorsals" were noted in Tris- donts were certainly present elsewhere in the tychius by Dick (1978). In many other fossil Permian, and Fritsch (1889, p. 97) described sharks (e.g., cladoselachians, symmoriids, some teeth from the xenacanth-bearing Gas- Goodrichthys) ventral elements are not cal- kohle and Kalksteine formations as Hybodus cified in the trunk region, although ribs were vestitus, which Zidek (1969) confirms are noted by Moy-Thomas (1936, p. 765) in 1982 MAISEY: HYBODONT SHARKS 23

"Ctenacanthus" costellatus. Interventrals coracoarcual and the last coracobranchial are absent in chimaeras (Shute, 1972). Su- muscles, and is perforated by at least one, praneurals and infrahaemals are unknown in possibly two foramina for the branchial ar- acanthodians, placoderms, and all non-neo- tery and diazonal nerve. These foramina are selachian sharks, but supraneurals are pres- close to articular processes for the pectoral ent in palaeoniscoids (see (Moy-Thomas and basals at the ventrolateral extremity of the Miles, 1971, for references), Acipenser (Ro- coracoid (fig. 12). There is a short, blunt pre- sen et al., 1981, fig. 56C), dipnoans (ibid., coracoid process anteriorly, best seen in H. fig. 54A), Eusthenopteron (ibid., fig. 56B), cassangensis, but probably present in H. and Recent sharks and rays (ibid., fig. 56A). fraasi and H. hauffianus. The scapular pro- Segmentation and constriction of the noto- cess is slender, and terminates close to the chord, and concomitant development of base of the anterior finspine, overlying the complete centra, has probably occurred in- first few ribs. In H. cassangensis a separate dependently in Recent elasmobranchs and in suprascapular element may have been pres- actinopterygians (on at least four separate ent (figs. 9A, 12A), but this has not been occasions in the latter, according to Rosen found in other hybodonts. et al., 1981). It is probable that ossified cen- The pectoral basals have a somewhat con- tra were also independently acquired by sar- stricted articulation with the scapulocora- copterygians, and Rosen et al. (1981, p. 248) coid and their arrangement is tribasal. As in consider that amniote and apodan centra Recent sharks, the propterygium of H. cas- have formed independently. The polyspon- sangensis has a larger articulation with the dylous perichordal rings of chimaeras are not glenoid fossa than either the mesopterygium regarded as true centra by Shute (1972, p. or metapterygium. The pectoral radials are 24). arranged in a regular pattern (fig. 12A, D). One radial series articulates with the propterygium, three with the mesopterygium, and THE APPENDICULAR SKELETON at least five preaxial series with the metap- Only the dorsal fins are known in H. bas- terygium. Each radial series apart from the anus (Woodward, 1916), but pectorals, pel- first in regularly jointed as in Recent sharks. vics, anal, caudal, and dorsals have been de- Distal radials of H. cassangensis are pointed scribed in other hybodonts (Fraas, 1896; and short, terminating a long way from the Brown, 1900; Jaekel, 1906; Koken, 1907; fin margins (i.e., the aplesodic condition) Brough, 1935; Teixeira, 1954, 1978). Both which presumably were supported by cera- dorsal fins bear a massive finspine, and have totrichia. Pectorals of H. hauffianus and H. a single triangular basal element which is fraasi are also tribasal (Brown, 1900, Taf. somewhat narrower dorsoventrally in the an- XV, fig. 1, Taf. XVI, fig. 1; Koken, 1907, terior fin. Only the posterior dorsal has a full Taf. 1). According to Koken (ibid.), the complement of calcified radials; in the ante- mesopterygium of H. hauffianus is larger rior fin there is evidence for at most one cal- than the other pectoral basals and carries cified radial, always at the posterior end of seven radial series, while the metapterygium the basal. The peculiar arrangement occurs only has three. in several Paleozoic sharks, e.g., "Cten- I suspect that the specimens on which this acanthus" costellatus, Goodrichthys, Tris- opinion was based were incomplete, causing tychius, and may be a synapomorphy at confusion as to the identity of the basal ele- some higher taxonomic level, although Har- ments. Koken (1907) agrees, however, that ris (1950) regarded this feature as a purely the propterygium carried one radial, and that functional difference from Recent sharks. the other radial series are jointed at least Where known, the pectoral girdle and fins once. As Koken (ibid.) notes, Brown (1900, of hybodonts resemble those of living elas- pl. 15, fig. 1) has reversed the identity of mobranchs. The anterior margin of the cor- propterygia and metapterygia in H. fraasi, acoid region has an elongate groove for the and produced a spurious argument for the 24 AMERICAN MUSEUM NOVITATES NO. 2724

A B

_ _-

C D scap gf

cbrf -z-

cor

pro mes

FIG. 12. Hybodont pectoral fins; (A) Hybodus cassangensis, from peel; (B) H. hauffianus (after Koken, 1907); (C) H. fraasi (after Brown, 1900; basal elements not in natural position, reversed view); (D) H. cassangensis, restoration (original). Not to scale.

progressive reduction of a segmented meta- in many Paleozoic elasmobranchs. Although pterygium (as in xenacanths) via forms like the arrangement of radials is not clear in Symmorium to Hybodus. either H. hauffianus or H. fraasi, they seem Distal radials of H. hauffianus are elon- to have been regularly jointed as in H. cas- gate, pointed, and aplesodic; and the fin sangensis, rather than as restored by Brown gained considerable support from cerato- (1900), Jaekel (1906), or Woodward (1916). trichia (Koken, 1907, Taf. 1). Distal radials Brown's (1900, fig. 3) figure of the pelvic of aplesodic Recent sharks are blunt; only in fins of H. hauffianus agrees in many respects plesodic lamnoids, carcharhinoids, and ba- with the arrangement in H. cassangensis toids are the distal radials more pointed, as (see fig. 13). Both species are represented 1982 MAISEY: HYBODONT SHARKS 25

FIG. 13. Hybodont pelvic fins; (A) Hybodus cassangensis, from peel; (B) H. hauffianus (after Brown, 1900); (A') H. cassangensis, restored; (B') H. hauffianus, restored. Not to scale.

best by males. The anteriormost radial is crushed in the specimen on which Teixeira's much larger than those behind it, especially interpretation was made. The peculiar biax- in H. cassangensis. Teixeira (1954, 1978) ial appearance of the pelvic probably results correctly interpreted one of these, but iden- from superposition of one fin on the other. tified the other as the pelvic girdle. In my There is a puboischiadic bar produced by fu- view (based on examination of peels of one sion of left and right halves of the pelvic gir- specimen), the pelvic girdle is poorly pre- dle, and penetrated distally by a few dia- served in H. cassangensis and has been zonal foramina. A comparable pelvic bar is 26 AMERICAN MUSEUM NOVITATES NO. 2724

also present in Recent sharks and rays. In fied as part of the anal fin by Teixeira, 1954, other fossil sharks, e.g., Cladoselache 1978), but its detailed morphology is ob- (Dean, 1909); "Ctenacanthus" costellatus scure. (Moy-Thomas, 1936); Tristychius arcuatus An anal fin is known in H. hauffianus, H. (Dick, 1978); Onychoselache traquairi (Dick fraasi, and Lissodus africanus, but the only and Maisey, 1980); xenacanths (Fritsch, detailed account of it remains that of Koken 1889, 1895; Jaekel, 1895, 1906); and Cobel- (1907) for H. hauffianus. As noted above, odus (Zangerl and Case, 1976), the pelvic the "anal fin" of H. cassangensis mentioned girdle consists of two separate halves. This by Teixeira (1978) is probably part of the ter- condition is also found in Helodus (Patter- minal clasper cartilage. son, 1965); placoderms (Moy-Thomas and The caudal fin of hybodonts is heterocer- Miles, 1971; Denison, 1978); actinistians cal and, as in Recent sharks, has endoskel- (Moy-Thomas and Miles, 1971); Eusthe- etal support from hypurals in its hypochord- nopteron (Andrews and Westoll, 1970); and al lobe (fig. 14D, E). This characteristic has palaeoniscoids (Aldinger, 1937). Fusion of been used as a synapomorphy of hybodont the pelvic half-girdles in Hybodus and Re- and Recent sharks (e.g., Compagno, 1973, cent elasmobranchs is consequently regard- 1977), on the assumption that the more lu- ed as a synapomorphy of these forms (Com- nate tail of cladodont sharks (e.g., fig. 14A- pagno, 1973). C) is primitive. A weak hypochordal skele- There is an elongate pelvic metapterygium ton occurs in some Paleozoic sharks such as with which about 10 jointed radials articulate Tristychius and Onychoselache, which would in H. cassangensis, but there are fewer in be united with hybodonts and Recent sharks H. hauffianus. However, according to on this basis. However, a similarly shaped Brown (1900, fig. 3 and pl. 16, fig. 1), there caudal fin skeleton characterizes many os- are four or five more segments behind the teichthyans, acanthodians, some placo- metapterygium which also bear jointed ra- derms, and also some agnathans. It is there- dials, although his reconstruction has more fore more parsimonious to regard the caudal segments and radial series than the specimen endoskeleton of hybodonts and living sharks on which it was based. From the specimens, as reflecting a primitive gnathostome condi- H. cassangensis and H. hauffianus had sim- tion. Cladodont lunate caudal fins have been ilar radial patterns. The anteriormost three compared functionally with those of scom- or four radials seem to have articulated di- broids (Harris, 1950; Compagno, 1977); yet rectly with the ends of the pelvic girdle. Of the scombroid tail is not generally regarded these, the anteriormost radial is enlarged and as a primitive osteichthyan one, and among unjointed. Behind this is a series of about 12 both sharks and osteichthyans such a lunate jointed radials. Each radial series is jointed tail is probably derived (Harris, 1950; once, so there is a series of longer proximal Schaeffer and Williams, 1977). and shorter distal radials. In H. cassangensis seven or eight radials articulate proxi- THE DERMAL SKELETON mally with a single metapterygium, but in H. hauffianus the posterior part of this is seg- TEETH: In Recent sharks the tooth base is mented off, so the largest piece of the me- penetrated by enlarged foramina and canals tapterygium carries only five radials. The re- which are specialized for transmitting nerve maining four or five radials articulate with fibers and blood vessels into the tooth. The four jointed metapterygial segments. majority of Mesozoic hybodont teeth lack Behind this there are three (H. hauffianus) such specialized foramina (Patterson, 1966), or four (H. cassangensis) larger intermediate although they are present in Polyacrodus segments before the long basal (myxiptery- spp. (Johnson, 1981), Hybodus cf. plicatilis gial) cartilage. There is a terminal cartilage (Rieppel, 1981), and in several other forms complex in both H. hauffianus (Brown, (see Johnson, 1981, p. 8 for examples). I had 1900, fig. 3) and in H. cassangensis (identi- earlier (Maisey, 1975) suggested that the ab- 1982 MAISEY: HYBODONT SHARKS 27

