a new stethacanthid chondrichthyan from the lower carboniferous of ...

Journal of Vertebrate Paleontology 21(3):438–459, September 2001 q 2001 by the Society of Vertebrate Paleontology

A NEW STETHACANTHID CHONDRICHTHYAN FROM THE LOWER CARBONIFEROUS OF BEARSDEN, SCOTLAND M. I. COATES*1 and S. E. K. SEQUEIRA*2 Department of Biology, Darwin Building, University College London, Gower Street, London WC1E 6BT, United Kingdom

ABSTRACT—Exceptionally complete material of a new stethacanthid chondrichthyan, Akmonistion zangerli, gen. et sp. nov., formerly attributed to the ill-defined genera Cladodus and Stethacanthus, is described from the Manse Burn Formation (Serpukhovian, Lower Carboniferous) of Bearsden, Scotland. Distinctive features of A. zangerli include a neurocranium with broad supraorbital shelves; a short otico-occipital division with persistent fissure and Y-shaped basicranial canal; scalloped jaw margins for 6–7 tooth files along each ramus; a pectoral-level, osteodentinous dorsal spine with an outer layer of acellular bone extending onto a brush-complex of up to 160% of neurocranial length; a heterosquamous condition ranging from minute, button-shaped, flank scales to the extraordinarily long-crowned scales of the brush apex; and a sharply up-turned caudal axis associated with a broad hypochordal lobe. The functional implications of this anatomy are discussed briefly. The rudimentary mineralization of the axial skeleton and small size of the paired fins (relative to most neoselachian proportions) are contrasted with the massive, keel-like, spine and brush complex: Akmonistion zangerli was unsuited for sudden acceleration and sustained high-speed pursuit of prey. Cladistic analysis places Akmonistion and other stethacanthid genera in close relation to the symmoriids. These taxa are located within the basal radiation of the chondrichthyan crowngroup, but more detailed affinities are uncertain. They may represent a plesion series on the holocephalan stem lineage, or a discrete clade branching from the base of the elasmobranch lineage.

INTRODUCTION Stethacanthus is one of the most widely known of Paleozoic chondrichthyan genera, mostly because of its unusual spine and ‘brush’ complex (Lund, 1974, 1985a; Zangerl, 1981, 1984; Williams, 1985; Coates et al., 1998). Unfortunately, this prominence among early sharks (sensu lato) is not matched by detailed knowledge of its skeletal anatomy or a clearly defined taxonomic diagnosis. The several species of Stethacanthus erected in the late nineteenth century were based upon isolated spines (Newberry, 1889), and the first associated skeletal remains, from the Mississippian of Montana (Lund, 1974, 1985a) and the Devonian and Mississippian of Ohio (Zangerl, 1981; Williams, 1985), were undescribed until a century or so later. Lund (1974, 1985a) and Williams (1985) compiled the most detailed reviews of the genus, and it is noteworthy that Williams chose to retain many specimens within the type species, Stethacanthus altonensis (St. John and Worthen, 1875), pending the discovery of additional material to resolve questions of taxonomic diversity, sexual dimorphism, and sampling bias. Therefore, it is worth emphasizing that in the following systematic description, comparisons with all material described as Stethacanthus altonensis should be treated with caution. Further reports of material attributed to Stethacanthus have extended its range to the Mississippian of Oklahoma (Zidek, 1993) and the Lower Tournaisian of Central Russia (Lebedev, 1996). In 1982, Wood announced the discovery of an Upper Carboniferous (basal Namurian/Serpukhovian) fish fauna from Bearsden, Scotland. His report included a photograph of HM (Hunterian Museum, Glasgow University) V8246 (Fig. 1C), one of the most complete Paleozoic chondrichthyan specimens ever discovered and a principal subject of the present article. Significantly, this individual resembles Zangerl’s (1981:fig. 81) reconstruction of S. altonensis, and information from the pub* Current Address: 1, Department of Organismal Biology and Anatomy, the University of Chicago, 1027 East 57th Street, Chicago, Illinois 60637-1508; 2, Department of Biology, Birkbeck College, University of London, Malet Street, London WC1E 7HX, United Kingdom.

lished photograph (Wood, 1982:fig. 2) was used to refine a subsequent restoration (Zangerl, 1984:fig. 1). However, while Zangerl (1984) referred to HMV8246 as ‘‘cf. Stethacanthus,’’ Wood (1982) identified the specimen as either S. altonensis St. John and Worthen (1875) or Cladodus neilsoni Traquair (1898), and suggested that these species are synonymous. Neither of the reviews by Williams (1985) or Lund (1985a) refer to the Bearsden material. Since then, Coates and Sequeira (1998) completed a detailed comparative description of the Bearsden stethacanthid neurocranium, and, most recently, histological analysis of the spine and ‘brush’ complex revealed a remarkably well preserved, and so far unique, combination of skeletal tissues (Coates et al., 1998). The hypothesized synonymy of Cladodus neilsoni and the Bearsden stethacanthid has now been rejected, following redescription of the single specimen of C. neilsoni, NMS (National Museums of Scotland, Edinburgh) 1911.62.52 (Sequeira and Coates, 2000). Cladodus neilsoni has been removed from Cladodus (a nomen dubium, Chorn and Whetstone, 1978), distinguished from HMV8246 on the basis of differences in neurocranial, pectoral fin and branchial arch morphology, and placed within a new genus, Gutturensis Sequeira and Coates (2000). Key remaining questions about the Bearsden stethacanthid therefore concern its relationship to the species S. altonensis, the genus Stethacanthus, and the Family Stethacanthidae Lund (1974). The specific diagnosis for S. altonensis is rudimentary (Williams, 1985) because the lectotype is an isolated spine, designated by Lund (1974): FMNH (Field Museum of Natural History, Chicago) UC27404 (St. John and Worthen, 1875:pl. 19, fig. 1). The generic diagnosis (Williams, 1985) is similarly minimal, and refers to no more than spine shape, histology, the presence of a brush with specialized apical scales, and the association of these with a ‘‘medium sized cladodont shark’’ (assumed to mean any Paleozoic chondrichthyan with multicusped teeth with a lingual torus). The family level is defined a little more clearly following Zangerl’s (1990) revision, in which the Stethacanthidae includes symmoriids with ‘‘neural arch ele-

438

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN ments enhanced in the neck,’’ as well as presence of the distinctive spine and brush. In practice, the taxonomic level at which these spine morphologies are diagnostic is uncertain. Spines of this shape are present in specimens attributed to S. altonensis (Lund, 1974, 1985a; Williams, 1985) as well as the Bearsden species. But the Bearsden species differs from S. altonensis in terms of the structure and shape of the spinebrush complex, the arrangement of radials in the tail and paired fins, and, most significantly, the gross structure of the neurocranium. Consequently, the Bearsden species is easily diagnosed as a new taxon, but are these differences sufficient to support the erection of a new genus? To an extent, the choice between new species or new genus and species is arbitrary. Analyses of early chondrichthyans are likely to treat S. altonensis and the Bearsden species as distinct entities irrespective of decisions about taxonomic badging, but it is also likely that such analyses will generate trees in which these species emerge as sistergroups. Therefore, inclusion of the Bearsden species within Stethacanthus might be the simplest solution, but this would also, necessarily, increase the level of polymorphism within the genus. And the end result would achieve much the same as the addition of new species to illdefined genera such as Cladodus or Ctenacanthus. We regard this as inconsistent with the aim of the International Code of Zoological Nomenclature (Ride et al., 1999), which seeks to establish and promote taxonomic stability. An alternative, and preferred, course of action is to erect a new genus for the Bearsden species, and take advantage of the unusually detailed condition of the specimens to introduce a more precisely defined taxon to the existing list of early chondrichthyan genera. Possible synonymy with Stethacanthus can be tested elsewhere, following revision of material referred to S. altonensis and/or the discovery of new specimens, but this is beyond the scope of the current work. METHODS Bearsden material was prepared using an S. S. White industrial airbrasive unit, with sodium bicarbonate as the abrasive agent mixed minimally with a low-pressure airstream of less than 1.0 kg/cm2. Fine preparation was completed using mounted tungsten carbide needles. Fragile specimens were reinforced with a thin coating of methacrylate resin consolidant dissolved in acetone. Enhanced contrast photography of HM V8246, completed by Dr. Keith Ingham of the Hunterian Museum, University of Glasgow, was achieved by immersing the specimen in methanol. All specimen drawings were completed using Zeiss Stemi SV6 and Wilde M10 binocular microscopes with camera lucida attachment. ABBREVIATIONS Institutional AMNH, American Museum of Natural History, New York; CMNH, Cleveland Museum of Natural History, Cleveland, Ohio, USA; FMNH, Field Museum of Natural History, Chicago; U.S.A.; HM, Hunterian Museum, Glasgow University, Glasgow; MCZ, Museum of Comparative Zoology, Harvard University, Cambridge, Massachusetts, U.S.A.; NMMNH, New Mexico Museum of Natural History, Albuquerque; NMS, National Museums of Scotland, Edinburgh; UMZC, University Museum of Zoology, Cambridge University, Cambridge, England. Anatomical ac.cart, accessory cartilage; aep, anterior ethmoid process; a.h.rad, anterior hypochordal radial; apr, anterior process; art.c, articular crest; artp, articular facet for palatoquadrate; ax.crt, axial cartilage; ax.rad, axial radial; bbr, basibranchial; bp, basal plate; br, brush; cbr, ceratobranchial; chy, ceratohyal; cor, coracoid; cr, cartilaginous branchial ray; d.crt, delta cartilage; dlof, dorsolateral otic fossa; d.rad, distal

439

radial; dzf, diazonal foramen; ebr, epibranchial; ebrV, fifth epibranchial; endf, endolymphatic fossa; fcda, foramen/canal for dorsal aorta; fehy, foramen for efferent hyoidean artery; fhyp, hypophyseal/internal carotid foramen; fl, flange; flda, foramen for lateral dorsal aorta; fm, foramen magnum; foa, foramen for orbital branch of external carotid; fosn, foramen for occipitospinal nerve; fpal, foramen for palatine nerve or branch of orbital artery; grv, groove; gica, groove for internal carotid; gr.opt, optic groove; ha, haemal arch; h.rad, hypochordal radial; hsp, haemal spine; hy, hyomandibula; inp, internasal plate; jc, jugular canal; lof, lateral otic fossa; lop, lateral otic process; lor, lateral otic ridge; mc, Meckel’s cartilage; mdr, median dorsal ridge; mpt, metapterygium; mpt.c, metapterygial condyle; mxpt, mixopterygium; na, neural arch; nc, nasal capsule; nsp, neural spine; oaf, otic articular fossa; occ, occipital cotylus; ocr, occipital crest; oof, otico-occipital/metotic fissure; op, otic process; pbr, pharyngobranchial; pc, perichordal calcification; p.cor, procoracoid; pct.lv, pectoral level; pep, posterior ethmoid process; plv.lv, pelvic level; pof, preoccipital fossa; pop, postorbital process; p.pl, pelvic plate; ppr, posterior process; pq, palatoquadrate; pr, palatine ramus; p.rad, proximal radial; pro, preorbital process; psc, posterior semicircular canal; r, calcified rod; rad, radial; sc, scapula; sn.rad, supraneural radial; sp, spine; t.crt, terminal cartilage; trf.jc, foramen for trigeminofacialis nerve and jugular canal; trf/os, anterior part of trigeminofacialis opening/possible eyestalk insertion site; vmr, ventrolateral mandibular ridge; von, ventral otic notch. SYSTEMATIC PALEONTOLOGY Class CHONDRICHTHYES Huxley, 1880 Order SYMMORIDA Zangerl, 1981 (sensu Zangerl, 1990) Family STETHACANTHIDAE Lund, 1974 (sensu Zangerl, 1990) Genus AKMONISTION, gen. nov. Remarks For the purposes of the present work, the Stethacanthidae will be treated as monophyletic, despite indications from recent analyses that this taxon may be paraphyletic relative to the Holocephali (Coates and Sequeira, 2000). Type Species Akmonistion zangerli, sp. nov. Generic Diagnosis In the context of relevant hypotheses of early chondrichthyan interrelationships, Akmonistion may be diagnosed as follows (sources: Schaeffer, 1981; Zangerl, 1981, 1990; Young, 1982; Maisey, 1984; Lund, 1985a, b, 1986; Williams, 1985; Gaudin, 1991; Janvier, 1996; Coates and Sequeira, 1998, 2001). Symplesiomorphies at various levels: persistent otico-occipital fissure; endolymphatic fossa divided from fissure by posterior tectum; jaws amphistylic; palatoquadrate otic process with simple, convex, posterodorsal rim; hyoid arch articulates with lateral otic process; pectoral fin metapterygial plate articulates with 51 distal radials; paired and median fins plesodic; mandibular teeth with up to three cusplet pairs flanking primary cusp. Possible synapomorphies with symmoriids and stethacanthids: narrow suborbital shelf; hyomandibula short and subcrescentic; enamel and orthodentine discontinuous between neighbouring tooth cusps; oropharynx lined with single and compound buccopharyngeal denticles; scales single-crowned, non-growing, and restricted to particular body regions; single, anacanthous, radial-supported dorsal fin at pelvic level, with delta-shaped cartilage; pectoral fin with trailing ‘whip’; caudal axis steeply upturned; hypochordal lobe supported by elongate, distally splayed, radials. Synapomorphies with stethacanthids: basicranium encloses Y-shaped canal enclosing division of dorsal into lateral aortae;

