THE ACANTHODIAN FAUNA OF THE CRAVEN PEAKS BEDS (EARLY TO MIDDLE DEVONIAN), WESTERN QUEENSLAND CAROLE J. BURROW & GAVIN C. YOUNG Burrow, C.J. & Young, G.C. 2005 05 31. The acanthodian fauna of the Craven Peaks Beds (Early to Middle Devonian), western Queensland. Memoirs of the Queensland Museum 51(1): 3-25. Brisbane. ISSN 0079-8835. Two acanthodian faunas of different ages have been identified from calcareous lithologies assigned to the Cravens Peak Beds, Georgina Basin, western Queensland. A sparse fauna comprising scales of Nostolepis sp. cf. N. striata and Radioporacanthodes sp. indicates a Lochkovian or Pragian age for one shot-hole sample south of the Toko Range. A limestone outcrop further south in the Toomba Range yielded a more abundant fauna which includes dissociated remains of two new acanthodians Teneracanthus toombaensis gen. et sp. nov. and Machaeracanthus pectinatus sp. nov. Teneracanthus gen. nov. is an acanthodid acanthodian which is most similar to the early Frasnian Lodeacanthus from Latvia. M. pectinatus sp. nov. is based on isolated scales. Comparison with acanthodian and thelodontid microremains from well-dated sequences in the Broken River Group indicate a late Emsian to early Eifelian age for the limestone stratum. (cid:9633) Acanthodian, Mesacanthidae, Teneracanthus gen. nov., Machaeracanthidae fam. nov., Machaeracanthus, Cravens Peak Beds, Georgina Basin, Devonian. Carole J. Burrow, School of Integrative Biology, University of Queensland, St Lucia 4072, Australia [[email protected]]; Gavin C. Young, Dept, of Earth and Marine Sciences, Australian National University, Canberra, ACT, 0200, Australia; 1 August 2004. Devonian fish remains have been known from timeslice of the Devonian by Struckmeyer & the Amadeus and Georgina Basins of central Totterdell (1990: 34). Australia for nearly 50 years (Hills, 1959; Further collecting (GCY, 1977) revealed Gilbert-Tomlinson, 1968; Young, 1985, 1988). exceptionally well preserved but extremely Central Australian Devonian rocks are fragile vertebrate remains, which could be predominantly sandstones and siltstones which, extracted by acetic acid digestion. Based on this in the absence of marine invertebrate fossils, additional material. Young (1984) described have been interpreted as fluvial, lacustrine, or some pterichthyodid antiarch placoderm bones, aeolian deposits that were laid down at the initial at the time considered to be one of the oldest emergence of the central Australian landmass records of this group. Turner & Young (1987) from the ocean. The only known exception is a described enigmatic chondrichthyan teeth of the small limestone outcrop, previously assigned to shark Mcmurdodus whitei, which display the basal calcareous unit of the Cravens Peak advanced features and are comparable to those of Beds in the Georgina Basin of western the living hexanchid sharks. Mcmurdodus was Queensland. This limestone was first sampled for erected by White (1968) for a tooth from the microfossils to confirm its assumed Early Aztec Siltstone of Victoria Land, Antarctica, Ordovician age, but instead was found to contain which was compared with teeth of the hexanchid an assemblage of Devonian fish remains (Draper, Notidanus. The fossil record of hexanchids 1976). Various scales of thelodont agnathans and otherwise extends back only to the Jurassic, so if acanthodians, associated with eridostracans and correctly assigned, the Devonian Mcmurdodus is ostracods, were described by Turner et al. (1981) by far the oldest known neoselachian (Cappetta et from this limestone, and from five other (seismic al., 1994). Turner (1995) redescribed the thel- shot-point) localities in the Toko Range area, odontids, erecting Turinia gavinyoungi, and about 40 km to the north (Fig. 1). They assigned listed an associated fauna of shark scales, the fauna a probable Emsian age, and associated onychodont teeth, sarcopterygian scales, lepid- eridostracans and ostracods were used to infer a otrichia, and acanthodian remains including probable shallow marine environment of climatiid spines and Acanthodes-type and deposition. On this evidence, an elongate marine Machaeracanthus scales. incursion from the south was included on the Turner (1991) and Young (1993) noted micro- palaeogeographic map for the Emsian-Eifelian vertebrate assemblages described by Turner et al. 4 MEMOIRS OF THE QUEENSLAND MUSEUM C SECTION 1 southern Toomba Range metres ?BCU CZ> 7GYS-11 280 -- 7FAULT LEGEND FOR B: I [indet.) 