FIG. 14. Various shark caudal fins; (A) Cladoselache (after Dean, 1909); (B) Cobelodus (after Zan- gerl and Case, 1976); (C) Goodrichthys (after Moy-Thomas, 1936); (D) Hybodus (after Koken, 1907); (E) Isurus (after Garman, 1913). Not to scale. sence of specialized foramina was a primitive selachian teeth are far from unique in this condition. Enlarged nutritive foramina occur respect, which leaves the alternative hypoth- in many Paleozoic cladodont and xenacanth esis that the lack of specialized foramina in teeth, however (e.g., "Cladodus" occiden- some hybodonts is a derived condition. It talis, AMNH 8803; "C." ferox, AMNH remains to be seen whether the derived state 2414; Xenacanthus sp., AMNH 5601, 5408). represents a synapomorphy of certain hy- Significantly, similar specialized foramina bodonts or whether it has arisen indepen- also occur in teeth of Helodus simplex (a ho- dently in various hybodont lineages. locephalan), e.g., AMNH 4359. Thus neo- Reif (1973) considers the enameloid ultra- 28 AMERICAN MUSEUM NOVITATES NO. 2724 structure of neoselachians to be distinctive, dentine by the number and size of dentinal with an outer "shiny layer" ("Glanz- tubules (using the term "pallial dentine" in schicht") underlain by a parallel-fibered lay- the sense recommended by Rieppel). In view er and then by a tangled layer. Similar enam- of the morphological and histological simi- eloid ultrastructure is recognized in teeth of larities between these teeth, it is possible Huenichthys costatus (Reif, 1977) and Reifia that orthodont hybodonts form a monophy- minuta (Duffin, 1980), from the upper Trias- letic group. Comparison with other sharks sic. Tooth enameloid ultrastructure of hy- suggests, however, that the osteodont con- bodonts is different from that of neoselachi- dition, in which a pulp cavity is lacking, is ans, but no consistent morphology has yet derived. This view is supported, not only by been identified. According to Reif (1973), the very restricted distribution of osteodont Hybodus and Acrodus tooth enameloid is teeth among neoselachians, but also by the like that of Paleozoic cladodont teeth, in orthodont nature of xenacanth and "clado- comprising a single-crystallite layer not over- dont" teeth. The osteodont tooth morphol- lain by a shiny layer, whereas the enameloid ogy of Hybodus and Acrodus may conse- of Asteracanthus comprises a single-crystal- quently be regarded as a synapomorphy of lite layer underlain by pleromic hard tissue these genera. Also, the teeth of Asteracan- (a hypermineralized dentine matrix). thus may best be regarded as "modified os- Glikman (1964) attempted to classify teodont," in which a tubular dentinal mor- sharks on the basis of differences in tooth phology has been acquired (see Peyer, 1946, histology. One infraclass (Orthodonti) was Taf. 8, figs. 1-4). This is in accord with other characterized by teeth with a base of "rhi- similarities between Asteracanthus, Hybo- zodentine," and a crown of orthodentine, dus, and Acrodus noted here and elsewhere with a pulp cavity. This group includes (e.g., Maisey, 1978). Thus hybodonts with "cladodonts," xenacanths, Polyacrodus, orthodont tooth histology and specialized Palaeobates, and presumably other hybo- basal nutritive foramina (e.g., Polyacrodus) donts such as Lissodus and Lonchidion, as may simply be plesiomorphic. well as most neoselachians. The other infra- The genus Polyacrodus was originally de- class (Osteodonti) was characterized by fined on the basis of tooth histology (Jaekel, teeth without a pulp cavity, and mainly com- 1889), but external morphological criteria posed of trabecular osteodentine with many have also been used to identify Polyacrodus branching canals. This group includes lam- teeth (Stensio, 1921; Johnson, 1981). Hybo- noids and hybodonts such as Hybodus and dus hauffianus Fraas (1896) was referred to Acrodus. Glikman's (1964) scheme has been Polyacrodus by Jaekel (1906), but Koken criticized by Patterson (1966), on the (1907) and Stensio (1921) disagree with this grounds that it separated certain hybodonts proposal. I have examined teeth from the from others too widely, and by Compagno holotype of H. hauffianus (Staatliches Mu- (1973), who showed that some carcharhinoids seum fur Naturkunde, Stuttgart, no. 8503), (e.g., Dirrhizodon) have osteodont teeth, and concur with the view that this species while closely allied forms (e.g., Hemipristis) should not be included in Polyacrodus, on have orthodont ones. Nevertheless, since it both histological and morphological grounds is now possible to separate hybodont and (Maisey, in prep.). At present, therefore, no neoselachian teeth by means of differences associated Polyacrodus remains are known, in their enameloid ultrastructure, it may be and it is unknown whether this genus pos- possible to draw a valid distinction between sessed typical hybodont finspines and ce- osteodont and orthodont hybodonts. phalic spines. Johnson (1981) considers such According to Rieppel (1981), in Palaeo- an association is likely from his collecting bates, Polyacrodus, and Lonchidion teeth experience in the lower Permian of Texas, the osteodentine of the base is replaced in but admits there is no direct evidence. Sten- the crown by orthodentine, which is distin- sio (1921, 1932) suggested that Polyacrodus guished from a thin overlying layer of pallial and Nemacanthus are synonymous, but 1982 MAISEY: HYBODONT SHARKS 29

FIG. 15. Hybodont cephalic spines. Variation in arrangement on head; (A) H. basanus; (B) H. hauffianus; (C) H. delabechei. Not to scale.

Johnson's (1981) work seems to preclude this Mesozoic deposits, with two records from (see also Maisey, 1977). the Permian (Nielsen, 1932; Branson, 1933). Estes (1964) erected a new hybodont ge- However, I have been sent examples from nus, Lonchidion, for some isolated teeth, the late Pennsylvanian of Kansas (J. Chorn) finspines, and headspines from the Lance and Texas (N. Hotton) that were found in Formation of Wyoming. Subsequently Pat- deposits which have also yielded hybodont terson (1966) assigned numerous British finspines, and Johnson (1981) has described Wealden teeth to this genus, but noted that Permian hybodont cephalic spines. the "anterior teeth" described by Estes There is some variation in the arrangement (which have highly specialized root forami- of cephalic spines in different hybodont na) may pertain to a squatinoid or an orec- species (fig. 15), although with such a small toloboid. This was later verified by Herman sample it is impossible to make more than (1977), who provisionally identified them as general observations. It is generally believed Mesiteia, and by Case (1979) who placed that only male hybodonts possessed cephalic them in Chiloscyllium. Duffin (1981, and in spines. There is presently no evidence that prep.) has concluded that Lonchidion is a they were present in females, in that speci- synonym of Lissodus. mens with pelvic fins lacking claspers are CEPHALIC SPINES ("Sphenonchus"): It also devoid of cephalic spines. Whether ce- has long been recognized that the genus phalic spines developed at maturity or were Sphenonchus Agassiz (1837) is based on ce- also present injuveniles remains speculative. phalic spines which are now assigned to hyb- Acrodus anningiae may represent the juve- odont sharks (e.g., Charlesworth, 1839, nile of A. nobilis (Woodward, 1889b, p. 289). 1845; Day, 1864; Fraas, 1889). Hybodont ce- Several tolerably complete or partial heads phalic spines have received little attention, of A. anningiae have cephalic spines, e.g., although they display some variation which BM(NH) P3152 (a head with three spines may be of systematic value. These spines are preserved), 38125 (another head with three known from several well-represented Meso- spines), P2146 (a head with the bases of three zoic species, including Hybodus basanus, spines), and P2735 (a partial head with one H. hauffianus, H. delabechei, H. medius, spine). Acrodus anningiae, Asteracanthus ornatis- In his diagnosis of Hybodus, Woodward simus, and Lissodus africanus, as well as (1889b, p. 350) stated that there are "two from less complete but associated remains of large hook-shaped, semi-barbed dermal other species, e.g., H. minor, H. raricosta- spines immediately behind each orbit." Ac- tus, H. reticulatus (the type species of Hyb- rodus was said to differ only (p. 279) "in the odus) (Woodward, 1889a). Until fairly re- rounded, non-cuspidate character of the cently, cephalic spines of Sphenonchus-like teeth." Of Asteracanthus, there was less morphology were known principally from certainty (p. 307) "large hook-shaped, semi- 30 AMERICAN MUSEUM NOVITATES NO. 2724 barbed spines present upon the head." rangement probably corresponds most with These structures seem to be confined to the that of H. hauffianus. supratemporal region of the head. In H. bas- Hybodontid cephalic spines have a com- anus, however, no specimen seems to have plex morphology and histology, and yet more than one pair of cephalic spines. Two there is no agreed terminology for their var- pairs occur in other Hybodus spp. and in ious features. The following account there- Acrodus, Lissodus, and perhaps Asteracan- fore introduces some descriptive terms and thus spp., and the number of spines is not by also discusses variation among the features itself a useful taxonomic indicator. recognized (see fig. 16). Each cephalic spine In H. hauffianus one pair of cephalic consists of a large, curved basal platform and spines overlies the region of the lateral otic a strongly recurved, enameled crown. The process, the other lies closer to the endolym- majority of spines have a single retrorse barb phatic (parietal) fossa. In H. basanus the sin- near the apex. The basal plate is convex gle pair corresponds topographically to the anteroposteriorly, and may also be convex more laterally positioned spines of H. hauf- from side to side (H. delabechei, H. reticu- fianus, and the area on either side of the pa- latus, H. hauffianus), or concave (Aster- rietal fossa is overlain by epaxial muscle acanthus ornatissimus). The basal platform (e.g., BM[NH] 6356). Brown (1900, p. 160) is usually drawn out into distinct lateral, me- notes that one pair of H. hauffianus cephalic sial, and posterior lobes. The lateral and pos- spines is smaller than the other, but was un- terior lobes are separated by a lateral inden- certain about their arrangement. In H. de- tation, while the mesial and posterior lobes labechei, BM(NH) 39880 (a crushed head are separated by a mesial indentation. It is and part of the trunk), Woodward (1889b, p. necessary to distinguish mesial and lateral 260) noted "behind the orbit ... are fixed sides because the cephalic spine is generally two large recurved semi-barbed spines, upon asymmetrical and is borne on either side of triradiate bases .... Each of the anterior the head. Thus it becomes possible to distin- pair has two protuberances at the base of the guish left and right cephalic spines from fea- 'crown,' while in the posterior pair these tures in both the basal platform and the are absent." The posterior pair lies in the crown. vicinity of the lateral otic process, as in H. Except in more symmetrical cephalic basanus. The anterior pair is supraorbital in spines, the lateral lobe of the basal platform position, i.e., farther forward than in H. is directed somewhat more anterolaterally hauffianus, and farther from the endolym- than the mesial lobe, e.g., H. basanus, H. phatic fossa (figs. 3B, 15C). A specimen of delabechei, Asteracanthus ornatissimus. In H. medius (BM[NH] 41103; figured in the majority of Jurassic forms these lobes are Woodward, 1889b, pl. 11, fig. 1) has similar short and fairly stout, with about equal width cephalic spines to H. delabechei except that and length when measured from the crown they are apparently not barbed. At least one base, e.g., H. reticulatus, H. hauffianus, but has lateral protuberances on the crown, and in some (particularly Cretaceous) specimens may represent an anterior spine, while the lateral and mesial lobes are elongate, nar- another lacks these. The topographic posi- row, and recurved posteriorly. In Astera- tions of the spines in this specimen, although canthus ornatissimus the lobes are relatively disturbed, support this morphological simi- short and the recurved crown extends farther larity to H. delabechei cephalic spines. The posteriorly than the base (fig. 16A-F). The arrangement of cephalic spines in Lissodus Paleozoic cephalic spines have a broad basal africanus is unclear, but Brough (1935, p. 38) platform and small crown. The lateral, me- writes "there are two on each side of the sial, and posterior lobes merge into each oth- head and they are seen more or less in their er with only shallow marginal indentations natural position behind the orbits." If the separating them. The weak crown is only spines are located as suggested, their ar- slightly recurved, and in both KU 57406 from 1982 MAISEY: HYBODONT SHARKS 31

"I v 7/,

K FIG. 16. Hybodont cephalic spines; (A-F) Asteracanthus ornatissimus, from BM(NH) P12522, Ox- ford Clay, Peterborough, England, ?Left spine, slightly restored, in (A) lateral, (B) posterior, (C) mesial, (D) anterior, (E) dorsal, and (F) basal views; (G, H) cephalic spine, USNM 316515, lower Permian, Archer Co., Texas; (I) Arctacanthus (after Nielsen, 1932); (J, K) cephalic spine, KU57406, upper Pennsylvanian, Shawnee Co., Kansas.