440

JOURNAL OF VERTEBRATE PALEONTOLOGY, VOL. 21, NO. 3, 2001

whorls of fused, 3-cusp, tooth units occur close to level of jaw articulation; pectoral level subtriangular osteodentinous spine with distinct anterior shoulder seated on globular calcified cartilage baseplate; baseplate continuous with brush consisting of elongate calcified rods; lateral line scales C-shaped; large, cranial cap and brush apex scales include morphologies referrable to Lambdodus hamulus and Cladodus pattersoni. Autapomorphies of genus and species: neurocranium with broad supraorbital shelf; jaw margins scalloped with 6–7 recesses for tooth files along each ramus; teeth on upper and lower jaws aligned in precise, crown-to-crown, opposition with no dental interdigitation; level of jaw articulation ventral to tooth row level; vertebral column of 901 precaudal neural arches includes ca. 14 subrectangular cervical arches and ca. 9 small, trapezoidal, pre-caudal arches; spine histology includes outer layer of acellular bone; bone layer extends onto leading edges of baseplate and brush; male brush length reaches around 160% of neurocranium length; scales with longest crowns flank posterolateral third of brush apical platform; pectoral whip includes 221 axial cartilages; caudal neural and supraneural spines extended distally and leaf-shaped. Other characters of uncertain polarity: neurocranium with short otico-occipital division; hyoid and gill arches with elongate cartilaginous rays; male pelvic claspers with distal components consisting of non-prismatic calcified cartilage; fifth epibranchial antero-posteriorly broader than those of first to fourth gill arches. Etymology From the ancient Greek ‘akmon’ for anvil, and ‘istion’ for sail; these terms refer to the shape and location of the spine-brush complex. AKMONISTION ZANGERLI, sp. nov. (Figs. 1–4, 5A–C, F–J, 6–16) Holotype HMV8246 (Figs. 1B, C, 3A, 6, 7B, 8A, 9A, 10A, 11B, 12G, 14), a skeletally complete male specimen, approximately 620 mm long, preserved in lateral view; used as the source of body proportions in the reconstruction of Stethacanthus altonensis presented in Zangerl (1984). Referred Specimens NMS 1981.63.22C, NMS 1981.63.23A, and UMZC GN1047, the second most complete skeleton (Fig. 1A). Locality and Horizon Specimens of Akmonistion zangerli, sp. nov. were collected by S. P. Wood (Wood, 1982) from the Manse Burn Formation of the Bearsden locality, near Glasgow, Scotland, during the summers of 1981 and 1982. Assistance was provided by the Hunterian Museum, University of Glasgow, and the Nature Conservancy Council. The type locality is the Manse Burn, Ordnance Survey Grid reference NS 529427329-NS 53057325 (Clark, 1989). The Manse Burn Formation lies within the Pendleian (Serpukhovian) E1 Zone of the Lower Carboniferous, based on spore, conodont and goniatite analysis. The Manse Burn Formation consists of shales from the Top Hosie Limestone Marine Band to the base of the first thick sandstone (Clark, 1989); this complex is further divided into six members based on their lithology. There is an approximate correspondence between the Shrimp Member, Posidonia Member, Nodular Shale Member, Platey Shale Member, Betwixt Member, and Lingula Member, as listed by Clark, and those beds listed by Wood (1982). In addition to Akmonistion, the chondrichthyan component of the Bearsden fauna includes Denaea, Tristychius arcuatus (from the non-marine shale beds), Deltoptychius (Dick et al., 1986) and an isolated head attributed (doubtfully in our view) to Symmorium. Other vertebrates from this fauna include at least eight actinopterygian species, rhizodont and osteolepid fragments, a coelacanth skull, and Acanthodes specimens (Coates, 1988, 1993, 1998).

The lithology of the formation indicates that it was deposited under variable conditions of salinity with seasonal periodicity. All of the Bearsden chondrichthyans except for Tristychius derive from the finely laminated marine shales of the Shrimp Member, now known from several other localities in the western Midland Valley of Scotland (Clark, 1989). The Shrimp Member bears evidence (enterospirae 87Sr/86Sr ratios; trace element and rare earth analyses: Clark, 1989) of a sequentially marine and non-marine environment, subject to seasonal fluctuations. Diagnosis As for genus. Etymology The species is named ‘‘zangerli’’ after Rainer Zangerl, in recognition of his major contributions to palaeoichthyology. DESCRIPTION Neurocranium The neurocranium has been described in detail elsewhere (Coates and Sequeira, 1998); for a diagrammatic summary/reconstruction see Figure 2. Key features include: an otico-occipital region of about the same length as the sphenethmoid division; short but distinct lateral otic processes; a short, broad, endolymphatic fossa; a persistent otico-occipital fissure; Y-shaped basicranial canals indicating origin of the lateral aortae anterior to occipital level; very broad supraorbital shelves; and minimal suborbital shelves limited to small, ethmoidally located outgrowths. The dorsal aspect differs from those of Stethacanthus altonensis (Lund, 1974, 1985a) and cf. S. productus (Lund, 1985a) in the following details: the postorbital processes extend out further from the main body of the neurocranium; the supraorbital shelves are far more extensive; the endolymphatic fossa is laterally broad and well defined; the span of the olfactory capsules is narrower. Mandibular Arch The mandibular arch is preserved in 3 specimens: HM V8246, UCMZ GN1047 and NMS 1981. 63.23A (Fig. 3). In each case the left palatoquadrate and meckelian cartilage are flattened and exposed in lateral view. In HMV8246 and UCMZ GN1047 parts of the right side of the mandibular arch are exposed, but insufficiently well to add significantly to morphological description. The palatoquadrate resembles those of other stethacanthids (Lund, 1985a, b, 1986a; Williams, 1985), Cobelodus (Zangerl and Case, 1976), xenacanths, and living notidanids (Hotton, 1952), plus other amphistylic Paleozoic chondrichthyans. The expanded posterior of the palatoquadrate, the otic process (Fig. 3, op), is of about the same length as the suborbital palatine ramus (Fig. 3, pr). The ventral edge of the palatoquadrate is gently sigmoid in profile: concave below the otic process and convex below the palatine ramus. The posterodorsal edge of the otic process is strongly convex, somewhat thickened, and forms a distinct lip which appears to have been folded ventrally post mortem (NMS 1981.63.23A). There is no short, vertical process posterior to the quadrate region as in Stethacanthus altonensis (Lund, 1985a) and Orestiacanthus (Lund, 1984), or elevated facet or crest, cf. Tamiobatis vetustus (Williams, 1998). The otic articular fossa (Fig. 3, oaf) is restricted to the lateral surface, and resembles the V-shaped articular fossa of Stethacanthus altonensis (Williams, 1985:118), although there are significant differences. In the Bearsden species this fossa is anteroposteriorly narrower, and the anterior rim is continuous with the anterodorsal angle of the otic process. In S. altonensis the posterior edge of the fossa extends to the anterodorsal angle, and the leading fossa edge terminates in front of the main body of the otic process, creating a distinct step in the otic process outline, as shown in Williams’ (1985) and Lund’s (1985a) descriptions. Other Paleozoic chondrichthyan palatoquadrates showing a similar, laterally directed surface for articulation with the postorbital process include cladoselachians (Maisey, 1989a;

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN

441

FIGURE 1. Akmonistion zangerli, gen. et sp. nov. A, UCMZ GN1047, drawing of complete specimen; B, HMV8246, drawing of complete specimen; C, HMV8246, photographed in methanol (courtesy of Dr. Keith Ingham, Hunterian Museum, University of Glasgow). Scale bars equal 40 mm.

442


FIGURE 2. Akmonistion zangerli, reconstruction of neurocranium from Coates and Sequeira (1998), in A, ventral, B, dorsal, C, lateral, and D, posterior, aspects (reproduced with permission of the Royal Society of Edinburgh). Scale bar equals 10 mm.

pers. obs. CMNH 8207), Cobelodus (Zangerl and Case, 1976), and Symmorium and Denaea (Williams, 1985). The quadrate articular surface is not well preserved. The best example (from UCMZ GN1047) includes an outer, rounded, rim bordering a transverse slot in the quadrate base. This feature is not attributed easily to the presence of a condyle, and it may reflect the simpler (and perhaps specialized) mandibular hinge joint which Maisey (1989a) described in a cladoselachian (although Williams (pers. comm.) interprets this as a preservational artifact). The palatine ramus is proportioned similarly to that of Stethacanthus cf. S. altonensis (Lund, 1985a), and is much narrower, dorsoventrally, than those of ctenacanths (Williams, 1998) and xenacanths (Hotton, 1952). The anterior end is expanded dorsally (cf. Cobelodus, Zangerl and Case, 1976) to produce a flange (Fig. 3A, fl) which is assumed to have articulated with the ethmoid processes of the neurocranial internasal plate (Fig. 2, inp) (Coates and Sequeira, 1998). The labial (lateral) surface close to the ventral margin bears a series of at least six, near-hemispherical, humps. Corresponding concavities (‘‘scalloping,’’ Maisey, 1989a) on the lingual surface accommodated individual tooth files (families). Below each bulge the ventral palatoquadrate edge is reinforced, forming a short shelf, divided from neighbouring shelves by shallow, acute, grooves. These shelves probably functioned as tooth platforms; so far, no equivalent shelves have been identified in comparable species. Tooth file positions are arranged so that the fifth or sixth (from anterior) lies below the otic articular fossa. The posteriormost shelf is the shortest; approximately 1/3 to 1/2 the anteroposterior length of the largest. Each shelf is in precise 1:1 register with an equivalent shelf on the mandibular dorsal edge. This contrasts with conditions in taxa such as hybodontids (Maisey, 1983:fig. 18), where the off-set arrangements of teeth in upper and lower jaws result in dental interdigitation when the jaws are closed. As in the palatoquadrate (HM V8246), only the lateral surface of Meckel’s cartilage is exposed. This, too, bears a series of six or seven humps and platforms associated with tooth file positions (uncertainty over tooth file number results from uncertainty about presence of a shelf/platform at the anteriormost end of the mandible). Posterior to the tooth-bearing margin, which occupies no more than two thirds of the mandibular dorsal edge, the upper rim of Meckel’s cartilage slopes ventrally

FIGURE 3. Akmonistion zangerli, visceral skeleton. A, HMV8246, mandibular, hyoid, and branchial arches; B, UCMZ GN1047, mandibular, hyoid, and fragmented branchial arches; C, NMS 1981.63.23A, mandibular and hyoid arches. Scale bars equal 20 mm.

towards the point of jaw articulation. This contrasts with other Stethacanthus species (Lund, 1985a), cladoselachians (Maisey, 1989a), and Cobelodus (Zangerl and Case, 1976), in which the biting margin is either level with or slightly below the mandibular articulation. Other Meckel’s cartilages which possess a similar angle between the biting margin and the rim flanking the zone of adductor muscle insertion include those of Damocles (Lund, 1986) and Falcatus (Lund, 1985b, where this angle is described as a ‘‘coronoid process’’). A prominent rod-like ridge parallels the ventral mandibular edge in each of the three known examples of Meckel’s cartilage (Fig. 3, vmr). Comparison with three-dimensionally preserved material (cf. Maisey, 1989a:fig. 2; Orthacanthus casts and original material, pers. obs.) shows that this ridge forms the ventrolateral mandibular angle. Ridge thickness confers ridgidity to the entire jaw, and the lateral prominence of this structure probably provided a good insertion area for adductor musculature. The flange ventral to this ridge is assumed to have been directed medially in life. There is no retroarticular process (cf. Tristychius, Dick, 1978; Orthacanthus, Hotton, 1952) or flange

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN for the ceratohyal (cf. Falcatus, Lund, 1985b). Like the palatoquadrate, the articular region is insufficiently well preserved to identify articular surfaces. There are no traces of labial cartilages. Such cartilages are present in extant holocephalans and neoselachians, and, given the size and quality of preservation of HM V8246, it seems likely that evidence of such structures would have been preserved (cf. Tristychius, Dick, 1978; Hybodus, Maisey, 1982, 1983). Hyoid Arch and Gill Arches All three specimens include the left hyoid arch, but only UCMZ GN1047 preserves those of both sides (Fig. 3). Each hyomandibula (epihyal of Zangerl and Case, 1976; Maisey, 1989a) is flattened laterally, moderately arched and expanded anterodorsally (Fig. 3, hy). The degree of curvature parallels that of the posterodorsal edge of the palatoquadrate otic process; the overall shape is a half-crescent, and unlike the more linear epibranchials. When articulated with the neurocranium, it appears that little of the hyomandibula would have projected behind the palatoquadrate otic process (cf. HM V8246). A shallow groove runs parallel to the posterodorsal edge of the hyomandibula; a similar groove may be present in cladoselachians, but the significance of this feature is unclear. In Stethacanthus the hyomandibula is around half the length of the ceratohyal. This resembles the cladoselachian condition (Maisey, 1989a), but it contrasts with Cobelodus (Zangerl and Case, 1976) in which the hyomandibula is proportionately longer. There is no evidence of a separate (or fused) pharyngohyal (cf. discussions in Hotton, 1952; Zangerl and Case, 1976; Maisey, 1984, 1989a). The hyomandibula-ceratohyal articulation lay mesial and slightly posterior to the articulation between palatoquadrate and meckelian cartilages. As Dick (1978) comments on Tristychius, the entire articulation was probably padded with a thick connective tissue layer also serving as the origin for suspensory ligaments to the mandibular arch. The form of the hyoid articulation is better preserved than that of the mandibular arch. From HM V8246 it is fairly clear that it consists of a simple convex surface on the distal end of the hyomandibula and a correspondingly concave ceratohyal proximal surface. There is no interhyal; the possible interhyal identified in a cladoselachian specimen (Maisey, 1989a) appears to be the broken proximal end of the ceratohyal. The ceratohyal in Akmonistion zangerli (Fig. 3, chy) bears a well marked fossa in the lateral surface, just below the articulation with the hyomandibula. Otherwise the elongate ceratohyal is a slender, curved rod of near uniform thickness throughout its length. A subtriangular cartilage lies anterior to the ceratohyal and medial to the two halves of the lower jaw in HM V8246. This cartilage is probably the anteriormost (first) basibranchial (Fig. 3A, bbr). A similarly shaped and located first basibranchial occurs in the Bear Gulch Stethacanthus material (Lund, 1985b: fig. 6A). Thus far, no calcified basibranchials are known from cladoselachians, although a series of irregularly shaped basibranchials is associated with gill arches in Cobelodus (Zangerl and Case, 1976) and a variety are associated with the partial branchial skeletons of specimens attributed to Phoebodus (Williams, 1985; identification challenged by Ginter, 1998, who diagnoses Phoebodus as only a tooth-form taxon). There appear to be five gill arches present, the least disturbed series of which is displayed in HM V8246 (Fig. 3A). Preserved parts of the arches consist mostly of elongate, slender epibranchials (Fig. 3, ebr) and slightly longer, curved, ceratobranchials (Fig. 3, cbr). No hypobranchials have been identified. The three anteriormost epibranchials in HM V8246 each articulate with a slender, posteriorly directed, pharyngobranchial (Fig. 3, pbr). In general, the condition of these arches resembles that of cladoselachians (Maisey, 1989a). Close comparison may also be made with gill arch skeletons of the Bear Gulch stethacanthids