1 1 Early Palaeozoic —J— syncline ----FAULT Devonian owaterhole,bore,lake,soak etc. _ 240- D Mithaka W.H A! --FAULT 10 km a 200- 160- SECTION 2 Eurlthethera Soak SECTION 3 120- Toomba Bore if \ SP798 [Reference section] 80- 40- 0-1 ” — — — 7FAULT 7BCU 7GY6-11 bio-i r r i ^^ (=GE065/28)^\ $*(cid:9632) FIG. 1 A, Georgina Basin (GB) northern Australia; location of B indicated by arrow. B, S part of the Toko Syncline, geology from Hay River-Mt Whelan Special 1:250 000 sheet (Shergold, 1985), location of GY fossil localities collected 1977, and approx, localities for shot-point samples (prefix SP). GY24; locality for indeterminate spine (Young & Goujet, 2003). C, 3 measured sections through Cravens Peak Beds, modified from Draper (1976: fig. 4), and Turner et al. (1981: fig. 3): Section 1, S Toomba Range, about 100m S of GY6-11 (Fig. IB), Section 2, Eurithethera Soak (see text for stratigraphic; Section 3, reference section of Smith (1972) as modified by Draper (1976), about 100m E of Toomba Bore. DEVONIAN ACANTHODIANS 5 (1981) contained conflicting evidence of age, and Toomba Fault’. He originally considered it to be Young (1995: 20) discussed the possibility that probably unconformable on the Ordovician the assemblage could have mixed material from Mithaka Formation and to be unconformably older (shot-point samples) and younger (outcrop) overlain by sandstone and conglomerate of the horizons. Young (1996: 103) considered the basal Cravens Peak Beds, but later reinterpreted this to calcareous unit of the Cravens Peak Beds to be a conformable contact (Turner et al., 1981). A include a component younger (Eifelian) than the 280m thick measured section of Cravens Peak Early Devonian age assigned by Turner et al. Beds situated 100m south of the limestone (1981). Young & Turner (2000: fig. 4) gave alter¬ outcrop begins with a basal pebbly unit (Draper, native late Pragian-Emsian, or late Emsian- 1976: fig. 4, section 1), and the limestone occurs Eifelian, ages for the assemblage. Turner (1995, in gullies only about 10m from similar strata 1997) compared thelodontids from the outcrop assumed to be the base of the Cravens Peak Beds with younger (Middle Devonian) turiniid scales (see Young & Turner, 2000: 464). However, an from the Broken River Province, Queensland and alternative interpretation is that the whole the Aztec sequence of Victoria Land, Antarctica. sequence is overturned, i.e. younging to SW, not Turner et al. (1981) originally assigned the NE. At least two faults were identified higher up acanthodian fish scales to two taxa (Nostolepis in Draper’s measured section (beneath the 220 sp. and Gomphonchus sp.), but examination of and 260m levels; Draper, 1976: fig. 4), but these other remains, including spines and ossified were omitted from the published section (Turner scapulocoracoids, suggested a younger element et al., 1981: fig. 3). The faulted contact noted in in the fauna (Burrow, 2002). We suggest that at the original field assessment places uncertainty least some of the samples from west of the Toko on the stratigraphic relationship between the two Range are of ?late Lochkovian to early Pragian lithologies. Field observations (GCY, 1977) age, and the samples from the Toomba Range noted a nearby ‘basal’ conglomerate of the might be late Emsian or Eifelian. Cravens Peak Beds, but there are conglomeratic beds at many levels within the measured sections, FIELD OCCURRENCE so this evidence is not compelling. In addition, the supposed contact between elastics and Turner et al. (1981) dealt with material from limestone is parallel to mapped faults in the five shot-point samples, but only one sample vicinity, consistent with it being an unrecognised from the limestone outcrop. Shot-point samples, fault, the (assumed) top of section 1 of Draper obtained during a 1963 seismic survey of the (1976) finishes in non-exposure; the only fossils Toko Range for Phillips Petroleum, and first are one poorly preserved fish sample showing a recorded by Jones (in Reynolds & Pritchard, faint impression of tuberculate ornament, and a 1964), were calcareous marl, processed by possible arthropod impression, collected within washing, not acid digestion. They included some 20m of the (assumed) highest exposure. Ordovician conodonts (probably reworked, or However, in the reference section for the Cravens perhaps due to contamination; Turner et al., Peak Beds at Toomba Bore (Draper, 1976: fig. 4, 1981: 53). This is circumstantial evidence that section 3), the only fossil (a plate impression the shot-point localities were sampling the basal from the placoderm Wuttagoonaspis sp.) occurs contact between the Devonian sequence and at the base. Many localities yielding a diverse fish underlying Early Palaeozoic (P.J. Jones, pers. assemblage including Wuttagoonaspis sp. comm.). In contrast, the limestone is a solid rock, (Young & Goujet, 2003) also occur in the lower only broken down by normal acid digestion part of the sandstone sequence. Field work is techniques, so there is a lithological difference required to re-examine the outcrop relationships between the two sources. between Devonian limestone and sandstone in According to field data for the Georgina Basin this area, but the geological information just Project held at Geoscience Australia in Canberra, summarised provides circumstantial evidence the original limestone sample (74710577), consistent with recognition