Kansas and USNM 316515 from Texas there (see above), and many other isolated exam- are two pairs of cusps, on either side of the ples in the British Museum (Natural History) principal one (fig. 16G, H, J, K). Woodward collections. Many specimens lack extra (1889b, p. 259) noted cusplike "protuber- cusps, however, including Asteracanthus ances" adjacent to the crown of H. dela- ornatissimus (BM[NH] P12522), Hybodus bechei anterior cephalic spines, and similar minor (P2788), and most Cretaceous spines cusps occur in H. reticulatus, H. medius including those described by Woodward 32 AMERICAN MUSEUM NOVITATES NO. 2724

(1916, pl. 1, fig. 4), Estes (1964, p. 9), Pat- and perhaps (e) development of a distal barb terson (1966, p. 329, figs. 26, 27), and Ca- (not present in KU 57406, unknown in petta and Case (1975, p. 5, pl. 1, figs. 3-6). USNM 316515). It is quite likely that some Although we cannot rule out the possibility of these tendencies were repeated in differ- that all cephalic spines with extra cusps were ent hybodontid lineages, but the possibility anterior ones and those without were pos- remains that detailed morphological studies terior, as in H. delabechei, such a statement of these cephalic spines will produce useful needs to be corroborated by more complete systematic data. fossils. Some curious "Sphenonchus"-like spines The crown of Mesozoic hybodont cephalic were described by Nielsen (1932, p. 53, fig. spines is posteriorly recurved, usually with 5, pl. 1, figs. 2-5) from the Permian of east a sigmoidal profile like that of some modern Greenland (fig. 16I) and Branson (1933, p. shark teeth, e.g., Odontaspis, but with a ter- 175, fig. 1) described similar spines from the minal barb. The barb is connected to the tip middle Phosphoria Formation of Wyoming. of the crown by a posterior crest. A short Both forms are now placed in Nielsen's lateral crest curves away from the barb be- genus Arctacanthus (Branson, 1934). Sub- fore disappearing high up on the crown. A sequently more complete specimens were re- more extensive mesial crest also extends ported from east Greenland (Bendix- from the barb, and passes farther down the Almgreen, 1975). Nielsen (1932) thought the crown before breaking up into several striae. spines were from chimaeroids; Branson This mesial crest is absent from the cephalic (1934) regarded them as rostral teeth of a spine crown in AMNH 6642, although the shark. Both Woodward (1934) and Bendix- short lateral crest is present. A dorsal crest Almgreen (1975) considered the Arctacan- runs from the spine tip, down much of the thus spines to be cephalic ones like those of crown's length, and also breaks up into sev- hybodontids. The Arctacanthus spines are eral striae. In highly asymmetrical spines, considerably more ornamented than such as BM(NH) P12522 (Asteracanthus) "Sphenonchus" spines by retrorse barbs. this crest is displaced toward the lateral side Apart from this feature, and the rather of the spine. In more symmetrical examples, straight crown, Arctacanthus and "Sphen- e.g., AMNH 6642 (Hybodus sp.) it is more onchus" spines are very similar. Unfortu- nearly median. nately the new material reported by Bendix- The crown base is usually striated. Many Almgreen (1975) has yet to be described in striae extend distally to merge with the dor- detail and the morphology of its basal plate sal and mesial crests, but rarely extend far is unknown. enough to meet the lateral crest. Even if the SCALES: The head and trunk of many Me- basal platform is symmetrical, it is often pos- sozoic hybodontids were covered by a dense sible to interpret left and right spines on the shagreen of coarse dermal scales. Of H. de- basis of the lateral and mesial crests, which labechei, Woodward (1889b, p. 260) wrote: are notably disparate in length. "the shagreen granules are conical in shape, In an evolutionary scenario for hybodont with ridges and deep furrows diverging from cephalic spines a gradual transformation the apex, and with a well-defined base; through time is noted toward (a) stronger dif- being, indeed, very suggestive of the small ferentiation of lateral, mesial, and posterior Carboniferous fossils named Petrodus. These lobes and concomitantly deeper marginal in- granules are largest upon the top of the head, dentations of the basal plate, (b) a flatter, less and are especially conspicuous between, and convex anteroposterior profile of the basal immediately in advance of, the orbits; they plate, (c) reduction and suppression of su- are much smaller behind the head, and tend pernumerary cusps, (d) greater elaboration towards fusion into groups of three." Of Ac- and ornamentation of the crown, with a rel- rodus anningiae, he wrote (p. 219): "the sha- ative increase in size over the basal plate, green is similar to that already described 1982 MAISEY: HYBODONT SHARKS 33 upon the head of Hybodus delabechei ... distribution among chondrichthyans and the largest tubercles being upon the frontal cannot be hybodontid synapomorphies, in- region, and the smallest behind; but the lat- cluding the presence of a basal plate, neck ter, so far as preserved, do not exhibit any canals, and basal canals. fusion into groups of three." Thus some hy- Reifs (1978b, p. 117) detailed account of bodontids possessed compound scales, while variation in scale morphology over different others had only simple scales. Possibly A. parts of the head in H. delabechei is based anningiae represents juvenile A. nobilis, as on flattened specimens, in which some parts Woodward (1889b) suggested, and the ab- of the head are difficult to study. However, sence of compound growing scales is growth my examination of uncrushed H. basanus related. Examination of H. basanus speci- heads (Maisey, in prep.) has essentially con- mens has also failed to reveal compound firmed the scale pattern described by Reif. scales. However, in view of Woodward's There are acuminate curved scales dorsally (1889b) comments regarding H. delabechei on the head and laterally over the palato- scales, we might not expect compound quadrates. The lower jaws are covered ex- scales on the head of H. basanus. We cannot ternally by cone-shaped scales. The inter- therefore be sure that the absence of com- mandibular area has both blunt scales pound scales in some hybodontids reflects (multicuspid in H. delabechei but not in H. anything but growth-related factors. basanus), and slender, curved pointed Woodward's description is too simple, scales. The oropharyngeal cavity (including however, according to an extensive review the "tongue") is covered by blunt unicuspid of the morphology and morphogenesis of scales in H. basanus. In H. delabechei there scales in Recent and fossil sharks (Reif, are unicuspid and multicuspid scales in the 1978b, p. 126), in which the following de- roof of the mouth. scription of hybodontid scales is given: Reif (1978b, p. 120) noted that even the "Growing or non-growing scales, very most complex compound hybodontid scales often with high pointed cusps. The cusps comprise six or fewer odontodes, a low num- point either to an apical or to a distal direc- ber which suggests that these scales were tion. The ridges running down from the cusp periodically shed, like non-growing scales. to the processes of the base are very numer- Also, the odontodes are too large to date ous; so are the neck canals. A neck is not from an ontogenetic time when the fish itself very well developed. The basis is flat or was much smaller. A complex mixed pattern slightly convex or concave. Histology can be of scale morphogenesis is suggested, with complex: there is an enameloid cap, the or- unicuspid scales which were replaced fairly thodentine of the crown is very thick; in the regularly (like placoid scales of Recent lower part of the crown and in the base the sharks), and compound scales to which orthodentine can grade into osteodentine. odontodes were periodically added up to a The basal plate is thin, it consists of acellular certain size, when they too were probably bone and has several basal canals." replaced. Certain differences are therefore recog- FINSPINES: The topic of elasmobranch fin- nized between non-growing hybodontid spine morphology is complex, and beyond scales and what are regarded as "'typical" the scope of this paper. Hybodont finspines placoid scales, which are restricted to neo- are unique in several respects (for descrip- selachians. Of all the scale characteristics tions see Stromer, 1927; Peyer, 1946; Patter- Reif (1978b) notes, some may represent hyb- son, 1966; Maisey, 1975, 1977, 1978, 1979). odontid synapomorphies, since they are ap- Finspines assigned to Hybodus copei Hay parently restricted to these sharks, including (1899) have been recorded from various for- the high-pointed cusps, ridged conical mations in the Wichita Group (lower Perm- crown, and absence of a pronounced neck. ian; Hussakof, 1911; Romer, 1942; Berman, Other characters have a more widespread 1970). Other Permian remains from Texas 34 AMERICAN MUSEUM NOVITATES NO. 2724 and Oklahoma (Simpson, 1974) are also un- is important to realize that this restoration is mistakably hybodontid. Lund (1970) erected a composite from various hybodonts rather Hybodus allegheniensis on the basis of iso- than a reconstruction of a given species. As lated finspine fragments (including a highly im- far as can be established, however, the bet- probable "pectoral" spine), teeth, and scales ter-known hybodonts agree in most details from the Duquesne Limestone (Pennsylva- of their skeletal anatomy. The bias of this nian), all of which seem to be from restoration is as follows: hybodonts even if their conspecificity is un- a. Head, jaws, and hyoid arch mainly proven. Some (but not all) of the specimens from H. basanus, some details con- referred to this species in the Carnegie Mu- firmed by H. hauffianus, H. fraasi. seum, Pittsburgh, are probably from hybo- b. Gill arches mainly from H. cassangen- donts. Hybodus allegheniensis is presently sis, with some features added from, or the earliest record of a hybodont shark, al- confirmed by, H. basanus, H. hauffi- though another form, apparently, pertaining anus, H. fraasi. to a different species, was found in Kansas c. Axial skeleton and caudal fin mainly and will be described elsewhere (Zidek, per- from H. hauffianus. sonal commun.). d. Dorsal fins from H. hauffianus, H. fraasi, H. basanus, H. cassangensis, REVISED RESTORATION Lissodus africanus. OF HYBODUS e. Pectoral fins from H. cassangensis; some features confirmed by H. fraasi, For many years the only available resto- H. hauffianus. rations of hybodonts have been those of f. Pelvic fins from H. cassangensis; some Hybodusfraasi (Brown, 1900, figs. 1, 2), H. features confirmed by H. hauffianus. hauffianus (Brown, 1900, figs. 4, 5B; Jaekel, g. Anal fin from H. hauffianus. 1906, fig. 2; Woodward, 1916, fig. 2), H. cassangensis (Teixeira, 1954, 1978), and Lisso- THE QUESTION OF A RELATIONSHIP dus africanus (Broom, 1909, pl. XII, figs. 1, BETWEEN HYBODUS AND 2; Brough, 1935, fig. 1, pl. III, fig. 2). These HETERODONTUS restorations suffer from varying degrees of inaccuracy and fantasy. Woodward's (1916) The idea of a close relationship between version is perhaps the most accurate, but has Heterodontus and Hybodus is so well en- a full complement of radials in the first dorsal trenched in the literature to merit special at- fin, dorsal and ventral intercalaries in the tention. Recent research into the dermal vertebral column, and irregularly jointed skeleton of sharks has shown that in scale pectoral radials. Brown's (1900, fig. 1) and and finspine morphology and in tooth enam- Jaekel's (1906, fig. 2) skeletal restorations eloid ultrastructure Heterodontus is much are schematic. Both indicate a full comple- closer to other Recent sharks than to Hyb- ment of anterior dorsal radials. According to odus (Maisey, 1978, 1979; Reif, 1973, 1978b). Brown (1900) H. fraasi has intercalaries, al- As in Hybodus, the neurocranium of Het- though in the specimens on which this inter- erodontus (Daniel, 1915, 1934; Holmgren, pretation is based the axial skeleton is diffi- 1941; Smith, 1942; Nobiling, 1977) is some- cult to interpret. Brown (1900) also figured what wedge-shaped in lateral view, with a unjointed radials in the paired fins. Jaekel short otico-occipital region. Both supraorbi- (1906) showed jointed pectoral and pelvic ra- tal and suborbital shelves are well devel- dials and no intercalaries. Brough (1935, fig. oped. In both genera, the suborbital shelf is 1) and Teixeira (1978, fig. 2) correctly showed extended anteriorly to form the floor of a only one calcified radial in the anterior dorsal strong preorbital articulation with the pala- fin. toquadrate. In Heterodontus, however, the A revised restoration of a morphotypic post-nasal wall is not expanded laterally into Hybodus skeleton is given here (fig. 17). It an ectethmoid process, and the ethmoidal Cl