443

(Lund, 1985a, b), of which Falcatus shows similarly shaped pharyngobranchials. These slender cartilages are quite unlike the more substantial pharyngobranchials of Tristychius (Dick, 1978), Hybodus, and recent sharks such as Heterodontus (Maisey, 1983:fig. 9). Conditions in Acanthodes (Miles, 1973; Gardiner, 1984) indicate that slender, posteriorly directed pharyngobranchials are plesiomorphic. The anteriormost four epibranchials are simple rods, but the fifth has a broader, more complex outline (Fig. 3A, ebrV). The distinct morphology of the fifth gill arch resembles the pattern of Tristychius (Dick, 1978), in which the fifth ceratobranchial is broader than the preceding four; similarly distinct fifth arches are present in many Recent sharks. Ceratobranchial 5 of Akmonistion is insufficiently well known for evidence of specialisation comparable to the epibranchial condition. The hyoid and gill arches bear elongate branchial rays, the extent of which is displayed most clearly in HM V8246 (Fig. 3A, cr). As noted by Allis (1923) and others, such elongate rays support septa extending beyond the lateralmost limit of primary gill lamellae. It is noteworthy that while these slender and delicate gill rays are well preserved, there are no traces of extrabranchial cartilages whatsoever. The distribution and length of these rays matches those of cladoselachians (Maisey, 1989a) and Chlamydoselachus (Allis, 1923), but are quite unlike Tristychius in which a dense, elongate array is associated with the hyoid arch while only short branchial rays extend from the gill arches (Dick, 1978). Mandibular and Pharyngeal Dentition Scalloping and associated tooth platforms along the margins of the palatoquadrate and Meckel’s cartilage indicate the presence of at least six, and probably seven, well spaced tooth files per jaw ramus. Individual teeth resemble those of Gutturensis neilsoni (Sequeira and Coates, 2000), as well as teeth attributed to Stethacanthus (Williams, 1985; Zidek, 1993). Williams (1985) identified such teeth as corresponding to the form taxon Cladodus exilis St. John and Worthen (1875). Most have a large central cusp flanked by two pairs of lateral cusplets (terminology after Cappetta, 1987) (Fig. 4). Each central cusp is recurved lingually and slightly sigmoid in lateral aspect. Smaller teeth may have only a single cusplet pair (Fig. 4C), while some of the largest teeth have a third pair adjacent to the the central cusp base (Fig. 4F). Marginal cusplets are taller than proximal cusplets, and are between one quarter and one fifth the height of the central cusp (unlike Ohio and Oklahoma material and C. exilis, in which marginal cusplets are half the height of the central cusp). All cusps bear fine, closely packed, cristae (unlike the more widely spaced cristae of the Oklahoma teeth), and a more pronounced cutting edge divides labial from lingual surfaces. Where the central cusp broadens towards the base, further cristae intercalate between those extending from cusp base to apex (also unlike Ohio material and C. exilis, Williams, 1985). Cusplet cristae are fewer and more widely spaced than those of the central cusp, and at their bases they display a degree of twist, the rotation of which is mirrored on either side of the central cusp. Cusps consist of orthodentine, coated with what appears to be a thin enameloid monolayer. Lund (1985b) reported the presence of an enameloid layer on the Bear Gulch Stethacanthus teeth, and that the fluted profile of the external (enameloid) surface is not reflected at the subjacent enameloid-orthodentine junction. This description matches conditions revealed by broken cusp surfaces in Akmonistion. It is also noteworthy that in Akmonistion, each cusp is histologically separate from its neighbor. Cusp enameloid and orthodentine in the largest teeth is separated from the root by a short neck region, consisting of osteodentine/acellular bone identical to that of the root. Tooth roots (bases) are sub-elliptical in basal and apical views, and almost biconvex in labial view, with cusps positioned along the labial margin. Lateral margins are reflexed dor-

444


FIGURE 4. Akmonistion zangerli, UCMZ GN1047, mandibular teeth, all drawn to the same scale (note size difference between specimens E and C). A, labio-basal view; B, labial view; C and D in profile; E and F in labial view; G in lingual view. Scale bar equals 2 mm.

sally beyond the marginal cusps, and the labial margin projects aborally as a near-rectangular tongue beneath the central cusp. Unlike the Ohio Stethacanthus (Williams, 1985), in which this tongue extends to the level of proximal cusplets, tongue width is equal to or less than central cusp width. The labial surface is otherwise marked by a curved row of marginal foramina bordering the central cusp base. The convex orolingual surface bears a central protuberance (button) of about the same width as the central cusp (Fig. 4G), from which it is divided by a trough. When teeth are articulated in-file, this trough accommodates the tongue of the following tooth. Numerous small foramina pierce the orolingual surface, predominantly around the margins and cusp bases, and three, larger, vascular canals enter the lingual face of the aforementioned protuberence. The aboral root surface is concave and mostly featureless, except for the rectangular tongue projection and a thickened, raised, lingual rim (Fig. 4A). As such, the root is anaulacorhizous. Small foramina are distributed mostly adjacent to the lingual rim, while three or four larger foramina lie central to the shallow concavity. Three examples of tooth whorls with fused roots are present on UCMZ GN1047 (Fig. 5A–C). Each subunit of a whorl resembles a three-cusped tooth (a central cusp flanked on each side by a cusplet). The most complete specimen (Fig. 5B) consists of at least seven tooth units, the smallest and earliest of which is less than half the size of the largest, and ontogeneti-

cally youngest, member of the series. The fused roots of these units form a tightly curled base which turns through almost 360 degrees, so that the largest tooth is beginning to over-grow the smallest. Comparable tooth whorls attributed to Stethacanthus are described from Ohio (Williams, 1985:pl. 9, fig. 6) and Oklahoma (Zidek, 1993:fig. 1B) (Fig. 5D, E). Both of these are interpreted as symphysial (cf. edestid chondrichthyans, Zangerl, 1981), consist of up to five monocuspid teeth, and show no appreciable changes in tooth size. In fact, none of these whorls is preserved in near-articulation with a supporting skeletal structure, and their orientation in life is unclear. The whorls in UCMZ GN1047 are preserved below the rear half of the lower jaw, and it seems equally likely that they may have been located at the front of the gill skeleton, as may be the case for the smaller whorls known in Falcatus (Lund, 1985a; Zidek, 1993). The buccopharyngeal region in HM V8246 and UCMZ GN1047 includes a dense mat of buccopharyngeal denticles (Fig. 5F–J). These seem to have lined the entire mouth and pharynx, extending back from the mandibular margin to the pharyngoesophageal boundary. However, within this region there is no evidence of any more precisely regionalized pattern of the kinds documented in neoselachians (Nelson, 1970). Neither is there any evidence indicating that different varieties of compound denticle predominated in particular regions of the oropharynx. Buccopharyngeal denticle shapes are varied, and range from single-cusped cones to single-rowed (Fig. 5H, ‘‘Stemmatias simplex,’’ Williams, 1985), and, more rarely, double-rowed (Fig. 5I, ‘‘S. bicristus,’’ Williams, 1985), compound forms. From external appearance, the histology of these denticles is identical to that of the mandibular teeth. Each cusp is slightly recurved, bears cristae and lateral carinae, is closely appressed to its neighbor, and, if not solitary, is a member of a size-graded series. The largest may be more than triple the linear dimensions of the smallest, and the root and cusp of a larger member is recessed to partly envelop the base of a preceding, smaller, member. From the variety of forms preserved, there is every reason to believe that, unlike the squamation, this buccopharyngeal mat consisted of growing denticles; i.e., they were not formed in a single morphogenetic event (Reif, 1985). Instead, newer, larger, cusps were added to each compound denticle throughout its lifetime. The root of each denticle is pierced by a row of marginal foramina adjacent to the cusp base. This feature seems to be absent from the otherwise similar denticles from the Oklahoma Stethacanthus (Zidek, 1993:fig. 2). The maximum number of denticle units in any single compound series is around eight. Vertebral Column Almost the entire vertebral column is preserved in HM V8246 (Fig. 6). Preservation of this quality is exceptional for a Paleozoic chondrichthyan of this size. Superficially, the column resembles that of Cobelodus. The notochord is unconstricted, and while there are traces of perichordal calcification, these are confined to the tail. Most of the 901 pre-caudal neural arches (basidorsals or neurapophyses) consist of slender rods of prismatic calcified cartilage. Left and right sides of each arch are fused dorsally in at least the anterior third to half of the vertebral series. All arches are preserved in close association with anterior and posterior neighbours, but there are no signs of zygapophyseal surfaces or other interconnections. As in Cobelodus (Zangerl and Case, 1976), the vertebral series can be subdivided into approximately four regions: cervical (occiput to pectoral level), thoracic (pectoral to pelvic levels), peduncular (pelvic to caudal), and caudal (included within the tail description). The cervical region extends from vertebra 1 to about 14 (Fig. 6, pct.lv marks cervical–thoracic boundary), and compares more closely to the 15 cervical arches in Cobelodus (Zangerl and Case, 1976) than to the 9 or 10 in Falcatus (Lund, 1985b)


445

FIGURE 5. Tooth whorls and buccopharyngeal denticles. A–C, Akmonistion zangerli UCMZ GN1047, tooth whorls with fused bases: D and E attributed to Stethacanthus; D, redrawn from Zidek, 1993:fig. 1B, and E, from Williams, 1985:pl. 9, fig. 6. F–I, Akmonistion zangerli, UCMZ GN1047, buccopharyngeal denticles: all match form-taxon Stemmatias simplex except specimen I, matching S. bicristus; J, Akmonistion zangerli, NMS 1981.63.23A, buccopharyngeal denticle. Scale bars equal 1 mm.

or the 6 or 7 in Damocles (Lund, 1986). All cervical neural arches are subrectangular in lateral view; each is anteroposteriorly broad and, like the rest of the vertebrae, sloped posteriorly. Likewise, all cervical arches have a large foramen which probably enclosed one of the segmental nerve roots. Lund (1985b) suggests that the anteroposterior breadth of cervical arches is caused by incorporation of dorsal intercalaries or arcualia, but conditions in the new material fail to shed further light on this hypothesis. Each cervical neural arch apex is drawn into a short prong; in arches 1–6 this is directed anteriorly, and in 71 this is directed posteriorly. A gradual depression in total arch height flanking this switch in prong orientation

FIGURE 6.

accommodates the ventralmost extent of the dorsal spine baseplate (Fig. 15B; cf. Falcatus, Lund, 1985b:fig. 6B; Tristychius, Dick, 1978:text-fig. 26). The thoracic region is the most uniform of the vertebral column, and includes neural arches from around vertebra 15 to about 55 (Fig. 6, plv.lv marks thoracic–penduncular transition). All are anteroposteriorly narrow, closely packed, have a short, posteriorly directed spine, and most exhibit a foramen for a segmental nerve root. The thoracic series in HM V8246 curves towards the posteroventral edge of the spine-brush complex and the proximal radials (basals) of the pelvic-level dorsal fin. This curvature probably results from post mortem distortion follow-

Akmonistion zangerli, HMV8246. Complete, pre-caudal, vertebral column; anterior to left of figure. Scale bar equals 20 mm.

446


FIGURE 7. Akmonistion zangerli, spine and brush complexes. A, UCMZ GN1047; B, HMV8246; C, reconstruction, note that long-crowned scales are oriented laterally (see Fig. 15A); anterior to left of figure. Scale bars equal 30 mm.

ing contraction of ligaments and muscles anchoring dorsal outgrowths. The peduncular region is slightly longer than the thoracic, and extends from vertebra 56 to around 89 or 90, the origin of the tail (Fig. 6, cau.lv). Such proportions are closer to Cladoselache (after Zangerl, 1981) than Cobelodus or Symmorium, in which the (reconstructed) peduncle is shorter than the thoracic region. In comparison with thoracic arches, those of the peduncle are inclined more posteriorly and diminish in size towards the root of the tail. Around five of the posteriormost segments of the penduncle bear haemal arches (perhaps including basiventrals and haemapophyses). Each of these is preserved as a slender cartilage rod, but the distal parts of the anteriormost four are missing or damaged. It is therefore unknown if these formed a fused arch, and/or extended to articulate with the expanded, anteriormost radial of the hypochordal lobe of the tail (Fig. 6, a.h.rad).

Spine-Brush Complex and Dorsal Fin The Bearsden material includes several examples of the so-called ‘‘spine-brush’’ (Zangerl, 1981) complex (Fig. 7). In HM V8246 the entire structure appears to have moved very little relative to the axial skeleton post mortem (Fig. 1B, C). The spine and basal plate are situated above the branchial region, while the brush projects above the pectoral girdle. Coates et al. (1998) analyzed the histology of the spine, brush, and basal plate, and reached the following conclusions: (1), the spine (Fig. 7, sp) consists of osteonal dentine surrounded by acellular bone and lacks any enamel-like surface tissue; (2), the brush (Fig. 7, br) and basal plate (Fig. 7, bp) consist of non-prismatic globular calcified cartilage, the peripheral regions of which include a meshwork of crystal fibre bundles; and (3), a thin, acellular bone layer of variable thickness coats the leading edge and base of the brush, and edges of the basal plate. These results complement and extend upon previous in-

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN terpretations of spine-brush histology (Zangerl, 1984; Williams, 1985). The spine of A. zangerli resembles those attributed to S. altonensis (cf. Williams, 1985; Lund, 1984), and, to a lesser extent, S. productus (cf. Lund, 1984). All specimens of the spine of A. zangerli are crushed laterally. The overall form is of a right-angled triangle with a concave hypotenuse facing anterodorsally; the external surface is completely unornamented, and there is a prominent ‘‘horizontal shoulder’’ (cf. Lund, 1974:168; Williams, 1985:121). Just below the spine apex, a slight ridge passes from the anterior surface in a posteroventral direction across the lateral surface towards the rear edge. The posterior edge, most easily seen in UCMZ GN 1047 (Fig. 7A) is rounded, and there is no evidence of a vertical median ridge flanked by sulci, cf. S. altonensis (Williams, 1985; Lund, 1984:fig. 1). Unlike certain examples attributed to S. altonensis (Williams, 1985), no small denticles are associated with the anterior surface of the spine. Horizontal sections through a spine (NMS 1981.63.24) show that the gross morphology of the internal cavity and walls resembles that of a specimen attributed to S. depressus, with no indication of ontogenetic development from separate shaft and mantle regions (Zangerl, 1984). An arc of large canals, assumed to have enclosed a nutritive vascular supply, lies about half way between inner and outer surfaces of the spine wall. The topographic relation of the spine to the cartilaginous basal plate in A. zangerli is as described for S. altonensis (Williams, 1985). However, as reported by Coates et al. (1998), the acellular bone of the spine is continuous with that of the keelshaped basal plate, where it forms an anterior carina, and the globular calcified cartilage of the basal plate extends into the cavity of the spine. Furthermore, although the basal plate and spine are separated from the base of the brush complex in all Bearsden specimens, in each case this division lies in a different position and results from post mortem breakage. In life, the acellular bone of the basal plate and spine was continuous with that of the brush, and these three units must have formed a single, rigid, unit. There is no evidence of the condylar articulation between basal plate and brush described by Lund (1974) as present in S. altonensis. Brush structure has been described and discussed repeatedly (Lund, 1974; Zangerl, 1981, 1984; Wood, 1982; Williams, 1985; Zidek, 1993; Coates et al., 1998). In A. zangerli, both principal specimens of which are male, the brush expands dorsally to produce a denticle-bearing platform of around 160% of neurocranial length. The ‘‘fibres’’ of the brush are hollow rods (Williams, 1985) consisting of globular calcified cartilage (Coates et al., 1998). These rods are superficial to an interior which is not resolvable into separate units. In anterior, leading edge, and ventral regions, rods merge into a consolidated mass of bone-covered calcified cartilage. Anteriormost rods are generally better preserved than those that are close to the trailing edge. Irregularly shaped blocks of material resembling bonecovered calcified cartilage are located posteroventrally to all brush specimens (Fig. 7, ac.cart). Like similarly positioned structures in Damocles and Falcatus, these are interpreted as accessory supports for the complex (cf. Lund, 1985b, 1986). The brush platform of Akmonistion is laterally narrow and bears an array of prominent, long-crowned scales. In a male S. altonensis from Bear Gulch (MV 2830) there are up to nine scale rows across the brush anterior, tapering to four or five across the rear, and the posterior median scales bear the longest crowns (Lund, 1974). Alternatively, the probable female from the Logan Quarry Shale (FMNH PF2207) has up to seven rows across the brush and the posterior lateral crowns are longest (Williams, 1985:pl. 10). Compared with these, scale number and distribution in a male Akmonistion brush resembles that of the male S. altonensis, while crown length distribution resem-

447

FIGURE 8. Akmonistion zangerli, dorsal fins. A, HMV8246; B, UCMZ GN1047; anterior to left of figure. Scale bars equal 20 mm.

bles the female. In Akmonistion, scale bases are packed closely together, forming a continuous mosaic of rows aligned diagonally relative to the anteroposterior midline (reconstruction shown in Fig. 15A). Anteriorly, and at the posterior extremity of the platform, these rows are narrow and the scales small. The largest scales are situated laterally, along the posterior third of the platform edge. Counting straight across, from platform side to side, there may be as many as nine (small) scales anteriorly, and as few as four posteriorly. Scale morphologies are described in the squamation section. Finally, in both well preserved examples of the brush complex, a slender rod of what appears to be acellular bone lies within or projects from the anterior third of the brush apex (Fig. 7, r). The functional and structural significance of this structure is obscure, but it corresponds topographically to the ‘‘dorsal rod’’ supported by the spine in Damocles (Lund, 1986). Conjunction (Patterson, 1982) of rod and brush indicates that former is not a direct homologue of the latter (cf. Lund, 1986). The dorsal fin (Fig. 8) is situated at pelvic level, in a position identical to that of the single dorsal fin in anacanthous symmoriids (Zangerl, 1981) and the second dorsal fin in genera such as Goodrichthys (Moy-Thomas, 1936), Tristychius (Dick, 1978), and Onychoselache (Dick and Maisey, 1981). As in all Stethacanthus specimens (Lund, 1974; Zangerl, 1981; Williams, 1985), both examples of dorsal fins from A. zangerli are disrupted: in HM V8246 proximal radials are obscured by the vertebral column and the posterior of the fin is scattered; in UMCZ GN1047 the fin radials are clustered but oriented randomly. Conditions in Cobelodus and Symmorium (Zangerl and Case, 1976; Zangerl, 1981; Williams, 1985) probably provide the closest comparison. In HM V8246, the dorsal fin includes at least 21 radials divided into proximal and distal series. This compares closely with the estimated 24 radials in S. altonensis (Lund, 1974; Williams, 1985). The fin outline was probably subtriangular (radial length increases from the first to at least the fourteenth from the leading edge); there is no direct evidence for the short, posterior, proximal extension included in Zangerl’s (1984) restoration. Proximal radials (basals) (Fig. 8, p.rad) consist of laterally flattened cylinders (probably caused by post mortem compression), and each articulates with a single distal radial. Distal radials (Fig. 8, d.rad) are much longer than the proximal series, and neither branch nor articulate with any further, more distal series. Most of the distal radials of UCMZ GN1047 have

448


FIGURE 9.

Akmonistion zangerli, caudal fin. A, HMV8246; B, reconstruction; anterior to left of figure. Scale bar equals 30 mm.

rotated so that anterior or posterior surfaces are visible. These show that the distal radials (unlike the cylindrical proximal series) are calcified so that the proximal, dorsal and ventral surfaces form a slender triangular frame, while the central region is uncalcified and thus remains hollow. This morphology is characteristic of distal radials in all fins of A. zangerli. A delta-shaped cartilage (Fig. 8, d.crt), the homologue of Zangerl and Case’s (1976) ‘‘‘V’-shaped element,’’ lies posterior to the fin in UMCZ GN1047. The cartilage apex points dorsally in this disrupted example, but in less disturbed Cobelodus and Symmorium specimens it points caudally, with the plate oriented sagittally. Corroboration for this orientation comes from Falcatus (Lund, 1985b:fig. 13), where incompletely separated dorsal fin radials are continuous posteriorly with a triangular cartilage plate sharing distinct characteristics with symmoriid delta cartilages (Zangerl, 1981). In all examples the near triangular outline is acute, and the shortest side, facing anteriorly, is concave. The arms of the ‘V’ consist of the thickly mineralized dorsal and ventral plate edges. In A. zangerli the center of the cartilage is uncalcified; anteriorly there appears to have been a calcified bar connecting dorsal and ventral edges, but this is broken in UMCZ GN1047. Foramina perforate this bar dorsally and ventrally. The Caudal Fin and Axial Skeleton The heterocercal tail is well preserved in HM V8246 (Fig. 9A) and resembles the superficially symmetrical, semilunate tails described in Symmorium, Denaea, and Cobelodus (Zangerl, 1981; Williams, 1985). As in most of the well preserved specimens of these genera, the elongate hypochordal radials are grouped as a narrow and distally acute ventral caudal lobe, with the posteriormost radials separated from the caudal haemal spines. However, unlike other restorations, the tail of A. zangerli is restored so that all radials articulate closely with proximal endoskeletal structures. As a result, the radials of the ventral caudal lobe are splayed distally, the apparent surface area of the ventral lobe is increased, and the near-symmetry of the total caudal outline is lost (Fig. 9B). The restored hypochordal skeleton thus resembles more closely those of Falcatus (Lund, 1985b) and Damocles (Lund, 1986). Furthermore, Symmorium, Cobelodus, and Denaea probably had a similarly splayed arrangement, unlike current restorations (Zangerl, 1981), but rather similar to the

condition of the hypochordal lobe in a specimen of Cobelodus aculeatus (NMMNH P-19182) from the Kinney Quarry, New Mexico (Zidek, 1992:fig. 3A). The transition between peduncle and caudal regions occurs at about vertebral segment 91. The anterior caudal boundary is identified by the anteriormost vertebra supporting a hypochordal radial. This lies seven or eight segments in front of the narrow region in which the vertebral column (and notochord) turns dorsally to form the upper lobe of the tail. The region of hypochordal outgrowth thus extends anteriorly relative to epichordal outgrowth. The abbreviated heterocercal tails of Falcatus (Lund, 1985b) and Damocles (Lund, 1986) are thus different in that dorsal and ventral skeletal outgrowth levels coincide. Neural arches show a significant change in morphology at the level of the anteriormost point of hypochordal outgrowth. The spindle-shapes of the thoracic and penduncle regions are replaced by a series of eight or nine low, rhombic, plates (Fig. 9A, na). As in more anterior arches, each of these includes a segmental nerve foramen. A horizontally directed rod lies just above this series in HM V8246, but this seems to be a displaced hypochordal radial rather than epineural structure. No neural arch series similar to these is known in other early chondrichthyans, although this may result from incomplete preservation of otherwise similar tails. From the point of chordal upturn, at around the hundredth vertebral segment, neural arches either articulate distally with supraneural radials (Fig. 9A, sn.rad) or extend as lanceolate neural spines (Fig. 9A, n.sp). The preserved, prismatic calcified layer of these cartilages is notably thinner than elswhere in the endoskeleton. At least 23 radials and spines precede the posterodorsal caudal apex. Superficially, the broad spines resemble those of Cladoselache, and corroborate observations of similarly expanded epichordal supports in Stethacanthus (Lund, 1974; Williams, 1985) and Stethacanthulus (Zangerl, 1990), unlike the slender epichordal rays of Cobelodus, Denaea, and Symmorium. When reconstructed in A. zangerli it becomes apparent that the expanded apices overlapped and stiffened the dorsal leading edge of the caudal fin. The anteriormost hypochordal radial (Fig. 9A, a.h.rad) forms a subtriangular cut-water that articulated proximally with two


449

FIGURE 10. Akmonistion zangerli, pectoral girdle and fin. A, HMV8246; B, UCMZ GN1047; C, reconstruction; anterior to left of figure. Scale bars equal 20 mm.

or more haemal arches. Posterior to this the haemal arches and hypochordal radials (Fig. 9A, h.rad) extend to form the ventral lobe of the tail. Around twelve closely spaced radials contribute to the leading edge, and about eight, more widely spaced, to the trailing edge. Proximally, the haemal arches extend in length for the anteriomost five of the caudal region, then decrease through to the fourteenth. Haemal arches nine to fourteen are proximodistally short and resemble those described as ‘boomerang shaped’ in Symmorium (Williams, 1985). Eleventh and twelfth arches have the most squat profiles, are situated at a transformation in the proximal curvature of these arches, and may bear foramina (but these might equally be artifactual depressions cleared incompletely of matrix). From the fifteenth caudal haemal arch and more posteriorly, haemal spines (Fig. 9A, h.sp) are extended, and articulate with radials reaching the trailing hypochordal perimeter. More anterior arches articulate mostly with a series of short proximal radials. Despite the high

quality of preservation, it is not currently possible to distinguish pre-ural from ural caudal zones. Caudal haemal arches and spines 23–31 are clearly defined, but more posterior arches are incompletely separated from each other, and form a raggededged ribbon of prismatic cartilage. The space occupied by the caudal stretch of notochord is almost undistorted and mostly empty. However, it includes a chain of small calcifications which are interpreted as having formed within the perichordal sheath (Fig. 9A, pc). Finally, it is noteworthy that at the point of caudal dorsal upturn, there appears to be a significant loss of segmental register between neural and haemal arches. Given the quality of preservation, this is unlikely to be artifactual. Pectoral Girdle Details of pectoral girdle morphology are displayed most clearly on UCMZ GN1047 (Fig. 10B), although in-life orientation is preserved better on HM V8246 (Fig. 10A). Each half of the girdle consists of a tall scapulocoracoid and a