-o

C) 0

VL)

0s 0

N Ct

z v~ 1.

..z 36 AMERICAN MUSEUM NOVITATES NO. 2724

canal is absent. Nor does the neurocranial of rostrum and paired rostral rods, position floor of Heterodontus extend anteriorly into of the "orbital" process of the mesial surface a strong median intemasal keel, as it does in of the palatoquadrate (not, however, corre- Hybodus. Instead the internasal plate of sponding to the orbital process of certain Heterodontus is gently rounded and tapers Recent sharks; Maisey, 1980), and extent an- anteriorly into a rudimentary rostrum (ac- teriorly of the jaws below the ethmoidal re- cording to Holmgren, 1941, p. 44, this is seen gion. If Hybodus and Heterodontus are more clearly in embryos than in adults). Het- closely allied, we might expect to find at least erodontus has relatively large, well-devel- some of these characteristics in Hybodus. oped olfactory capsules, and a much narrow- However, the systematic value of these char- er internasal septum than Hybodus. As acters would depend on their being identified Holmgren (1941, p. 47) has noted, the eth- as synapomorphies. A single carotid foramen moidal region of Heterodontus (and chilo- is not restricted to Heterodontus and Hyb- scyllids) is elongate and downturned. This is odus, but is widespread among Recent and not the case in H. basanus (fig. 2C). fossil sharks. A separate hyomandibular VII While Holmgren (1941) concluded that foramen also occurs in many Recent and Heterodontus and chiloscyllids were allied some fossil sharks; it is doubtful whether to galeoid sharks (a still popular view, e.g., Hybodus had a separate foramen. The ros- Compagno, 1973, 1977), synapomorphies trum and paired lateral rostral bars are ab- uniting these taxa are not convincing. sent in several Recent sharks. The condition Among the many cranial similarities noted in almost all fossil sharks is unknown, al- by Holmgren (1941, p. 47, but excluding though it seems unlikely that such rostral ele- those he found common to all sharks), the ments were present in many cases. The jaws absence of a precerebral fossa (but not the of many fossil sharks extend below the eth- fontanelle) and the position of the articular moid region, and we find the same condition fossa for the hyomandibula in the anterior in some Recent forms, e.g., Chlamydosela- part of the otic region may be synapomor- chus and Squatina. Heterodontus and Hyb- phies of Heterodontus, chiloscyllids and odus are similar in that the "palatobasal" galeoids (sensu Holmgren, 1941). The re- articulation between the palatoquadrate and maining similarities between Heterodontus intemasal septum is well developed. This and galeoids listed by Holmgren (1941, p. 47) similarity may be of functional rather than also occur in other sharks and cannot there- phylogenetic significance, however, as there fore be regarded as synapomorphies of the are important differences in the ethmoidal forms under discussion. By placing Hetero- articulations of these genera (see below). dontus within his galeomorph group, Com- While the immediate relationship of Heter- pagno (1973, 1977) creates problems in defin- odontus to other Recent sharks is open to ing galeomorphs since Heterodontus and question, it shares several apparently de- orectolobids lack "typical" galeoid (i.e., car- rived characters (none of which occur in H. charhinoid and/or lamnoid) characters. basanus) with them, including: Nevertheless, the agreements between Het- erodontus and orectoloboid cranial anatomy 1. An outward swing of the aortic cephalic (Holmgren, 1941; Compagno, 1973) suggest circuit, resulting in a more direct course close relationship between these forms, even of the efferent hyoidean artery to the if their galeomorph affinity is questionable. internal carotid. According to Holmgren (1941, p. 47) 2. No caudal internasal keel in the eth- several cranial characteristics distinguish moid region. Heterodontus from galeoids, including the 3. Postorbital processes reduced ventral- well-developed "palatobasal articulation," ly. presence of a single carotid foramen, sepa- 4. Auditory capsules located lateral to the rate hyomandibular VII nerve foramen, lack occiput. 1982 MAISEY: HYBODONT SHARKS 37