450


subtriangular procoracoid cartilage connecting with the anteroventral surface of the coracoid region. The scapular process is a near-parallel sided cartilage sheet, the dorsal apex of which expands into anterodorsal and posterodorsal processes (Fig. 10, sc, apr, ppr). From HM V8246 it appears that the long axis of the scapular process slopes posteriorly relative to the axial skeleton. The scapular process base is thickened mediolaterally and expanded in a mostly posterior direction to form and support the articular surface for the pectoral fin radials. A large foramen (Fig. 10, dzf) perforates the cartilage just above the articular area. This probably enclosed the diazonal nerve and brachial artery. The articular surface consists of a well formed, robust, rounded crest (Fig. 10, art.c). Oriented horizontally, the crest extends anteriorly from the posteriormost prominence for almost three quarters of the total scapulocoracoid width. The posterior end of the articular crest is expanded to form a condylar surface at the articulation with the metapterygium (Fig. 10, mpt. c). The narrow cartilage surface immediately ventral to this crest is smooth and featureless. The medial face of the scapulocoracoid is exposed on UCMZ GN1047. The articular region is penetrated dorsally by the slotshaped diazonal nerve foramen. As in Cobelodus (Zangerl and Case, 1976), the medial surface of the scapular process is dominated by a dorsoventrally oriented trough, the posterior edge of which forms a prominent, anteriorly convex, curved ridge. The coracoid region (Fig. 10, cor) is convex anteriorly and concave posteriorly, and, in life, curved medially. The procoracoid cartilage (Fig. 10, pcor) is assumed to be homologous with the similarly positioned cartilage plates in xenacanths and the ‘claw-shaped elements’ of Cobelodus, Symmorium and Denaea. However, unlike these, the procoracoids are not directed posteriorly, and appear to have moved little from their former articulated position in HM V8246. Pectoral Fin Pectoral fins are preserved most completely in HMV8246 (Fig. 10A), where those of left and right sides are superpositioned. General fin morphology resembles those of Cobelodus (Zangerl and Case, 1976) and Denaea (Williams, 1985). Eight or nine short proximal radials articulate with the girdle directly; rather fewer than the thirteen in Stethacanthus altonensis (Zangerl, 1981). In UCMZ GN1047 these consist of anteroposteriorly flattened cylinders, slightly expanded at proximal and distal articular surfaces. Each proximal radial articulates with a single distal radial, except the anteriormost, which may terminate at the fin leading edge. The fifteen or so distal radials are much longer than the proximal series. Distal radials neither branch nor articulate with any further radials. The posteriormost proximal endoskeletal fin support, usually identified as the metapterygium (Fig. 10, mpt), is a broad, neartriangular plate of about the same anteroposterior length as the scapulocoracoid articular ridge. The proximal metapterygial articular surface is concave and faces anteriorly, towards the condyle on the rear of the scapulocoracoid. The distal metapterygial edge bears a stepped series of six articular facets for distal radials. These facets are not separated by deep grooves, as in Stethacanthus altonensis (Zangerl, 1981, 1984), Symmorium (Williams, 1985), and Falcatus (Lund, 1985b). The posterior metapterygial surface articulates with the anteriormost of a series of about 22 axial radials (Fig. 10, ax.rad) forming a whiplike pectoral fin extension. The anteriomost three axial radials are shorter than more posterior members of the series; this resembles conditions in the similar pectoral whips of Cobelodus and several other primitive chondrichthyans. Unlike Gutturensis neilsoni (Sequeira and Coates, 2000), no distal radials articulate with these anterior cartilages. From the barely disturbed skeleton of HM V8246 it appears that the complete whip was almost as long, proportionally, as that of Denaea (Williams, 1985). The complete distal extent of the pectoral fin is uncertain. In

FIGURE 11. Akmonistion zangerli, pelvic girdle, fin and clasper. A, UCMZ GN1047, anterior to right of figure; B, HMV8246, anterior to left of figure; C, reconstruction, anterior to right of figure. Scale bars equal 20 mm.

HM V8246 a band of tiny skin denticles beyond the posterior distal radials (Fig. 10A) provides some indication of fin size (employed to predict fin proportions in the reconstruction shown in Fig. 16). Despite the otherwise exceptional preservation of this specimen, there are no traces of ceratotrichia. Pelvic Girdle The pelvic girdle consists of simple sheets of prismatic cartilage, each of which is a subtriangular, rounded, plate (Fig. 11, p.pl). The most completely preserved examples are present on UMCZ GN1047. The anterior edge of each pelvic plate is slightly concave, and the posterior is convex. The girdle halves are reconstructed as lying horizontally in the ventral body-wall musculature. The lateral edge, articulating with the proximal radials of the fin, bears separate articular facets. There is no suggestion of any symphysial union between left and right pelvic plates. Five diazonal nerve foramina (Fig. 11, dzf) pierce the pelvic plate adjacent to the lateral edge, and these are accompanied by a larger and slightly more posteromedially situated foramen. A similarly large foramen is present in the pelvic plates of Cobelodus and Denaea (Williams, 1985), but is lacking from the Bear Gulch Stethacanthus altonensis (Lund, 1984:fig. 9). Pelvic Fin Pelvic fins of male specimens include eleven or twelve proximal radials articulating directly with the girdle (Fig. 11). In HM V8246 the posteriormost, premetapterygial, proximal radial bifurcates and articulates with two distal radials.

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN Distal radials are most completely preserved in HM V8246, and, like those elsewhere, are unbranched. Two distal radials probably articulated with the metapterygium (Fig. 11B, mpt), which includes a large right-angled notch accommodating the most proximal of a series of around seven axial or ‘basal’ cartilages (Fig. 11, ax.crt). Each of these articulates with further short, broad, radials; two per segment for the larger members of the series. Axial cartilages, the myxopterygium and terminal cartilages seem to consist of globular calcified cartilage like that of the spine-brush complex (Coates et al., 1998), rather than prismatic cartilage. The rod shaped myxopterygium (Fig. 11, mxpt) is gently curved, and nearly the same length as the pelvic fin and axial cartilages combined. All myxopterygia show transverse fractures suggesting incomplete division into or construction from a segmented pattern (cf. segmented myxopterygia in Falcatus, Lund, 1985b). However, the irregularity of fracture distribution within and between specimens indicates that these apparent segments may be preservational artifacts. The myxopterygial core was unmineralized, and the rod is preserved as a flattened cylinder broken along the dorsomedial edge. Calcification is thickest distally and ventromesially, while proximally and dorsolaterally mineralization consists of no more than small, isolated, calcified ‘islands.’ A broad, shallow, groove (Fig. 11A, grv) passes across the dorsolateral myxopterygial surface in a proximoventral to distodorsal direction. This groove is undescribed in other fossil claspers. Comparison with clasper anatomy in Chlamydoselachus (Smith, 1937) suggests that the groove may reflect the location of an erector/expansor muscle, or the site of a venous sinus extending from the iliac vein. Terminal clasper parts include at least two and probably three or more cartilages (Fig. 11, t.crt). Unlike Cobelodus and Denaea (Williams, 1985), none of these are continuous with the myxopterygium. In both preserved clasper pairs, the most prominent terminal cartilage is ovoid with a convex surface marked by a shallow groove, oriented anteroposteriorly. This convex unit is faced by a similarly shaped concave cartilage, with fragments of a third protruding from between these valvelike structures in UCMZ GN1047. At least one smaller cartilage lies between the proximal surfaces of the terminal cartilages and the distal end of the myxopterygium. The concave terminal cartilage resembles a proportionally larger concave component in Falcatus. It is noteworthy that there are no plates consisting of or supporting acute scales, as in Damocles (Lund, 1986) or primitive holocephalans (Lund, 1982). Scales Although Akmonistion and other stethacanthids could be described as predominantly scale-less, this would overlook the fact that they are among the most remarkably heterosquamous chondrichthyans discovered thus far. In fact, Akmonistion retains four distinct scale regions: (1), the cranial cap scales (Fig. 12A–F); (2), the brush apical scales (Fig. 13); (3), an area of miniscule flank scales (Fig. 12H); and (4), lateral line scales (Fig. 12G). All scales are single crowned and are therefore inferred to be non-growing, i.e., formed in a single morphogenetic event (Reif, 1985). Each scale could also be described as consisting of a single odontode (Ørvig, 1977). This is strikingly dissimilar to the ‘polyodontode’ (Ørvig, 1977) scales of genera such as Tamiobatis (Williams, 1998) and Diplodoselache (Dick, 1981), in which growth occurred by the apparent welding of newly formed scales to pre-existing examples (Reif, 1978, 1985). The scales of Akmonistion might also be described as ‘placoid,’ implying that they match traditional or more precise descriptions of neoselachian scales (e.g., Romer and Parsons, 1977:158; Reif, 1985:16). But Akmonistion scales lack a key placoid feature: no crown has been demonstrated to consist of anything other than dentine: an enamel and/ or enameloid cap appears to be absent. Otherwise, these scales are noteworthy for the massive bases, consisting of acellular

451

FIGURE 12. Akmonistion zangerli, scales from the cranial cap, lateral line, and flank region. A–C: cranial cap scales from UCMZ GN1047, anterior to right of figure; single cranial cap scale from UCMZ GN1047 in D, lateral view (anterior to right), E, basal view (anterior to top), and F, apical view (anterior to top); G, lateral line scales from HM V8246, anterior to left of figure; H, flank scales from UCMZ GN1047. A–F, scale bar equals 2 mm; G and H, scale bar equals 1 mm.

bone, that characterise examples from cranial cap and brush apical regions. Most details of the cranial cap scales and brush apical scales match the description of corresponding material attributed to S. altonensis by Williams (1985), and thus include forms resembling Lambdodus hamulus (St. John and Worthen, 1875; Figs. 12B, C, 13E) and Cladodus pattersoni (Newberry, 1889; Figs. 12D–F, 13A–D). These scales are an order of magnitude larger than lateral line and other body scales, and the crowns have distinct carinae close to the apex. The cranial cap scale pavement has disintegrated in all Bearsden specimens. AMNH 1734, attributed to S. altonensis and figured by Zangerl (1981) and Williams (1985), thus remains the best source of information about scale distribution in a closely comparable region. Smaller scales are situated anteriorly and laterally, larger scales medially and posteriorly, and all crowns are directed posteriorly (pers. obs., MIC). Individual cranial cap scales are well preserved in HM V8246, UCMZ GN1047, and NMS 1981.63.23A. Scales resembling Lambdodus hamulus (Fig. 12B, C) have a smooth, recurved, crown, arching back over an extended, quadrangular, base with a coarse, fluted and ridged surface. A scale with an apparently simpler, more equilateral base (Fig. 12D–F), therefore resembling Cladodus pattersoni, has been removed from the matrix. This reveals a grooved underside with a pair of large, centrally placed, basal canal foramina (Fig. 12E). Smaller foramina, assumed to open into neck canals, are clustered above and below a central prominence on the posterior surface. In basal and apical views, the modified square outline (regarded by Reif, 1985, as the general and possibly plesiomorphic condition) of the scale base indicates the way in which it overlapped anterior and underlay posterior neighbours (assuming a posteriorly directed crown). This indicates that such scales were interarticulated within predominantly anteroposteriorly directed series. No neck

452


FIGURE 13. Akmonistion zangerli, scales from the brush apex of UCMZ GN1047. A, long-crowned, posterolateral scale; dorsal view; B, midline scale, lateral view, cf. Cladodus pattersoni; C, scale base of long-crowned, posterolateral scale in dorsal view; D, profile of long-crowned, posterolateral scale in anteroposterior aspect, ventrally directed surface to left of figure; E, scale from anterior of brush apex, cf. Lambdodus hamulus, lateral view, anterior to right. F, patch of associated scales from anterior of brush apex. All scale bars equal 2 mm; cross-hatching indicates broken crown surface.

canal foramina appear on the anterior surface, but smaller, nutritive foramina are clustered around the crown base. Lateral line scales (Fig. 12G) are preserved only in HM V8246, where a short row lies between the dorsal extremities of the neural arches and the ventral edge of the spine basal plate. The scales are remarkably small relative to total body size, measuring around 0.5 mm across. Each scale is triradiate, with a single cusp and a crescent-shaped base. As in the cranial cap and spine-brush scales, the crown is smooth and curved. The scale base consists of a pale material and the crown of darker material; a collar region where the crown meets the scale base is translucent, revealing dentine tubules. Two or more canals penetrate the scale base below the crown. Pairs of these scales are arranged so that the crescentic bases meet, and create a C-shaped incomplete ring around the course of the sensory canal. Again, this resembles conditions in Orestiacanthus (Lund, 1984). The miniscule body scales appear to be vestigial, Petroduslike, buttons, about 0.2 mm across with a subcircular base and minimal, apical, blister of dentine (Fig. 12H). These scales are known only from a patch covering the flank above the pelvic plate in UCMZ GN1047. Airbrasive preparation has removed much of their surface detail, and it is uncertain if they were present elsewhere. Williams (1985) reports the presence of similar denticles in the orbit region of AMNH 1734, while Lund

FIGURE 14.