5. Occipital demi-centrum incorporated Brough (1935) suggested that the presence, into the occiput (absent in Recent and per se, of finspines in both hybodonts and most fossil batoids). Heterodontus is meaningless, as these struc- 6. Hypophyseal duct closed in adults. tures are primitively present in various Pa- leozoic sharks. There are general similarities In some respects Heterodontus seems to be in the dental array of Heterodontus and cer- more derived than other Recent sharks. It tain hybodonts, particularly Asteracanthus lacks a separate epiphyseal opening in the and Acrodus. In all these forms the greater adult chondrocranium, behind the precere- part of the dentition forms a crushing tooth bral fossa, but an epiphyseal foramen occurs plate suitable for a durophagous habitus and (probably primitively) in squaloids, hexan- there are resemblances in their tooth histol- choids, and scyliorhinids. An epiphyseal ogy which reflect this functional similarity opening is absent in triakids, carcharhinoids (Agassiz, 1837; Peyer, 1946). There are also (with a few exceptions), lamnoids, and orec- important differences, however, including toloboids, and may represent a synapomor- the enameloid ultrastructure (Reif, 1973), phy between all these forms; Hybodus bas- and the arrangement of nutritive foramina in anus retains a separate epiphyseal opening. the anterior teeth (absent in Hybodus, Ac- Heterodontus is also derived in lacking rodus, and Asteracanthus). dorsal lateral aortae, so that its cranial blood Supposed similarities in the jaws of Hyb- supply is essentially hyoidean. In the course odus and hexanchoids are explained by (a) of dissecting various elasmobranchs, I have misidentification of the fossils Palaeospinax discovered this to be a fairly widespread oc- and Synechodus as hybodonts rather than as currence. Squalus is one of the few squa- primitive neoselachians (which caused great loids to retain lateral aortae, and these ves- confusion when Smith (1942) worked on sels also occur in Chlamydoselachus, Heterodontus), and (b) misinterpretation of hexanchoids, carcharhinoids, and lamnoids, the jaws and postorbital region of Hybodus but most other Recent elasmobranchs lack by Brown (1900) and Koken (1907). It is them. therefore understandable that Smith (1942) From the foregoing notes it is apparent was unable to reconcile the supposed simi- that H. basanus differs from Recent elas- larity between Hybodus and hexanchoid mobranchs in several respects. There is no jaws with the hypothesis that Hybodus is evidence here that Heterodontus is closer to more closely allied to Heterodontus. As it Hybodus than to other living sharks. Addi- turns out, Hybodus and Heterodontus jaws tionally none of the hybodont synapomor- are alike in having a strong ethmoidal artic- phies listed below occur in Heterodontus. A ulation and in lacking a postorbital articula- sister relationship between Hybodus and tion with the palatoquadrate (although these Heterodontus is therefore unparsimonious, sharks are not unique in either respect), but since it assumes that all unique hybodont the jaws of Heterodontus and Hybodus ar- characters have become suppressed in Het- ticulate with the neurocranium differently. In erodontus and that those characters shared both genera there is a strong ethmoidal artic- by Heterodontus and remaining living sharks ulation, with the dorsal margin of the pala- have either been lost or were never acquired toquadrate resting in a deep ethmoidal by Hybodus. groove. In Heterodontus there is no ethmo- The evidence that has been used in the palatine process, however, and no "rostral" past to promote a relationship between Het- articulation. Acrodus and Palaeobates are erodontus and hybodonts is thus equivocal. close to Hybodus in these regards, and their That hypothesis was largely based on such palatoquadrates are very similar in shape features as the presence of finspines, the de- (Kuhn, 1945; Rieppel, 1981). Asteracanthus velopment of low-crowned molariform teeth palatoquadrates resemble those of Hybodus and supposed similarities in jaw suspension. and Acrodus in having a deep adductor fossa 38 AMERICAN MUSEUM NOVITATES NO. 2724 but in addition have a well-developed lateral Palaeobates keuperinus, from the Keuper of process anteriorly which forms a basis for England, is founded on isolated teeth (Ac- the upper tooth plate. Such a structure is ab- rodus keuperinus Murchison and Strickland, sent in Heterodontus palatoquadrates. 1840), which were subsequently referred to Most Recent sharks and rays (including Palaeobates by Seilacher (1943). These Heterodontus), Hybodus basanus, and teeth lack specialized nutritive foramina probably other Hybodus, Acrodus, and As- (present in P. angustissimus). Some teeth of teracanthus species lack a postorbital artic- P. keuperinus were associated with fin- ulation between palatoquadrate and spines. These spines resemble those of Hyb- neurocranium. Unlike the palatoquadrates of odus (and P. angustissimus) in having Heterodontus, which articulate only with smooth ribbing and downcurved posterior the ethmoidal region, those of H. basanus denticles. One of these finspines (BM[NH] (and probably many hybodonts) remain close 46957), is associated with a large hybodont to the basicranium for much of their length, cephalic spine (Woodward, 1889a). Thus P. and the areas for jaw articulation with and keuperinus is probably a hybodont, but dif- attachment to the neurocranium are corre- ferences in finspine and tooth morphology spondingly much greater. Finally, the hy- from P. angustissimus suggest that it should omandibula plays an important role in jaw be removed from that genus (Rieppel, 1981). suspension of both Heterodontus and Hy- An almost complete fossil shark from British bodus (as in all elasmobranchs), but its re- Columbia was referred to Palaeobates by lationships to the back of the jaws and its Schaeffer and Mangus (1976) on the basis of shape are different in these genera. Thus dermal denticles, but its cranial anatomy and similarities in the jaws and teeth of Hetero- most important postcranial characters are dontus and Hybodus provide only equivocal unknown. evidence for an immediate relationship, and Teeth of Lonchidion from the lower Cre- are far outweighed by the numerous differ- taceous suggest hybodont affinity (Estes, ences in their cranial anatomy. It is more 1964; Patterson, 1966; Herman, 1977; Case, parsimonious to conclude that the few simi- 1979). They closely resemble teeth of the larities between the jaws of Heterodontus Triassic hybodont Lissodus (Estes, 1964; and hybodonts have arisen independently Patterson, 1966) and may be congeneric and are of functional rather than systematic (Duffin, 1981). Teeth of Lissodus are rarely significance. well preserved, however, and histological comparison with Lonchidion is difficult. The Permo-Triassic genus Polyacrodus is NOTES ON SOME PROBLEMATICAL still known only by teeth, but their associa- FOSSIL TAXA tion with hybodont finspines and cephalic Some fossil sharks represented by frag- spines (Stensio, 1921; Johnson, 1981) strong- mentary material suggest affinity with better ly suggests that Polyacrodus is a hybodont known Mesozoic hybodonts. Palaeobates (see earlier). was originally defined by teeth, which were Wodnika was originally considered to be distinguished on purely stratigraphical a hybodont (Miinster, 1843; Weigelt, 1930) grounds as Triassic species of Strophodus but recent discoveries of almost complete Agassiz (1837) (von Meyer, 1849). Histolog- specimens from the Permian Kupferschiefer ical studies (Jaekel, 1889; Stensio, 1921) sup- of Germany suggest that this genus is not ported a distinction between the teeth of closely related to Hybodus or to other hyb- these taxa and Jaekel attempted to separate odonts (Schaumberg, 1977). The Triassic ge- Palaeobates from other hybodonts. Stensio nus Carinacanthus was considered to be a (1921) thought they were closely related, hybodont by Bryant (1934). The holotype is however, and Rieppel (1981) has corrobo- a badly preserved postcranial skeleton. Cal- rated this view by describing associated re- cified ribs are not preserved but are suggest- mains of P. angustissimus, the type species. ed by faint impressions in the matrix. The 1982 MAISEY: HYBODONT SHARKS 39 postcranial skeletons of Wodnika and Cari- closely allied to it, and expressed doubt over nacanthus are otherwise similar, and and re- previous reports of calcified vertebral centra semble (perhaps primitively) those of Paleo- (e.g., Woodward, 1889a; Canavari, 1916). zoic sharks such as Goodrichthys and Recently, however, a well-preserved speci- "Ctenacanthus" costellatus. men of P. mortoni from the Kansas Chalk The Scottish lower Carboniferous (Visean) demonstrates that calcified centra are pres- sharks Tristychius and Onychoselache both ent (Stewart, 1980). Since these structures have calcified ribs which seem to have oc- are known among elasmobranchs only in Re- cupied a pleural position (see earlier discus- cent sharks and rays, their immediate fossil sion). The teeth of these taxa are distinguish- relatives and genera such as Palaeospinax able from each other, but both types of teeth and Synechodus (Compagno, 1973, 1977; resemble those of Mesozoic hybodonts more Maisey, 1975, 1977; Schaeffer and Williams, than those of other Paleozoic phalacanthous 1977), it is more parsimonious to regard Pty- sharks such as Goodrichthys, Ctenacanthus chodus as a close relative of these forms. At compressus, and "C." costellatus, which present, however, it is by no means clear have cladodont teeth. that Ptychodus is closer to some members "Ctenacanthus" vetustus is a late Devo- of this group than to others, as the following nian shark with finspines that resemble those review of the data will show. of Hybodus, except that the ribbing is broken up into pectinations anteriorly, and there A. Ptychoduslhybodont relationship are no posterior denticles. The spines have a convex posterior wall with a broad median Pro: gross morphology and histology of ridge, and the central cavity is keyhole- teeth (Agassiz, 1839; Owen, 1840; Casier, shaped and reduced by extensive deposits of 1953); absence of enlarged nutritive foramina orthodentine. Associated teeth are Orodus- in tooth bases (Patterson, 1966); similar like and may possess tubular dentine reach- enameloid ultrastructure in Ptychodus and ing the tooth surface without a continuous Asteracanthus (Reif, 1973). enameloid layer (a similar absence of enam- Con: presence of vertebral centra (Wood- eloid is noted in Wodnika teeth by Reif, per- ward, 1889a; Canavari, 1916; Stewart, 1980); sonal commun.). absence of hybodont finspines and cephalic The upper Cretaceous elasmobranch Pty- spines; presence of an upper symphyseal chodus is known mainly from distinctive but tooth row in Ptychodus (absent in Hybodus, isolated teeth, but tolerably complete denti- Acrodus, Asteracanthus); non-divergent tions have also been described (e.g., Wood- tooth files in Ptychodus (Woodward, 1887). ward, 1887; Williston, 1900; Canavari, 1916). On the basis of tooth morphology, Ptycho- B. Ptychoduslbatoid relationship dus was considered to be a "cestraciont" Pro: cyclospondylous vertebral centra (with Hybodus and Heterodontus) by Agas- (Woodward, 1889a; Canavari, 1916); denti- siz (1839), Owen (1840), and Casier (1953). tion of tooth plates, with mandibular rami in Woodward's (1887) discovery that small a straight line; straight tooth replacement "prehensile" anterior teeth were absent in files and a slightly wavy contour of the den- Ptychodus led him to remove that genus tition (Woodward, 1887). from the "cestracionts" (although Hybodus, Con: "unmodified" pectoral fin, no syn- which was by then also known to lack such arcual (Stewart, 1980). teeth, was still considered to be a "cestraciont"), and to suggest that Ptychodus was C. a batoid, subsequently (Woodward, 1889a) PtychoduslHeterodontus relationship placing the genus within the Myliobatidae. Pro: gross morphology and histology of Patterson (1966) reiterated the case for Pty- teeth (Agassiz, 1839; Owen, 1840; Casier, chodus being a hybodont, suggested that 1953). Hylaeobatis (known only from teeth) was Con: absence of finspines; different tooth 40 AMERICAN MUSEUM NOVITATES NO. 2724 enameloid ultrastructure (Reif, 1977); lack of G. Palatoquadrate morphology, with a "prehensile" anterior teeth and non-diver- large quadrate flange, deep adductor gent tooth replacement files (Woodward, fossa, and strong ethmopalatine artic- 1878); vertebrae of Ptychodus are cyclo- ulation. spondylous whereas those of Heterodontus H. Hyomandibula passes dorsal to caudal are asterospondylous (Woodward, 1889a). part of palatoquadrate. Woodward (1916) suggested that Hylaeo- I. Scales with several neck canals and batis, a genus founded on isolated teeth, was lacking a pronounced neck. closely allied to Ptychodus and that both J. Various aspects of finspine morpholo- genera were batoids. Patterson (1966) reached gy (see Maisey, 1978). a similar conclusion based on a comparative K. Teeth lack specialized nutritive foram- study of teeth, but considered that both gen- ina. era were hybodonts. Stewart (1980) consid- L. Cephalic spines with "Sphenonchus" ered Ptychodus to be a neoselachian and not morphology present. a hybodont, but retained Hylaeobatis as a Palaeobates angustissimus resembles hybodont, "since no evidence exists that it Hybodus, Acrodus, and Asteracanthus in is not." As Woodward (1916) and Patterson characters G, I, J, and L but not K. "Pa- (1966) noted, however, the histology and laeobates" keuperinus shares characters J, structure of Hylaeobatis and Ptychodus K, and L. Lissodus resembles Hybodus in teeth are very similar. The affinities of Pty- characters F, J, K, L, although many other chodus are thus no clearer now than they important features are unknown. were a century ago. Clearly it is not possible to establish the interrelationships of these taxa with any de- HYBODONT INTERRELATIONSHIPS gree of confidence. While they might be se- Hybodus and Acrodus are customarily quenced after a cladistic fashion, based on recognized as distinct genera on the basis of characters A to L, and while this sequence differences in their tooth morphology. Forms might reflect their actual relationships, all referred to these taxa are united by the fol- that would really be created is a list of taxa lowing characters, which are presently un- which are known in progressively less detail; known in other sharks: it is conceivable that all these taxa shared many or all of the characteristics listed A. Large, downturned postorbital pro- above, and that only in Hybodus and Acro- cess. dus are the data tolerably complete. Hybo- B. Distinctively inflated and long jugular dus shares the following characteristics with canal. Recent sharks and rays: C. Massive ethmopalatine process ventral to ectethmoid process. I. Ectethmoid process present on the D. Otic capsules lie between postorbital postnasal wall. processes. II. Pelvic girdle forms a continuous pu- E. Lateral otic process positioned imme- boischiadic bar. diately behind and in part dorsal to III. Gap present between basihyal and ba- postorbital process. sibranchials, and hypobranchials di- F. Complex system of many large labial rected posteriorly toward the midline. cartilages. None of these characters is known in other It is possible that some of these characters taxa allied to Hybodus. The last character are shared by other taxa such as Asteracan- (III) is important since Nelson (1969) con- thus and Palaeobates, but this is presently cludes that it is diagnostic (synapomorphic) unknown. Asteracanthus shares several for elasmobranchs. The branchial skeleton is characters with Hybodus and Acrodus, in- organized similarly in xenacanths (e.g., Ko- cluding: ken, 1889; Fritsch, 1889, 1895; Jaekel, 1895, 1982 MAISEY: HYBODONT SHARKS 41