Akmonistion zangerli, cololite. Scale bar equals 5 mm.

(1984) describes general, if well spaced, covering of similar scales in the stethacanthid Orestiacanthus. Brush apical scales, like those of the cranial cap, also include forms comparable to Lambdodus hamulus (St. John and Worthen, 1875) (Fig. 13E) and Cladodus pattersoni (Newberry, 1889) (Fig. 13B), but the morphological variety is more diverse. Crowns are directed posteriorly, except for those situated along the posterolateral margin which face laterally (Figs. 7, 15). Scale size varies considerably from anterior to posterior of the brush apex, the overall shape of which is slender and streamlined. From HM V8246 and UCMZ GN1047, the rear third of the brush apex seems to have been no more than four scales across. The smallest scales, some 2 mm long, are clustered around and over the leading edge (Fig. 13F) forming an anterior buckler, whereas the largest scales have crowns which are about 20 mm long (equivalent to almost one third of total brush height), overhanging the posterior third of the brush lateral surface (Fig. 13A, D). Posterior scales closer to the midline have crowns which are no longer than 5mm. Bases of the largest scales (Fig. 13A, C, D) have a concave apical surface with prolonged buttresses flanking, but not ventral to, the crown base. The dorsoventrally deep sides of these scale bases are crenelated, and adjacent bases seem to have interdigitated, conferring greater rigidity to the total platform surface. Each of the longest scale crowns emerges dorsally from its base, but bends sideways and extends outwards with a slight sinusoidal curve (Fig. 13D). Gut Trace and Cololite An area of debris posterior to the branchial skeleton and pectoral girdle in HMV8246 probably represents remains of stomach/foregut contents. The extent of this region is shown in Figure 1C (light shaded with a dark border). Debris includes small fragments (around 1–2 mm across) of arthropod cuticle and actinopterygian scales (pers. comm., Neil Clark; pers. obs., MIC.). A second region of preserved gut contents lies directly above the pelvic plates and ventral to the dorsal fin. This squat bolus (Fig. 14) measures about 15 mm along its longest axis. The posterior end is blunt,

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN two or three deep incisions and several minor striae indicate its spiral form, and it resembles many isolated coprolites found in the fish-bearing beds of the Manse Burn Formation. However, the position of this particular example indicates that it had not been expelled from the intestinal tract prior to fossilisation, hence use of the term cololite instead of coprolite (Hunt, 1992). DISCUSSION The skeletal reconstruction of Akmonistion (Fig. 15) and its restoration in life (Fig. 16) illustrate the unusual proportions of this species. There is no close analogue of the tail morphology among Recent chondrichthyans. High aspect ratio tails of extant pelagic genera (e.g., Rhincodon, Isurus, Lamna, Carcharhinus) share the presence of a steeply upturned caudal lobe, but none possesses the larger, and endoskeletally supported, hypochordal lobe. The unconsolidated vertebral column suggests a degree of anguiliform axial flexibility like that of Recent lampreys, and the apparently small pectorals and pelvics suggest paired fin proportions similar to those of modern pseudotriakid and scyliorhinid sharks (pers. comm., M. Gottfried). Hydrodynamic performance must have been affected by the spine and brush complex. The structure probably generated considerable drag, and the broad lateral surface area and anterior position of this keel-like outgrowth would reduce yaw to a minimum. Median fin distribution is configured inappropriately for sudden acceleration, and, given the size of the spine and brush complex, so too is sustained, high speed, pursuit of agile prey. Akmonistion is thus envisaged as a slow, steady, swimmer, suited for opportunistic scavenging and preying on benthic invertebrates. This contrasts strongly with the inferred swimming and prey capture modes of the Devonian genus Cladoselache (Williams, 1990). The function of the spine and brush complex remains obscure, although it is now clear that in Akmonistion, at least, it was rigid, immobile, and unlikely to have served as an accessory clasper and/or a jaw-like threat display (see discussions in Zangerl, 1984; Lund, 1985b; Williams, 1985; Coates et al., 1998). No Recent chondrichthyan possesses a similar range and distribution of specialized scales, although otherwise reduced squamations with scatterings or discrete patches of enlarged, thorn-like scales, occur in numerous skates and rays (most notably the bowmouth guitarfish, Rhina ancylostoma) as well as the bramble shark, Echinorhinus (Reif, 1985). The unusual squamation of Echinorhinus has special relevance for Reif’s (1980, 1985) model of chondrichthyan selforganizing scale patterns. This model hypothesizes that each (placoid) scale is surrounded by an inhibitory field preventing new scales forming in close proximity to existing scales. Echinorhinus, however, breaks the general rule (and apparent neoselachian synapomorphy) of regular scale spacing, because its large, thorn-like, scales can occur with bases fused or welded together (Reif, 1985:pl. 15). This implies absence or suppression of the hypothesized inhibitory field. Resemblance to conditions in primitive chondrichthyans, including xenacanths (Hampe, 1997) as well as the cranial cap and brush apex scales of stethacanthids, suggests that such suppression has been a repeated phenomenon in chondrichthyan phylogeny. The largest scales of stethacanthids have been compared to teeth, as if they shared a specialized developmental relationship like that proposed for the teeth and extra-oral dermal denticles in Denticeps clupeoides (Sire et al., 1998). This suggestion is rejected for the following reasons: morphology, because comparisons between Figures 4, 5, 12, and 13 yield numerous differences clustering scales and teeth of Akmonistion into distinct sets; histology, because microscopical inspection of broken surfaces indicates that enamel/enameloid is absent from scales whereas it forms the outermost layer of tooth crowns; and ontogeny, because no evidence suggests that the largest scales

453

resulted from a serial-replacement ontogeny resembling that of chondrichthyan (and other primitive gnathostome) dentitions. The phylogenetic position of Akmonistion and other stethacanthids among the Chondrichthyes is a moot point. According to Zangerl (1990), the family Stethacanthidae includes Stethacanthus (Newberry), Orestiacanthus (Lund), Stethacanthulus (Zangerl), and Bethacanthulus (Zangerl). Falcatus (Lund) and Damocles (Lund) are placed in a separate family, the Falcatidae (Zangerl, 1990). As noted in the introduction, Akmonistion clearly fits within Zangerl’s (1990) definition of the Stethacanthidae rather than the Falcatidae, but these families can be united at a more inclusive level by the presence of a specialized pectoral level dorsal spine without any closely associated pectoral level dorsal fin (in males, at least). In an earlier scheme, Lund (1985a) grouped Orestiacanthus with the Falcatus-like genera on the basis of a shared, sagitally compressed, spine morphology, while Stethacanthus clustered with the symmoriids because of shared features of the pelvic plate and metapterygium. More recently, Lund and Grogan (1997:fig. 6; Node 48) found several synapomorphies to unite Falcatus and Stethacanthus as the Stethacanthidae, which their analysis placed at the apex of a Paleozoic shark clade extending from the base of the elasmobranch stem-lineage (Lund and Grogan, 1997:fig. 6; elasmobranchs distal to Node 57). Synapomorphies of the Stethacanthidae sensu Lund and Grogan (1997) include the presence of scales restricted to specialized areas (ethmo-rostral, supraorbital, and dorsal fin-crest zones); ringshaped lateral line scales; a sexually dimorphic first dorsal fin with an unornamented triangular spine; the first dorsal fin developing at puberty; an elongate pectoral post-metapterygial axis; and absence of an anal fin and pelvic basipterygium. However, Lund and Grogan’s analysis focused on chimaeriform interrelationships and they included few non-holocephalan chondrichthyans. Most notably, Cladoselache and the symmoriids (sensu Zangerl, 1981) are not included, although these taxa exhibit many of the hypothesized stethacanthid apomorphies. It is also relevant to the present work that they characterized the brush complex as a dorsal fin, an interpretation that we, like Williams (1985), reject. Here, the brush complex is characterized as a specialized basal plate outgrowth, and is thus mutually exclusive of radial and ray supported fins (Coates et al., 1998; Coates and Sequeira, 2001). Importantly, Lund and Grogan’s (1997) cladogram resembles several other hypotheses of primitive chondrichthyan phylogeny which place stethacanthids as stemgroup elasmobranchs. These include examples by Young (1982:text-fig. 9B), Gaudin (1991:fig. 2), Coates and Sequeira (1998:fig. 9), and Maisey (2001:fig. 2). Two major alternative positions for stethacanthids have also been proposed, either as stem-group holocephalans (Janvier, 1996:fig. 4.39, left hand side) or as stem-group chondrichthyans (Janvier, 1996:fig. 4.39, right hand side; this resembles Maisey’s (1984) hypothesis of symmoriids and Cladoselache as stem-group chondrichthyans). The remainder of this discussion summarizes the results of an ongoing reevaluation of all three alternatives. New data from Akmonistion has contributed to a data matrix (appendices 1–3) of 86 non-additive binary coded characters (Pleijel, 1995) for an in-group including 19 chondrichthyan genera and 3 non-chondrichthyans (Coates and Sequeira, 2001). Using PAUP 3.1.1 (Swofford, 1993; since replicated using PAUP 4.0b3, Swofford, 1999) with heuristic search options, one tree was found of 161 steps (Fig. 17A). The primary feature of this result is that stethacanthids and symmoriids emerge as stem-group holocephalans (cf. Janvier’s [1996] novel conjucture). Tree support statistics are detailed in the figure caption, but it is noteworthy that Bremer support values (Bremer, 1994) are low: at four extra steps all resolution is lost above node D (Fig. 17) in a strict consensus of the resultant 2064 trees.

454


FIGURE 15. Akmonistion zangerli, reconstruction in A, dorsal, and B, lateral, aspects. Scale bar equals 50 mm. Pectoral fins in A extended horizontally; in B flexed ventrally; elongate gill rays shown extending from hyoid arch but omitted from gill arches.


455

FIGURE 16. Akmonistion zangerli, reconstruction in life. Extent of gill covers based upon branchial ray length and comparable conditions in Chlamydoselachus (Smith, 1937); likewise tooth families around jaw margins, which, based upon preserved size range (Fig. 5), probably retained worn teeth displaced externally relative to the functional position (cf. Gutturensis, Sequeira and Coates, 2000). Pectoral fins flexed ventrally as in Figure 15. Scale bar equals 50 mm.

FIGURE 17. Cladograms, adapted from Coates and Sequeira (in press). A, single shortest tree, 161 steps, consistency index: 0.53; B, single tree resulting from character reweighting (rescaled consistency index and best fit options) using 2,064 trees saved at a maximum of 165 steps. Underlined figures: Bremer support values; double figures: 50% majority rule values for nodes in set of 2,064 trees saved at maximum of 165 steps. Box encloses holocephalan/chimaeroid taxa.

Preliminary investigation of the data set relocated stethacanthids and symmoriids as a monophyletic clade of stem-group elasmobranchs, leaving only genera above node K (holocephalans) branching from node D. This transformation, resembling the tree topologies proposed by Young (1982) and others, increased tree length by only 4 steps (MacCLADE v. 3.05, Maddison and Maddison, 1993). Although this suboptimal solution is one of the many trees of 165 steps or less, a nearidentical branching sequence was found after reanalysis of characters re-weighted according to data from all 2064 trees. This tree, shown in Figure 17B, is near-identical to the described transformation, except for the more basal position of Cladoselache. Alternatively, when symmoriids and stethacanthids are shifted to a stem-group chondrichthyan position, tree length increases to 169 steps. This configuration has not been investigated further. Characters uniting Akmonistion with the holocephalans at node H (Fig. 17A) include the following: 4, ring- or C-shaped sensory canal scales; 6, an anteriorly concave spine shape in lateral view (reversed to absent at node K); 13, a calcified anterior dorsal fin basal plate; 50, calcified perichordal centra. Characters at node E uniting symmoriids with stethacanthids plus holocephalans include the following: 2, scales reduced or absent; 8, a spineless pelvic level dorsal fin; 15, a delta-shaped cartilage at pelvic level (reversed to absent at node K); 19, a lunate tail (reversed to absent at node I); 27, a broad pectoral fin insertion (reversed to absent at node I); 52, a laterally directed fossa on the palatoquadrate otic process (reversed to absent at node K); 58, a semicrescent-shaped hyomandibula (unknown in genera included above node K). The hypothesis of symmoriids and stethacanthids as stemholocephalans therefore depends upon a series of character states unknown and implied state reversals in the apical group of Harpagofututor, Helodus and Ischyodus. In the alternative and arguably more conventional tree topology (Fig. 17B), characters 15, 19, 27, 52, and 58 perform significantly better as synapomorphies uniting the monophyletic clade of stethacanthids and symmoriids. However, the character set joining this clade with elasmobranchs above the divergence from holocephalans is less robust. These include characters 65, emergence of the glossopharyngeal nerve through the metotic fissure (reversed in Falcatus, Damocles, and at node R); 73, dorsomedially directed endolymphatic ducts (unknown in included fossil