1906; Reis, 1897) but is not arranged in this from Hybodus in at least some skeletal char- way in chimaeras (Nelson, 1969), Cladose- acters. At present, therefore, "Sphenon- lache (Dean, 1909), or Cobelodus (Zangerl chus" cephalic spines offer the most reliable and Case, 1976). On this basis, xenacanths way of recognizing a hybodont shark, even form a sister group to other elasmobranchs from fragmentary remains. (hybodonts plus Recent sharks and rays), but Living elasmobranchs are united by nu- Cobelodus and Cladoselache fall outside merous characters which do not occur in this group, as do chimaeras. Similarities in Hybodus. None of the features identified as the neurocrania of ctenacanths (including hybodont synapomorphies occur in living Tamiobatis?) and xenacanths (Schaeffer, elasmobranchs. All hypotheses which have 1981) suggest that Ctenacanthus also had its united Hybodus with certain Recent elas- branchial arches arranged according to the mobranchs to form a sister group of remain- elasmobranch pattern, a prediction which ing Recent forms are therefore rejected as may become testable as further remains are unparsimonious. These rejected hypotheses described. include: CONCLUSIONS a. Batoids are a sister group to hybodonts and Recent sharks (e.g., Regan, 1906; Of Agassiz's (1837) original "Hybo- White, 1937; Romer, 1945; Berg, 1955; dontes," two genera (Hybodus and Sphe- Patterson, 1967). nonchus) are founded on dermal elements b. Hexanchoids are a sister group to hy- (teeth and cephalic spines) that have subse- bodonts and other Recent elasmo- quently been recognized as belonging to a branchs (e.g., Brown, 1900; Goodrich, unique group of sharks. Unfortunately, the 1909, 1930). characters used by Agassiz to distinguish c. Hybodonts are more closely allied to Hybodus are ambiguous if used as synapo- Heterodontus than to any other Recent morphies either for that genus or for hybo- elasmobranchs (e.g., Woodward, 1889a; donts generally. Reif's (1973) work on tooth Goodrich, 1909, 1930; Young, 1962; enameloid ultrastructure may lead to identi- Patterson, 1967). fication of unambiguous synapomorphies in d. Hybodonts and Heterodontus are a sis- hybodont teeth, while his work on scale mor- ter group to all other Recent elasmo- phology (Reif, 1978b) suggests some other branchs (a variation of c; e.g., Young, features which may be hybodont synapo- 1962). morphies. Cephalic spines of "Sphenon- chus" morphology have been found associ- The most parsimonious hypothesis of re- ated with various teeth and finspines which lationship is that hybodonts are a monophy- resemble those of Hybodus and allied genera letic sister group of all Recent elasmo- (even where teeth possess specialized nutri- branchs (Brough, 1935; Moy-Thomas, 1939a, tive foramina). These cephalic spines are 1939b), with which they share an ectethmoid therefore regarded as a synapomorphy for a process, a unique arrangement of basibran- more inclusive concept of hybodont sharks chials and hypobranchials, and a continuous than is defined by the absence of nutritive puboischiadic bar. Some Paleozoic sharks canals in the teeth. Using the cephalic spine with long, Hybodus-like ribs have an un- as a means of recognizing a hybodont is not fused pelvic girdle (e.g., Tristychius, Ony- totally inconsistent with the Agassizian con- choselache), which outgroup comparison cept of hybodonts. Although various cranial with other gnathostomes suggests is the and postcranial characters are unique to hy- primitive condition. If Hybodus, Tristychius, bodonts, it is unlikely that these features will and Onychoselache belong to a monophylet- ever become known in more than a handful ic group, from which Recent elasmobranchs of examples. There is already evidence that are excluded on the basis of rib morphology some hybodonts (e.g., Palaeobates) differ (Dick, 1978; Dick and Maisey, 1980), the 42 AMERICAN MUSEUM NOVITATES NO. 2724 fused puboischiadic bar of Hybodus and vanced" hybodonts (Schaeffer, Recent elasmobranchs was probably ac- 1967b; but see Schaeffer, 1981). quired independently. It is more parsimon- ii. Structure of neurocranium (like hex- ious to regard pleural ribs of sharks as anchoids, according to Brown, 1900; primitive (i.e., Tristychius and Onychose- similar in many respects to Recent lache are sister groups to Hybodus and sharks, according to Schaeffer, 1967b; Recent elasmobranchs) since only two apo- like galeomorphs, according to Com- morphic states need arise (fusion of pelvic pagno, 1973). girdle in the ancestors of hybodonts and Re- iii. Structure ofjaws (e.g., Brown, 1900; cent elasmobranchs; subsequent modifica- Regan, 1906; Koken, 1907; Goodrich, tion of ribs in Recent elasmobranchs). The 1909; Smith, 1942). alternative hypothesis (Dick, 1978; Dick and iv. Similarities between teeth of Hybo- Maisey, 1980) requires four apomorphic dus and Synechodus (Woodward, characters (acquisition of long pleural ribs by 1888a); external morphology of teeth hybodonts; of intermuscular ribs by neose- (Agassiz, 1837). lachians; of a puboischiadic bar in Hybodus; v. Presence of cyclospondylous and of a puboischiadic bar in neoselachians). Al- weakly asterospondylous vertebrae though Tristychius and Onychoselache may in Synechodus and Palaeospinax be primitively allied to Hybodus, therefore, (e.g., Woodward, 1889a; Regan, these genera probably do not form a mono- 1906). phyletic group unless Recent elasmobranchs are included. LITERATURE CITED Many characters which have been used to suggest a relationship between hybodonts Agassiz, L. 1833-1844. Recherches sur les poissons fos- and Recent elasmobranchs can now be re- siles. Neuchatel, 5 vols. 1420 pp., with jected on the following grounds: supplement. Aldinger, H. A. Plesiomorphic characters (occurring out- 1937. Permische Ganoidfische aus Ostgron- side hybodonts and Recent sharks, per- land. Medd. Gr0nland, vol. 102, pp. haps representing synapomorphies of 5-392. higher taxa): Andrews, S. M. i. Presence of finspines (Regan, 1906). 1977. The axial skeleton of the coelacanth, ii. Tribasal pectoral endoskeleton (Re- Latimeria. In Andrews, S. M., R. S. gan, 1906; Brough, 1935). Miles, and A. D. Walker (eds.), Prob- iii. Division and reduction of radials in lems in vertebrate evolution. London, Academic Press, pp. 271-288. paired fins (Schaeffer, 1967b). Andrews, S. M., and T. S. Westoll iv. Presence of pelvic basipterygium 1970. The postcranial skeleton of Eusthe- (Regan, 1906). nopteron foordi Whiteaves. Trans. Roy. v. Acquisition of haemal elements along Soc. Edinburgh, vol. 68, pp. 207-329. entire length of notochord (Schaeffer, Balfour, F. M. 1967b). 1878. A monograph on the development of vi. Lack of epichordal and hypochordal elasmobranch fishes. London, xi + 295 radials in tail (Schaeffer, 1967b). PP. B. Convergent characters: Bendix-Almgreen, S. E. i. Tooth (especially toward 1975. The paleovertebrate fauna of Green- morphology land. In Escher, A., and W. S. Watt a durophagous habitus; e.g., Peyer, (eds.), Geology of Greenland, pp. 537- 1946; Schaeffer, 1967b). 573. Copenhagen, Geol. Surv. Green- C. Spurious or ambiguous characters: land. i. Form of rostrum (e.g., Regan, 1906); Berg, L. S. lack of rostrum (e.g., Goodrich, 1955. Classification of fishes and fish-like ver- 1909); "enlarged" rostrum of "ad- tebrates, living and fossils. 2nd ed., cor- 1982 MAISEY: HYBODONT SHARKS 43

rected and enlarged (in Russian). Trav. Charlesworth, E. Inst. Zool., Acad. Sci., USSR, Lenin- 1839. Illustrated zoological notices-on the grad, pp. 1-286. remains of a species of Hybodus from Berman, D. S. Lyme Regis. Mag. Nat. Hist., n.s., vol. 1970. Vertebrate fossils from the Lueders 3, pp. 242-248. Formation, lower Permian of north-cen- 1845. On the fossil bodies regarded by M. tral Texas. Univ. Calif. Publ. Geol. Sci., Agassiz as the teeth of a fish, and upon vol. 86, pp. 1-39. which he has founded his supposed ge- Blot, J. nus Sphenonchus. Report Brit. Assoc. 1969. Holocephales et elasmobranches. Sys- Adv. Sci., 15th meeting, part 2, p. 56. tematique. In Piveteau, J. (ed.), Traite Compagno, L. J. V. de Paleontologie, vol. 4, pp. 702-776. 1973. Interrelationships of living elasmo- Masson et Cie, Paris. branchs. In Greenwood, P. H., R. S. Branson, C. C. Miles, and C. Patterson (eds.), Interre- 1933. Fish fauna of the middle Phosphoria lationships of fishes. Lin. Soc. London, Formation. Jour. Geol., vol. 41, no. 2, vol. 53, pp. 15-61. Academic Press, pp. 179-183. London. 1934. Permian sharks of Wyoming and East 1977. Phyletic relationships of living sharks Greenland. Science, vol. 79, no. 2054, and rays. Amer. Zool., vol. 17, pp. 303- p. 431. 322. Broom, R. Daniel, J. F. 1909. Fossil fishes of the upper Karroo Beds 1915. The anatomy of Heterodontus francis- of South Africa. Ann. S. African Mus., ci. II. The endoskeleton. Jour. Morph., vol. 7, pp. 251-269. vol. 26, no. 3, pp. 447-494. Brough, J. 1934. The elasmobranch fishes. Univ. Calif. 1935. On the structure and relationships of Press, Berkeley, 3rd ed., xi + 332 pp. the hybodont sharks. Mem. Manchester Day, E. C. H. Lit. Phil. Soc., vol. 79, no. 4, pp. 35- 1864. On Acrodus anningiae; with remarks 48. upon the affinities of the genera Acro- Brown, C. dus and Hybodus. Geol. Mag., no. 1, 1900. Ueber das Genus Hybodus und seine pp. 57-65. systematische Stellung. Palaeontograph- Dean, B. ica, vol. 46, pp. 149-174. 1907. Notes on acanthodian sharks. Amer. Bryant, W. L. Jour. Anat., vol. 7, pp. 209-222. 1934. New fishes from the Triassic of Penn- 1909. Studies on fossil fishes (sharks, chimae- sylvania. Proc. Amer. Phil. Soc., vol. roids and arthrodires). Mem. Amer. 73, no. 5, pp. 319-326. Mus. Nat. Hist., vol. 9, pp. 209-287. Canavari, M. Denison, R. H. 1916. Descrizione di un notevole esemplare di 1968. The evolutionary significance of the ear- Ptychodus Agassiz trovato nel calcare liest known lungfish, Uranolophus. In bianco della Creta superiore di Gallio 0rvig, T. (ed.), Nobel Symposium 4, nei Sette Comuni (Veneto). Palaeon- Current problems of lower vertebrate togr., Ital., vol. 22, pp. 35-102. phylogeny, pp. 247-257. Almquist and Cappetta, H., and G. R. Case Wiksell, Stockholm. 1975. Contribution a l'etude des selaciens du 1978. Handbook of paleoichthyology. Volume Groupe Monmouth (Campanien-Maes- 2: Placodermi. Gustav Fischer Verlag, trichtien) du New Jersey. Palaeonto- Stuttgart, New York, 128 pp. graphica, vol. 151, pp. 1-46. Devillers, C. Case, G. R. 1954. Les c6tes. In Grasse, P. P. (ed.), Traite 1979. Additional fish records from the Judith de zoologie. Paris, Masson et Cie, vol. River Formation (Campanian) of Mon- 12, pp. 673-697. tana. Geobios, no. 12, fasc. 2, pp. 223- Dick, J. R. F. 233. 1978. On the Carboniferous shark Tristychius Casier, E. arcuatus Agassiz from Scotland. Trans. 1953. Origine des Ptychodontes. Mem. Inst. Roy. Soc. Edinburgh, vol. 70, no. 4, pp. Sci. Nat. Belgique, vol. 49, pp. 1-51. 63-109. 44 AMERICAN MUSEUM NOVITATES NO. 2724

Dick, J. R. F., and J. G. Maisey Paleontology, II, p. 292. Israel Program 1980. The Carboniferous shark Onychose- for Scientific Translations, Jerusalem. lache traquairi. Palaeontology, vol. 23, Goppert, E. no. 2, pp. 262-374. 1895. Untersuchungen zur Morphologie der Doderlein, L. Fischrippen. Morph. Jahrb., vol. 23, pp. 1889. Das Skelet von Pleuracanthus. Zool. 145-217. Anz., vol. 12, pp. 123-127. Goodey, T. Duffin, C. 1910. A contribution to the skeletal anatomy 1980. A new euselachian shark from the upper of the frilled shark, Chlamydoselachus Triassic of Germany. Neues Jahrb. anguineus Garm. Proc. Zool. Soc. Lon- Geol., Paliaont., Heft 1, pp. 1-16. don, 1910, pp. 540-571. 1981. Comments on the selachian genus Dor- Goodrich, E. S. atodus Schmidt (1861) (Upper Triassic, 1909. Cyclostomes and fishes. In Lancaster, Germany). Neues Jahrb. Geol., Pa- E. R. (ed.), A treatise on zoology, vol. liont., Heft 5, pp. 289-302. 9, Vertebrata craniata, 518 pp. London, Egerton, P. M. G. A+C Black. 1845. Description of a Hybodus found by Mr. 1930. Studies on the structure and develop- Boscawen Ibbetson in the Isle of Wight. ment of vertebrates. Macmillan, Lon- Quart. Jour. Geol. Soc., vol. 1, pp. 197- don, xxx + 837 pp. 199. Haller, G. Emelianov, S. W. 1926. Ueber die Entwicklung, den Bau, und 1935. Die Morphologie der Fischrippen. Zool. die Mechanik des Kieferapparates des Jahrb. (Anat.), vol. 60, pp. 132-262. Dornhais (Acanthias vulgaris). Zeitschr. Estes, R. Mikrosk. Anat. Forsch., vol. 5, pp. 1964. Fossil vertebrates from the late Creta- 749-793. ceous Lance Formation, eastern Wyo- Hamdy, A. R. ming. Bull. Dept. Geol., Univ. Calif., 1973. The phylogenetic significance of the ex- vol. 49, pp. 1-187. istence of a sixth branchial arch in Gal- Fraas, E. eocerdo cuvier. Proc. Egyptian Acad. 1889. Kopfstacheln von Hybodus und Acro- Sci., vol. 26, pp. 83-85. dus, sogenannte Ceratodus heteromor- Harris, J. E. phus Ag. Jahreshefte Ver. Vaterl. Na- 1950. Diademodus hydei, a new fossil shark turk. Wiirttemberg, vol. 45, pp. 233- from the Cleveland Shale. Proc. Zool. 240. Soc. London, vol. 120, no. 4, pp. 683- 1896. Neue Selachier Reste aus dem oberen 697. Lias von Holzmaden in Wiirttemberg. Hay, 0. P. Jahreshefte Ver. Vaterl. Naturk. Wiirt- 1899. On some changes in the names, generic temberg, vol. 52, pp. 1-25. and specific, of certain fossil fishes. Fritsch, A. Amer. Natural., vol. 33, pp. 783-792. 1889. Fauna der Gaskohle und der Kalksteine Herman, J. der Permaformation Bohmens. Z. Stego- 1977. Les selaciens des terrains neocretaces cephali, Dipnoi, Selachii. Prague, vol. et paleocenes de Belgique et des con- II, 114 pp. trees limitrophes. Mem. Serv. Geol. 1895. Fauna der Gaskohle und der Kalksteine Belge, Bruxelles, Mem. no. 15, 450 pp. der Permaformation Bohmens. Prague, Holmgren, N. vol. 3, 132 pp. 1941. Studies on the head in fishes. Part 2. Garman, S. W. Comparative anatomy of the adult se- 1913. The Plagiostoma (sharks, skates and lachian skull with remarks on the dorsal rays). Mem. Mus. Comp. Zool., vol. 36, fins in sharks. Acta Zool., vol. 22, pp. 528 pp. 1-100. Gegenbaur, C. 1943. Studies on the head in fishes. Part 4. 1898. Vergleichende Anatomie der Wirbel- General morphology of the head in fish. tiere mit Beriicksichtigung der Wirbel- Acta Zool., vol. 24, pp. 1-188. losen. Leipzig, xiv + 978 pp. Hussakof, L. Glikman, L. S. 1911. The Permian fishes of North America. 1964. Subclass Elasmobranchii (sharks). In Carnegie Inst. Washington Publ. no. Obruchev, D. V. (ed.), Fundamentals of 146, pp. 153-178. 1982 MAISEY: HYBODONT SHARKS 45

Jaekel, 0. selachians. Neues Jahrb. Geol. Pa- 1889. Die Selachier aus dem oberen Mus- liiont., Heft 9, pp. 553-567. chelkalk Lothringens. Abhandl. Geol. 1977. The fossil selachian fishes Palaeospinax Specialk. Elsass-Lothringen, vol. 3, pp. Egerton, 1872 and Nemacanthus Agas- 273-332. siz, 1873. Zool. Jour. Linn. Soc., vol. 1895. Ueber die Organisation der Pleuracan- 60, pp. 259-273. thiden. Sitzberg. Gesellsch. Natur- 1978. Growth and form of finspines in hybo- forsch. Freunde, Berlin, 1895, pp. 69- dont sharks. Palaeontology, vol. 21, no. 85. 3, pp. 657-666. 1898. Ueber Hybodus Agassiz. Sitzberg. Ge- 1979. Finspine morphogenesis in squalid and sellsch. Naturforsch. Freunde, Berlin, heterodontid sharks. Zool. Jour. Linn. 1898, pp. 135-146. Soc., vol. 66, pp. 161-183. 1906. Neue Rekonstruktionen von Pleur- 1980. An evaluation of jaw suspension in acanthus sessilis und von Polyacrodus sharks. Amer. Mus. Novitates, no. (Hybodus) hauffianus. Sitzberg. Ge- 2706, pp. 1-17. sellsch. Naturforsch. Freunde, Berlin, Meyer, H. von 1906, pp. 155-159. 1849-1851. Fische, Crustaceen, Echinoder- Jarvik, E. men und andere Versteinerungen aus 1952. On the fish-like tail in the ichthyostegid dem Muschelkalk Oberschlesiens. Pa- stegocephalians with descriptions of a laeontographica, vol. 1, pp. 216-242 new stegocephalian and a new crossop- (1849), pp. 243-279 (1851). terygian from the upper Devonian of Miles, R. S. East Greenland. Meddel. Gr0nland, 1967. Observations on the ptyctodont fish, vol. 114, no. 12, pp. 1-90. Rhamphodopsis Watson. Zool. Jour. Johnson, G. Linn. Soc., vol. 47, pp. 99-120. 1981. Hybodontoidei (Chondrichthyes) from 1970. Remarks on the vertebral column and the Wichita-Albany Group (early Perm- caudal fin of acanthodian fishes. Le- ian) of Texas. Jour. Vert. Paleo., vol. thaia, vol. 3, pp. 343-362. 1, no. 1, pp. 1-41. Miles, R. S., and T. S. Westoll Jollie, M. 1968. The placoderm fish Coccosteus cuspi- 1962. Chordate morphology. Chapman and datus Miller ex-Agassiz from the middle Hall, London, xiv + 478 pp. Old Red Sandstone of Scotland. Part 1. Koken, E. Descriptive morphology. Trans. Roy. 1889. Ueber Pleuracanthus Ag. oder Xena- Soc. Edinburgh, vol. 67, pp. 373-476. canthus. Sitzberg. Gesellsch. Natur- Moss, S. A. forsch. Freunde, Berlin, 1889, pp. 77- 1962. The mechanisms of upper jaw protru- 94. sion in sharks. Amer. Zool., vol. 2, p. 1907. Ueber Hybodus. Geol. Palaeont. Ab- 542. handl., n.s., 1907, pp. 261-276. 1972. The feeding mechanism of sharks of the family Carcharhinidae. Jour. Zool. Lon- Kuhn, E. don, vol. 167, pp. 423-436. 1945. Ueber Acrodus-Funde aus dem Grenz- Moy-Thomas, J. A. bitumenhorizont der anisischen Stufe 1934. The structure and affinities of Tarrasius der Trias des Monte San Giorgio (Kt. problematicus Traquair. Proc. Zool. Tessin). Eclogae geol. Helvetiae, vol. Soc. London, 1934, pp. 367-376. 38, no. 2, pp. 662-673. 1936. The structure and affinities of the fossil Lund, R. elasmobranch fishes from the lower 1970. Fossil fishes from southwestern Penn- Carboniferous rocks of Glencartholme, sylvania. Ann. Carnegie Mus., vol. 41, Eskdale. Proc. Zool. Soc. London, pp. 231-261. 1936, pp. 761-788. Luther, A. 1939a. The early evolution and relationships of 1908. Untersuchungen ueber die vom N. tri- the elasmobranchs. Biol. Rev., vol. 14, geminus innervierte Muskulatur der Se- pp. 1-26. lachier (Haie und Rochen). Acta Soc. 1939b. Palaeozoic fishes. Chemical Publ. Co., Sci. Fenn., vol. 36, pp. 1-168. Inc., New York, 149 pp. Maisey, J. G. Moy-Thomas, J. A., and R. S. Miles 1975. The interrelationships of phalacanthous 1971. Palaeozoic fishes. 2nd ed., Philadelphia 46 AMERICAN MUSEUM NOVITATES NO. 2724