456


holocephalans); 76, hyoid articulation at the posterolateral angle of the otic capsule (reversed in Falcatus and Damocles; plesiomorphic holocephalan condition unknown); 79, palatoquadrate articulates on the rear of the postorbital process (plesiomorphic holocephalan condition unknown); and 86, presence of a precerebral fontanelle (possible homology with the holocephalan ethmoid canal is uncertain; once again the plesiomorphic holocephalan condition is unknown). The character state distribution for this branching sequence (Fig. 17B) is clearly not more satisfactory than that of the shortest tree (Fig. 17A). In conclusion, available evidence indicates that stethacanthids are crowngroup chondrichthyans rather than members of the stem lineage. Attempts to more fully resolve early chondrichthyan interrelationships will need to include more taxa and characters. The present matrix lacks characters describing the dentition, and characters describing the squamation are minimal. Primitive holocephalan conditions may be informed by the addition of taxa such as iniopterygians and eugeneodontids (reviewed in Zangerl, 1981; Janvier, 1996), and redescriptions of genera such as Pucapampella (Maisey, 2001) and Antarctilamna (Young, 1982; existing referred material may include more than one taxon, and neurocranial specimens require comprehensive reinterpretation, pers obs. MIC) may yet shed light on conditions preceding the basal divergence of modern chondrichthyan clades. ACKNOWLEDGMENTS We thank the staffs of the Hunterian Museum, University of Glasgow, the Zoology Museum, University of Cambridge, the National Museum of Scotland, Edinburgh, the Natural History Museum, London, the American Museum of Natural History, New York, the Cleveland Museum of Natural History, The Field Museum, Chicago, and the Carnegie Museum of Natural History, Pittsburgh, for access to collections, loan of specimens, and permission to prepare selected material. Particular thanks are due to Dr. J. A. Clack, University of Cambridge, for allowing extensive use of laboratory facilities, and Drs. J. G. Maisey, R. Lund, and E. D. Grogan for valuable discussions about early sharks. Drs. J. Long, M. Williams, O. Hampe, and M. Gottfried reviewed the manuscript and provided valuable comments. This research was funded by BBSRC advanced research fellowship B/94/AF/1945; project S05829. LITERATURE CITED Allis, E. P., Jr. 1923. The cranial anatomy of Chlamydoselachus anguineus. Acta Zoologica 4:123–221. Bremer, K. 1994. Branch support and tree stability. Cladistics 10:295– 304. Cappetta, H. 1987. Chondrichthyes II: Mesozoic and Cenozoic Elasmobranchii. H. P. Schultze (ed.), Handbook of Paleoichthyology, 3B. Gustav Fischer Verlag, Stuttgart, New York, 193 pp. Chorn, J., and K. N. Whetstone. 1978. On the use of the term nomen vanum in taxonomy. Journal of Paleontology 52:494. Clark, N. D. L. 1989. A study of Namurian crustacean-bearing shale from the western Midland Valley of Scotland, Ph.D. dissertation, University of Glasgow, 271 pp. Coates, M. I. 1988. A new fauna of Namurian (Upper Carboniferous) fish from Bearsden, Glasgow, Ph.D. dissertation. University of Newcastle Upon Tyne, 308 pp. ——— 1993. New actinopterygian fish from the Namurian Manse Burne Formation of Bearsden, Scotland. Palaeontology 36:123– 146. ——— 1994. The origin of vertebrate limbs; pp. 169–180 in M. Akam, P. Holland, P. Ingham, and G. Wray (eds.), The Evolution of Developmental Mechanisms. Development Supplement 1994. ——— 1998. Actinopterygians from the Namurian of Bearsden, Scotland, with comments on early actinopterygian neurocrania. Zoological Journal of the Linnean Society 122:27–59. ———, and S. E. K. Sequeira. 1998. The braincase of a primitive shark.

Transactions of the Royal Society of Edinburgh: Earth Sciences 89: 63–85. ———, and ——— 2001. Early sharks and primitive gnathostome interrelationships; pp. 241–262 in P. E. Ahlberg (ed.), Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny and Development. Taylor and Francis, for the Systematics Association, London. ———, ———, I. J. Sansom, and M. M. Smith. 1998. Spines and tissues of ancient sharks. Nature 396:729–730. Denison, R. H. 1979. Acanthodii. H.-P. Schultze (ed.), Handbook of Paleoichthyology, 5. Gustav Fischer Verlag, Stuttgart, New York, 62 pp. Dick, J. R. F. 1978. On the Carboniferous shark Tristychius arcuatus Agassiz from Scotland. Transactions of the Royal Society of Edinburgh 70:63–109. ——— 1981. Diplodoselachi woodi gen. et sp. nov., an early Carboniferous shark from the Midland Valley of Scotland. Transactions of the Royal Society of Edinburgh, Earth Sciences 72:99–103. ———, M. I. Coates, and W. D. I. Rolfe. 1986. Fossil sharks. Geology Today 2:82–84. ———, and J. G. Maisey. 1981. The Scottish Lower Carboniferous shark Onychoselache traquairi. Palaeontology 23:363–374. Gardiner, B. G. 1984. The relationships of the palaeoniscid fishes, a review based on new specimens of Mimia and Moythomasia from the Upper Devonian of Western Australia. Bulletin of the British Museum (Natural History), Geology 37:173–428. Ginter, M. 1998. Taxonomic problems with Carboniferous ‘‘Cladodontlevel’’ sharks’ teeth. Ichthyolith Issues Special Publication 4:14– 16. Goujet, D. 1984. Les poissons Placodermes du Spitsberg. Arthrodires Dolichothoraci de la Formation de Wood Bay (De´vonien infe´rieur). Cahiers de Paleóntologie, Centre national de la Reche`rche scientifique, Paris, 284 pp. Gross, W. 1937. Das Kopfskelett von Cladodus wildungensis Jaekel. 1. Teil. Endocranium und Palatoquadratum. Senckenbergiana 19:80– 107. Hampe, O. 1997. Zur funktionellen Deutung des Dorsalstachels und de Placoidschuppen der Xenacanthida (Chondrichthyes: Elasmobranchii; Unterperm). Neues Jahrbuch fu¨r Geologie und Palaöntologie, Abhandlungen 206:29–51. Harris, J. E. 1938. I. The dorsal spine of Cladoselache. II. The neurocranium and jaws of Cladoselache. Scientific Publications of the Cleveland Museum of Natural History 8:1–12. Heidtke, U. 1982. Der Xenacanthide Orthacanthus senckenbergianus aus dem pfa¨lzischen Rotliegende (Unter-Perm). Polichia 70:65–86. Hotton, N. 1952. Jaws and teeth of American xenacanth sharks. Journal of Paleontology 26:489–500. Hunt, A. P. 1992. Late Pennsylvanian coprolites from the Kinney Brick Quarry, central New Mexico, with notes on the classification and utility of coprolites; pp. 221–230 in J. Zidek (ed.), Geology and Paleontology of the Kinney Brick Quarry, Late Pennsylvanian, central New Mexico. Bulletin 138, New Mexico Bureau of Mines and Mineral Resources, Soccorro. Huxley, T. H. 1880. On the application of the laws of evolution to the arrangement of the Vertebrata and more particularly of the Mammalia. Proceedings of the Zoological Society of London 1880:649– 662. Janvier, P. 1996. Early Vertebrates. Oxford Science Publications, Oxford, 393 pp. Lebedev, O. A. 1996. Fish assemblages in the Tournaisian-Viseán environments of the East European Platform. Geological Society Special Publication 107:387–415. Lund, R. 1974. Stethacanthus altonensis (Elasmobranchii) from the Bear Gulch Limestone of Montana. Annals of the Carnegie Museum 45:161–178. ——— 1982. Harpagofututor volsellorhinus new genus and species (Chondrichthyes, Chondrenchelyiformes) from the Namurian Bear Gulch Limestone, Chondrenchelys problematica Traquair (Visean), and their sexual dimorphism. Journal of Paleontology 56:938–958. ——— 1984. On the spines of the Stethacanthidae (Chondrichthyes), with a description of a new genus from the Mississippian Bear Gulch Limestone. Geobios 17:281–295. ——— 1985a. Stethacanthid elasmobranch remains from the Bear Gulch Limestone (Namurian E2b) of Montana. American Museum Novitates 2828:1–24.

COATES AND SEQUEIRA—NEW SCOTTISH CHONDRICHTHYAN ——— 1985b. The morphology of Falcatus falcatus St. John & Worthen, a Mississippian stethacanthid chondrichthyan from the Bear Gulch Limestone of Montana. Journal of Vertebrate Paleontology 5:1–19. ——— 1986a. On Damocles serratus, nov. gen. et sp. (Elasmobranchii: Cladodontida) from the Upper Mississippian Bear Gulch Limestone of Montana. Journal of Vertebrate Paleontology 6:12–19. ———, and E. D. Grogan. 1997. Relationships of the Chimaeriformes and the basal radiation of the Chondrichthyes. Reviews in Fish Biology and Fisheries 7:65–123. Maddison, W. P., and D. R. Maddison. 1993. MacCLADE: Analysis of Phylogeny and Character Evolution. Version 3.0.5. Sinaur Associates, Sunderland, Massachusetts. Maisey, J. G. 1982. The anatomy and interrelationships of Mesozoic hybodont sharks. American Museum Novitates 2724:1–48. ——— 1983. Cranial anatomy of Hybodus basanus Egerton from the Lower Cretaceous of England. American Museum Novitates 2758: 1–64. ——— 1984. Chondrichthyan phylogeny: a look at the evidence. Journal of Vertebrate Paleontology 4:359–371. ——— 1989a. Visceral skeleton and musculature of a late Devonian shark. Journal of Vertebrate Paleontology 9:174–190. ——— 1989b. Hamiltonichthys mapesi, g. & sp. nov. (Chondrichthyes; Elasmobranchii), from the Upper Pennsylvanian of Kansas. American Museum Novitates 2931:1–42. ——— 2001. A primitive chondrichthyan braincase from the Middle Devonian of Bolivia; pp. 263–288 in P. E. Ahlberg (ed.), Major Events in Early Vertebrate Evolution: Palaeontology, Phylogeny and Development. Taylor and Francis, for the Systematics Association, London. Miles, R. S. 1968. Jaw articulation and suspension in Acanthodes and their significance. Nobel Symposium 4:109–127. ——— 1973. Relationships of acanthodians; pp. 63–103 in P. H. Greenwood, R. S. Miles and C. Patterson, (eds.), Zoological Journal of the Linnean Society 53, Supplement 1. Moy-Thomas, J. A. 1936. The structure and affinities of the fossil elasmobranch fishes from the Lower Carboniferous of Glencartholm, Eskdale. Proceedings of the Zoological Society of London B 1936: 762–788. Nelson, G. J. 1970. Pharyngeal Denticles (Placoid Scales) of Sharks, with Notes on the Dermal Skeleton of Vertebrates. American Museum Novitates 2415:1–26. Newberry, J. S. 1889. The Paleozoic fishes of North America. United States Geological Survey Monograph 16:1–340. Ørvig, T. 1977. A survey of odontodes (dermal teeth) from developmental, structural, functional, and phyletic points of view; pp. 53– 75 in S. M. Andrews, R. S. Miles, and A. D. Walker (eds.), Problems in Vertebrate Evolution. Academic Press, London. Patterson, C. 1965. The phylogeny of the chimaeroids. Philosophical Transactions of the Royal Society of London B 249:101–219. ——— 1982. Morphological characters and homology; pp. 21–74 in K. A. Joysey and A. E. Friday (eds.), Problems of Phylogenetic Reconstruction. Academic Press for the Systematics Association, London. Pleijel, F. 1995. On character coding for phylogeny reconstruction. Cladistics 11:309–315. Reif, W.-E. 1978. Types of morphogenesis of the dermal skeleton in fossil sharks. Palaöntologische Zeitschrift 52:235–257. ——— 1980. A model of morphogenetic processes in the dermal skeleton of elasmobranchs. Neues Jahrbuch fu¨r Geologie und Palaëontologie Abhandungen 159:339–359. ——— 1985. Squamation and Ecology of Sharks. Courier Forschungsinstitut Senckenberg 78:1–255. Ride, W. D. L., H. G. Cogger, C. Dupuis, O. Kraus, A. Minelli, F. C. Thompson, and P. K. Tubbs. 1999. International Code of Zoological Nomenclature. 4th ed. The International Trust for Zoological Nomenclature 1999, The Natural History Museum, London, 306 pp. Romer, A. S., and T. S. Parsons. 1977. The Vertebrate Body. 5th ed. W. B. Saunders Company, Philadelphia, 624 pp. Schaeffer, B. 1981. The xenacanth shark neurocranium, with comments on elasmobranch monophyly. Bulletin of the American Museum of Natural History 169:1–66. Schaumberg, G. 1982. Hopleacanthus richelsdorfensis n. g. n.sp., ein Euselachier aus dem permischen Kupferschiefer von Hessen (WDeutschland). Palaöntologische Zeitschrift 56:235–257.

457

Sequeira, S. E. K., and M. I. Coates. 2000. Reassessment of ‘Cladodus’ neilsoni Traquair: a primitive shark from the Carboniferous of East Kilbride, Scotland. Palaeontology 43:152–172. Sire, J.-Y., S. Marin, and X. Allizard. 1998. Comparison of Teeth and Dermal Denticles (Odontodes) in the Teleost Denticeps clupeoides (Clupeomorpha). Journal of Morphology 237:237–255. Smith, B. G. 1937. The anatomy of the frilled shark Chlamydoselachus anguineus Garman; Article VI, pp. 335–505 in E. W. Gudger (ed.), The Bashford Dean Memorial Volume Archaic Fishes. The American Museum of Natural History, New York. Stensio¨, E. A. 1937. Notes on the endocranium of a Devonian Cladodus. Bulletin of the Geological Institute of Uppsala 27:128–144. ——— 1963. Anatomical studies on the arthrodiran head. Part 1. Kungliga Svenska Vetenskapsakademiens Hanlingar 9:1–419. Swofford, D. L. 1993. Phylogenetic Analysis Using Parsimony, Version 3.1.1. Illinois Natural History Survey, Campaign, Illinois. ——— 1999. Phylogenetic Analysis Using Parsimony (and other methods). Version 4.0b3. Sinauer Associates, Sunderland, Massachusetts. St. John, O., and A. H. Worthen, 1875. Description of fossil fishes. Geological Survey of Illinois 6:245–488. Traquair, R. H. 1898. On Cladodus neilsoni (Traquair), from the Carboniferous Limestone of East Kilbride. Transactions of the Geological Society, Glasgow 11:41–50. Williams, M. E. 1985. The ‘‘cladodont level’’ sharks of the Pennsylvanian Black Shales of central North America. Palaeontographica 190:83–158. ——— 1990. Feeding behaviour in Cleveland Shale fishes; pp. 273– 290 in A. J. Boucot (ed.), Evolutionary Paleobiology of Behaviour and Coevolution. Elsevier, Amsterdam. Williams, M. E. 1998. A new specimen of Tamiobatis vetustus (Chondrichthyes, Ctenacanthoidea) from the Late Devonian Cleveland Shale of Ohio. Journal of Vertebrate Paleontology 18:251–260. Wood, S. P. 1982. New basal Namurian (Upper Carboniferous) fishes and crustaceans found near Glasgow. Nature 297:574–577. Young, G. C. 1982. Devonian sharks from south-eastern Australia and Antarctica. Palaeontology 25:817–843. Zangerl, R. 1981. Chondrichthyes 1: Paleozoic Elasmobranchii. H. P. Schultze (ed.), Handbook of Paleoichthyology, 3A. Gustav Fischer Verlag, Stuttgart, New York, 115 pp. ——— 1984. On the microscopic anatomy and possible function of the spine-‘‘brush’’ complex of Stethacanthus (Elasmobranchii: Symmoriida). Journal of Vertebrate Paleontology 4:372–378. ——— 1990. Two new stethacanthid sharks (Stethacanthidae, Symmoriida) from the Pennsylvanian of Indiana, USA. Palaeontographica A 213:115–141. ———, and G. R. Case. 1976. Cobelodus aculeatus (Cope), an anacanthous shark from Pennsylvanian Black Shales of North America. Palaeontographica A 154:107–57. Zidek, J. 1992. Late Pennsylvanian Chondrichthyes, Acanthodii, and deep-bodied Actinopterygii from the Kinney Quarrey, Manzanita Mountains, New Mexico; pp. 145–182 in J. Zidek (ed.), Geology and Paleontology of the Kinney Brick Quarry, Late Pennsylvanian, central New Mexico. Bulletin 138, New Mexico Bureau of Mines and Mineral Resources, Soccorro. ——— 1993. A large stethacanthid shark (Elasmobranchii, Symmoriida) from the Mississippian of Oklahoma. Oklahoma Geology Notes 53:4–15. Received 10 June 2000; accepted 4 January 2001.

APPENDIX 1 Characters and Character-States Used in the Phylogenetic Analysis 1. 2. 3. 4.

Prismatic cartilage. Absent (0); present (1). Body mostly scale-less. Absent (0); present (1). Lateral line passes through scales. Absent (0); present (1). Ring or C-shaped scales enclosing sensory canals. Absent (0); present (1). 5. Dorsal fin spine consists mostly of vascularised osteodentine, with no clear division into two or three discrete layers as in most elasmobranch examples. Absent (0); present (1). 6. Physonemid spine shape (anteriorly concave profile in lateral view;

458

7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37.

38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.


posterior opening proximodistally extensive). Absent (0); present (1). Spine precedes dorsal fins. Absent (0); present (1). Spine present in only anteriormost dorsal fin position; pelvic level dorsal fin spineless. Absent (0); present (1). Cephalic spines. Absent (0); present (1). Two dorsal fins. Absent (0); present (1). Second or single dorsal fin opposite pelvis. Absent (0); present (1). Single dorsal fin elongate, extending from near-pectoral level to anal or pre-anal fin level. Absent (0); present (1). Anterior dorsal fin base plate calcified. Absent (0); present (1). Posterior dorsal fin baseplate calcified. Absent (0); present (1). Posterior dorsal fin with delta-shaped cartilage. Absent (0); present (1). Anal fin. Absent (0); present (1). Anal fin supported by radials external to body wall. Absent (0); present (1). Anal fin with double base-plate. Absent (0); present (1). Caudal axis upturned steeply, supporting high aspect ratio (lunate) heterocercal tail; elongate hypochordal radials unsegmented or segmented only proximally. Absent (0); present (1). Horizontal or near-horizontal notochordal caudal extremity. Absent (0); present (1). Caudal neural and/or supraneural spines extended. Absent (0); present (1). Ventral lobe of tail supported by hypochordal radials extending beyond level of body wall. Absent (0); present (1). Scapular blade. Absent (0); present (1). Scapular anterodorsal and posterodorsal processes. Absent (0); present (1). Procoracoid directed posteriorly. Absent (0); present (1). Supra-articular pectoral foramina: numerous (0); single or absent (1). Pectoral articular surface: narrow/short (stenobasal) (0); broad/long (eurybasal) (1). Pectoral articular surface (and fin) elevated. Absent (0); present (1). Paired fin radials barely extend beyond level of body wall. Absent (0); present (1). Dibasal pectoral fin endoskeleton. Absent (0); present (1). Tribasal pectoral fin endoskeleton. Absent (0); present (1). Anteriormost proximal radial (propterygium) broad. Absent (0); present (1). Anteriormost distal pectoral radial largest of series. Absent (0); present (1). Middle of three proximal radials (mesopterygium) articulates with 31 distal radials. Absent (0); present (1). Posteriormost radial (metapterygium) broad and articulates directly with 51 distal radials. Absent (0); present (1). Proximal articular facet of metapterygium directed anteriorly. Absent (0); present (1). Metapterygium connects with distinct series of distal radials articulating proximodistally to form a short or long axis. Absent (0); present (1). Axial radials articulate with pre- and post-axial radials. Absent (0); present (1). Pelvic plate semicircular with anterolateral concavity. Absent (0); present (1). Pelvic girdle with narrow anteromedial process and single diazonal foramen. Absent (0); present (1). Fused puboischiadic bar. Absent (0); present (1). Anteriormost pelvic radial broader than posterior, pre-metapterygial members of series. Absent (0); present (1). Four or fewer radials articulate directly with pelvic girdle. Absent (0); present (1). Claspers in males. Absent (0); present (1). Myxopterygial claspers. Absent (0); present (1). Claspers with clawed terminus. Absent (0); present (1). Claspers with ‘toothed’ plate. Absent (0); present (1). Fused anterior vertebral arches (synarcual) at anchorage of dorsal spine. Absent (0); present (1). Calcified ribs. Absent (0); present (1). Calcified perichordal ‘centra’ or rings. Absent (0); present (1). Palatoquadrate otic process expanded with an anterodorsal angle. Absent (0); present (1).

52. Laterally directed otic articular fossa on palatoquadrate. Absent (0); present (1). 53. Quadrate condyle. Absent (0); present (1). 54. Double condylar-glenoid mandibular joint. Absent (0); present (1). 55. Labial cartilages. Absent (0); present (1). 56. Holostyly. Absent (0); present (1). 57. Hyoid rays elongate, supporting opercular flap. Absent (0); present (1). 58. Hyomandibula semicrescent shaped. Absent (0); present (1). 59. Interhyal. Absent (0); present (1). 60. Hypobranchial orientation. Medially or anteriorly (0); largely posteriorly (1). 61. Basibranchials 1 and 2. In contact or close apposition (0); separated by a gap (1). 62. Large posteriorly projecting basibranchial copula. Absent (0); present (1). 63. Persistent metotic/otico-occipital fissure. Absent (0); present (1). 64. Ventral cranial fissure. Absent (0); present (1). 65. Glossopharyngeal nerve exits through metotic fissure. Absent (0); present (1). 66. Glossopharyngeal nerve foramen situated posteroventral to otic capsule and anterior to metotic fissure. Absent (0); present (1). 67. Glossopharyngeal nerve foramen exits dorsally, posterior to otic capsule. Absent (0); present (1). 68. Canal for lateral dorsal aorta within basicranial cartilage Absent (0); present (1). 69. Canal for lateral dorsal aorta long. Absent (0); present (1). 70. Otico-occipital proportions: greater than length of ethmo-orbital portion (0); equal to or less than ethmo-orbital portion (1). 71. Posterior tectum. Absent (0); present (1). 72. Occipital unit wedged between rear of otic capsules. Absent (0); present (1). 73. Endolymphatic ducts directed dorsally and fossae close to midline. Absent (0); present (1). 74. Endolymphatic ducts exit into slot-shaped median fossa. Absent (0); present (1). 75. Dorsal ridge posterior grades smoothly into occipital roof, with no horizontal crests. Absent (0); present (1). 76. Hyoid articular area on posterolateral angle of otic capsule. Absent (0); present (1). 77. Prominent lateral otic process. Absent (0); present (1). 78. Lateral commissure expanded anteroposteriorly. Absent (0); present (1). 79. Articulation for palatoquadrate on rear of postorbital process. Absent (0); present (1). 80. Positions of foramina for nerves II, III and IV in orbit. Nerve foramen II situated mid-orbit; III and IV posterior to II (0); nerve foramen II situated mid-orbit, IV anterior to II (1). 81. Myodome for superior oblique muscle situated anterodorsally. Absent (0); present (1). 82. Broad suborbital shelf. Absent (0); present (1). 83. Suborbital shelf expanded anterolaterally. Absent (0); present (1). 84. Palatobasal process. Absent (0); present (1). 85. Antorbital process. Absent (0); present (1). 86. Precerebral fontanelle. Absent (0); present (1).

APPENDIX 2 Data Matrix of Early Chondrichthyans and Outgroups Character-state distributions for taxa obtained from original observations and/or the following literature sources: Norselaspis, Janvier, 1996; Kujdanowiaspis, Stensio¨, 1963, Goujet, 1984; Acanthodes, Miles, 1968, 1973, Denison, 1979, Coates, 1994; Mimia, Gardiner, 1984; Diplodoselache, Dick, 1981; Orthacanthus, Hotton, 1952, Schaeffer, 1981, Heidtke, 1982; ctenacanth (Tamiobatis), Moy-Thomas, 1936, Zangerl, 1981, Williams, 1998; ‘‘C.’’ wildungensis, ‘‘C.’’ hassiacus, Stensio¨, 1937, Gross, 1937, Schaeffer, 1981; Hopleacanthus, Schaumberg, 1982; Tristychius, Dick, 1978; Onychoselache, Dick and Maisey, 1981; Hamiltonichthys, Maisey, 1989b; Hybodus, Maisey, 1982, 1983; Cladoselache, Harris, 1938, Zangerl, 1981, Maisey, 1989a, Williams, 1998; Cobelodus, Zangerl and Case, 1976; Denaea, Zangerl, 1981, Williams, 1985; Stethacanthus, Zangerl, 1981, 1984, Coates and Sequeira, 1998; Falcatus, Lund, 1985b; Damocles, Lund, 1986a; Harpagofututor, Lund, 1982; Helodus, Patterson, 1965; Ischyodus, Patterson, 1965.


459

APPENDIX 3 Character-taxon Matrix used in Phylogenetic Analysis. Characters Taxon Norselaspis Kujdanowias. Acanthodes Mimia Cladoselache Cobelodus Denaea Akmonistion Falcatus Damocles Harpagofututor Helodus Ischyodus Ctenacanth ‘‘C.’’ wild. ‘‘C.’’ hass. Hopleacanthus Tristychius Onychoselache Hamilton. Hybodus. Diplodosel. Orthacanthus

1

1 0

2 0

3 0

4 0

5 0

6 0

7 0

8 0

8 6

00000000000000000000000000000????????0000000000?0000000?000???00000000?00000000?00000? 0000?0000000000000000000000000?000000000000100000000100000?000000000000000000100010000 00100010000000010000011001001010000000?????0000000101100?110001101000100??00001??0010? 00100010000000010000000001001000000000000000000000101100001000110100010000000000100100 100010110110000000101111011000000010100000000000001100?011000??????111?????1??1????00? 110000010010001000100111111000000011101000011100001111?0?1000010100111?????1001??0000? 1100000100100010001001111?10000000111010000111000011???0?1????10100??1??100??01?10000? 1101111101101010001011110110000000111000000110000111???011000010100111101001001?10000? 11011111011010100001111101000000001110000001100001111??011000000000111????00001?100001 110111110010101000011111010100000011??000011101001111??0?1????00000??1?????0001?1???0? 1101000100?000000001111101010000101110000011101001001??11?????00000111???000000?1???00 100110110110100000010?100100010010110000001110?10?001??1??????00000001????0000001???00 1101101101101000000111110100010010110000001110?101001??1??????00000001????00000?1???00 10000010011011011000?11?0??0001000111100000110000010???0??????101001101111011110111001 1?????????????????????????????????????????????????101??0??????10100110??110111101110?1 1?????????????????????????????????????????????????101??0???????0??0111?11??111101110?? 10000010?110???11?000?1001000010001?100?0?11100?01??1??0??????00000??1?????1111??10011 10000010011011011000001101000011010011010111100011001?1010011100000001?111101001110011 1000001011101101100001100100001101101001011110001000???00?????00000??1????11110?????11 100000101110?1011100001001000011011000011111100010001??000011100001101011111110??10011 1000001011101101110001100100001101100000110110001000111000011100001101?111111101110011 1000101?00?10001100011110100001000111?0000011000001011?0?00???????????????????1???0??? 1000101?00?1000110011111010000100001110000011010101011?0100??0101001101111011111110001