and Toronto, W. B. Saunders Co., 259 1967. Elasmobranchii (Chondrichthyes). In PP. Harland, W. B. et al. (eds.), The fossil Moy-Thomas, J. A., and T. S. Westoll record. Geol. Soc. London, p. 666. 1935. On the Permian coelacanth, Coelacan- Peyer, B. thus granulatus Ag. Geol. Mag., vol. 1946. Die schweizerischen Funde von Aster- 72, pp. 385-415. acanthus (Strophodus). Paliiont. Ab- Miunster, G. von handl., vol. 64, pp. 1-101. 1843. Nachtrag zu der Beschreibung einiger Regan, C. T. merkwiirdiger Fische aus den Kupfer- 1906. A classification of the selachian fishes. schiefern. Beitr. Petrafactenk., vol. 6, Proc. Zool. Soc. London, 1906, pp. pp. 47-52. 115-124. Murchison, R. I., and H. E. Strickland Reif, W. E. 1840. On the upper formations of the New 1973. Morphologie und Ultrastruktur des Hai- Red Sandstone system in Gloucester- "Schmelzes." Zool. Scripta, vol. 2, pp. shire, Worcestershire and Warwick- 231-250. shire. Trans. Geol. Soc. London, (2) 1977. Tooth enameloid as a taxonomic crite- V(2), pp. 331-348. rion. 1. A new euselachian shark from Nelson, G. the Rhaetic-Liassic boundary. Neues 1969. Gill arches and the phylogeny of fishes, Jahrb. Geol. Paliont., Heft 9, pp. 565- with notes on the classification of ver- 576. tebrates. Bull. Amer. Mus. Nat. Hist., 1978a. Tooth enameloid as a taxonomic crite- vol. 141, art. 4, pp. 477-552. rion. 2. Is "Dalatias" barnstonensis Nielsen, E. Sykes, 1971 (Triassic, England) a squal- 1932. Permo-Carboniferous fishes from East omorphic shark? Neues Jahrb. Geol. Greenland. Meddel. Gr0nland, vol. 86, Paliont., Heft 1, pp. 42-58. pp. 1-63. 1978b. Types of morphogenesis of the dermal Nobiling, G. skeleton in fossil sharks. Paliiont. 1977. Die Biomechanik des Kieferapparates Zeitschr., vol. 52, pp. 110-128. beim Stierkopfhai (Heterodontus por- Reis, 0. M. tusjacksoni + Heterodontus philippi). 1897. Das Skelett der Pleuracanthiden und Adv. Anat., Embryol. Cell Biol., vol. ihre systematischen Beziehungen. Ab- 52, pp. 3-52. handl. Senckenberg. Naturforsch. Ge- 0rvig, T. sellsch., vol. 20, pp. 57-155. 1960. New finds of acanthodians, arthrodires, Rieppel, 0. crossopterygians, ganoids, and dip- 1981. The hybodontiform sharks from the noans in the upper middle Devonian cal- Middle Triassic of Mte. San Giorgio, careous Flags (Oberer Plattenkalk) of Switzerland. Neues Jahrb. Geol. Pa- the Bergisch Gladbach Paffrath trough. laont. Abh. vol. 161, no. 3, pp. 324-353. Part 1. Paliaont. Zeitschr., vol. 34, pp. 295-335. Romer, A. S. 1942. Notes on certain American Palaeozoic Owen, R. vol. 1840-1845. Odontography; or a treatise on the fishes. Amer. Jour. Sci., 240, pp. 216-228. comparative anatomy of teeth, etc. (2 viii vols.). London, lxxiv + 655 pp. 1945. Vertebrate paleontology. 2nd ed., + Press. 1869. Description of a great part of a jaw with 687 pp. Univ. Chicago the teeth of Strophodus medius, Ow., Rosen, D. E., P. L. Forey, B. G. Gardiner, and from the Oolite of Caen, in Normandy. C. Patterson Geol. Mag., vol. 6, pp. 193-196, 235- 1981. Lungfishes, tetrapods, paleontology, 236. and plesiomorphy. Bull. Amer. Mus. Patterson, C. Nat. Hist., vol. 167, art. 4, pp. 159-276. 1965. The phylogeny of the chimaeroids. Phil. Schaeffer, B. Trans. Roy. Soc. (B), vol. 249, pp. 101- 1952. The Triassic coelacanth fish Diplurus, 219. with observations on the evolution ofthe 1966. The British Wealden sharks. Bull. Brit. Coelacanthini. Bull. Amer. Mus. Nat. Mus. (Nat. Hist.), Geol., vol. 11, no. 7, Hist., vol. 99, art. 2, pp. 25-78. pp. 281-350. 1967a. Late Triassic fishes from the western 1982 MAISEY: HYBODONT SHARKS 47

United States. Bull. Amer. Mus. Nat. 1929-1931. Meddl. Gr0nland, vol. 86, Hist., vol. 135, art. 6, pp. 283-342. pp. 1-212. 1967b. Comments on elasmobranch evolution. 1961. Permian vertebrates. In Raasch, G. 0. In Gilbert, P. W. et al. (eds.), Sharks, (ed.), Geology of the Arctic, pp. 231- skates and rays, pp. 3-35. The Johns 247. Univ. Toronto Press. Hopkins Press, Baltimore. 1969. Elasmobranchiomorphi; Placodermata, 1981. The xenacanth shark neurocranium, Arthrodires. In Piveteau, J. (ed.), Traite with comments on elasmobranch mon- de Paleontologie, vol. 4, no. 2, pp. 71- ophyly. Bull. Amer. Mus. Nat. Hist., 692. vol. 169, art. 1, pp. 1-66. Stewart, J. D. Schaeffer, B., and M. Mangus 1980. Reevaluation of the phylogenetic posi- 1976. An early Triassic fish assemblage from tion of the Ptychodontidae. Trans. Kan- British Columbia. Bull. Amer. Mus. sas Acad. Sci., vol. 83, no. 3, p. 154. Nat. Hist., vol. 156, art. 5, pp. 517-563. Stromer, E. Schaeffer, B., and M. Williams 1927. Ergebnisse der Forschungsreisen Prof. 1977. Relationships of fossil and living elas- E. Stromer's in den Wusten Agyptens. mobranchs. Amer. Zool., vol. 17, pp. II. Wirbeltier-Reste der Baharije Stufe 293-302. (unterstes Cenoman): 9. Die Plagiosto- Schauinsland, H. men. Abhandl. Bayer. Akad. Wis- 1906. Die Entwicklung der Wirbelsiiule nebst sensch., vol. 31, no. 5, pp. 1-64. Rippen und Brustbein. Hertwig's Hand- Teixeira, C. buch Vergl. u. Exp. Entwickl., Band 3, 1954. Sur un hybodontide du Karroo de pp. 339-572. l'Angola. Garcia de Orta, vol. 2, no. 2, Schaumberg, G. pp. 205-207. 1977. Der Richelsdorfer Kupferschiefer und 1956. Sur un hybodontide du Karroo de seine Fossilien. III. Die tierischen Fos- l'Angola. Rev. Fac. Cien. Lisboa, Ser. silien des Kupferschiefers. 2. Vertebra- C, vol. 5, pp. 135-136. ten. Aufschluss, vol. 28, pp. 297-352. 1978. Les poissons fossiles du Karroo du Lota Seilacher, A. et de longo en Angola. Centenario da 1943. Elasmobranch Reste aus dem oberen Academia das Ciencas de Lisboa, no. Muschelkalk und Keuper Wiirttem- II, pp. 275-300. bergs. Neues Jahrb. Min., vol. 50, pp. 256-292. Vidal, L. M. Shute, C. C. D. 1915. Nota geol6gica y paleontol6gica sobre 1972. The composition of vertebrae and the del Jurasico superior de la provincia de occipital region of the skull. In Joysey, Lrida. Bol. Inst. Geol. Min. de Es- K. A., and T. S. Kemp (eds.), Studies pania, vol. 36, pp. 17-55. in Vertebrate Evolution, pp. 21-34. Watson, D. M. S. Edinburgh, Oliver and Boyd. 1937. The acanthodian fishes. Phil. Trans. Simpson, L. S. Roy. Soc. London, ser. B, vol. 228, pp. 1974. Acanthodes and Hybodus in the Perm- 49-146. ian of Texas and Oklahoma. Jour. Pa- Weigelt, J. leont., vol. 48, pp. 1291-1293. 1930. Wichtige Fischreste aus dem Mansfeld- Smith, B. G. er Kupferschiefer. Leopoldina, vol. 6 1942. The heterodontid sharks: their natural (Walter Festschrift), pp. 601-624. history, and the external development Westoll, T. S. of Heterodontus japonicus based on 1941. The Permian fishes Dorypterus and notes and drawings by Bashford Dean. Lekanichthys. Proc. Zool. Soc. Lon- Amer. Mus. Nat. Hist., Bashford Dean don, vol. 3, pp. 39-58. Memorial volume: Archaic fishes, art. White, E. G. 8, pp. 649-770. 1937. Interrelationships of the elasmobranchs Stensio, E. with a key to the Order Galea. Bull. 1921. Triassic fishes from Spitzbergen. Part I, Amer. Mus. Nat. Hist., vol. 75, pp. 1- xxvii + 307 pp., Vienna. 138. 1932. Triassic fishes from East Greenland col- Williston, S. W. lected by the Danish expeditions in 1900. Cretaceous fishes. Selachians and pyc- 48 AMERICAN MUSEUM NOVITATES NO. 2724

nodonts. Univ. Geol. Surv. Kansas, gr. Palaeont. Soc. London, 1916, pp. 1- vol. 6, pp. 237-256. 48. Woodward, A. S. 1919. The fossil fishes of the English Wealden 1886a. On the relations of the mandibular and and Purbeck Formations. Part II. hyoid arches in a Cretaceous shark Monogr. Palaeont. Soc. London, 1919, (Hybodus dubrisiensis Mackie). Proc. pp. 105-148. Zool. Soc. London, 1886, pp. 218-224. 1934. Notes on some recently discovered Pa- 1886b. On the palaeontology of the selachian laeozoic fishes. Ann. Mag. Nat. Hist., genus Notidanus, Cuvier. Geol. Mag., vol. 13, pp. 526-528. n.s., vol. 3, pp. 205-217, 253-259. Young, J. Z. 1887. On the dentition and affinities of the se- 1962. The life of vertebrates, 2nd ed. Oxford lachian genus Ptychodus, Agassiz. Univ. Press, xvi + 820 pp. Quart. Jour. Geol. Soc. London, vol. Zangerl, R. 43, pp. 121-131. 1979. Interrelationship of early chondrichthy- 1 889a. Catalogue of the fossil fishes in the Brit- ans. Zool. Jour. Linn. Soc., suppl. 1, ish Museum (Natural History). Part I. vol. 53, pp. 1-14. London, British Mus. (Nat. Hist.), Zangerl, R., and G. R. Case xivii + 474 pp. 1976. Cobelodus aculeatus (Cope), an ana- 1889b. On head of Hybodus delabechei, asso- canthous shark from Pennsylvanian ciated with dorsal fin spines, from the black shales of North America. Palae- lower Lias of Lyme Regis, Dorsetshire. ontographica, vol. 154, pp. 107-157. Ann. Rept. Com. Yorks. Phil. Soc., Zidek, J. 1889, pp. 58-61. 1969. Upper Carboniferous hybodontoid re- 1891. The hybodont and cestraciont sharks of mains and problematic ichthyodorulites the Cretaceous period. Proc. Yorks. of Bohemia. Part 1: Hybodus vestitus. Geol. Poly. Soc., vol. 12, pp. 62-68. (asopis Naradnino Muzea, odd. Priro- 1898. Outlines of vertebrate palaeontology for dovedny, vol. 137, nos. 1-2, pp. 70-76. students of zoology. Cambridge Univ. Zittel, K. von Press, xvi + 470 pp. 1911. Grundzuege der Paliontologie, 2nd ed. 1916. The fossil fishes of the English Wealden II. Abt. Vertebrata, vii + 598 pp. R. and Purbeck Formations. Part I. Mono- Oldenburg Verlag, Miinchen, Berlin.