of different assumed initially to be from the Early Ordovician acanthodian faunas in the limestone and the Coolibah Formation, was collected by J.J. Draper shot-point samples. Two alternative stratigraphic on 8 August 1975 from an outcrop which ‘occurs positions for the ‘basal calcareous unit’ (BCU) along a fault’. Draper (1976: 3) stated that the are indicated in Section 1 of Fig. 1C. outcrop occurred ‘at the base of the scarp at the SE end of the Toomba Range; this outcrop has an Of eight taxa originally documented from the area of about 5m2 along a spur fault of the ‘basal calcareous unit’, five were listed by Turner 6 MEMOIRS OF THE QUEENSLAND MUSEUM et al. (1981: Fig. 4) as common to the limestone collecting; L, Natural History Museum and shot-point samples, but two thelodont taxa collection, Prague; SP, shot-point localities; from the former were reassigned to a single new UQL, University of Queensland Earth Sciences species by Turner (1995). The best preserved locality; UQY, University of Queensland Earth ostracods (.Healdianella and Bashkirinal) come Sciences (Queensland Museum) collection. from the shot-point samples, and the eridostracan SYSTEMATIC PALAEONTOLOGY Cryptophyllus is best represented only from the limestone sample (Turner et al., 1981: fig. 15). Class ACANTHODII Given the possibility that different levels may have been sampled, the poorly preserved REMARKS. Nostolepsis has usually been unfigured remains of these crustacean taxa are in assigned to the Climatiiformes Berg, 1940, now need of restudy (P.J. Jones, pers. comm.). considered to be a paraphyletic group (Janvier, Young & Goujet (2003) concluded that the 1996). Also, scales with Nostolepis-type Wuttagoonaspis fauna from the Cravens Peak histology have been described on articulated fish Beds and lower Dulcie Sandstone in the Georgina with dentigerous jaw bones (Valiukevicius, Basin was probably no older than Pragian, and no 2003), indicating they should be assigned to the younger than early Eifelian, consistent with the Ischnacanthiformes. These new taxa thus also suggested stratigraphic position of the limestone. throw doubt on the familial assignment of Whether this Wuttagoonaspis fauna is younger or Nostolepis, which was erected for isolated scales. older than the W.Jletcheri Ritchie, 1973 fauna in For these reasons, we have not assigned the type area in the Darling Basin, NSW (Mulga Nostolepis to an order or family. Downs Group) must await detailed analysis of Nostolepis Pander, 1856 faunal associations from many known but unstudied fish localities. The Wuttagoonaspis TYPE SPECIES. Nostolepis striata Pander, 1856. assemblage in the Darling Basin is recorded from Nostolepis sp. cf. N. striata many sites in the vicinity of the type locality (Fig. 2A,B) (Wuttagoona Station, about 60km NW of Cobar), plus some 150km to the S (Glen et al., 1987; MATERIAL. One scale CPC20088/3 from SP799 Young, pers. obs.), and some 360km to the W, in (Ipesavis/sulcatus CZ. ?late Lochkovian/early Pragian). the Barrier Ranges north of Broken Hill (Coco DESCRIPTION. Light amber coloured, about Range Formation; Neef et al., 1995). Correlation 1.0mm long, with about four crown ridges rising of these widespread localities is uncertain up from the short, slightly-inclined neck without detailed systematic work, but we note a anteriorly. Main central plane of crown range of undescribed acanthodian material from horizontal, smooth behind ridges; a narrow, various localities and cores in the Darling Basin lower ledge is preserved on one side. Scale base which represent at least several horizons. bowl-shaped, moderately swollen. Posterior OTHER LOCALITIES crown and base have been broken off. UQL4697 = BRJ103A. Section ~70m west of old road REMARKS. The scale is one that Turner et al. crossing of Digger's Creek; stratigraphically below Fish (1981: 60) referred to as “scales ... resemble Hill limestones; Bracteata Formation, ?late Emsian-early Nostolepis striata”. Based on morphology, the Eifelian. scale is comparable to nostolepid scales also UQL4704=BRJ104B. Mid-level in section/traverse along assigned to Nostolepis sp. cf. N. striata from the limestone outcrop from Digger’s Creek Crossing to road; Martins Well Limestone (Ipesavis/sulcatus CZ) GR 683 489 Burges 1:100000 sheet; Bracteata Formation, of the Broken River region, north Queensland late Emsian, probably serotinus or patulus CZ. (Turner et al., 2000; Burrow, 2002). BRJ133K. Jessey Springs; Chinaman Creek Limestone, late Emsian/Eifelian. Order ISCHNACANTHIFORMES Berg, 1940 Family PORACANTHODIDAE Vergoossen, Abbreviations. ANUV, Gavin Young collection, 1997 Australian National University; BMR, Bureau of Mineral Resources, now Geoscience Australia; Radioporacanthodes Vergoossen, 1999 BRJ, J. Jell Broken River Formation collection; Radioporacanthodes sp. CPC, Commonwealth Palaeontological (Fig. 2C,D) Collection, Canberra; CZ, Conodont Zone/s; GY, MATERIAL. Two scales CPC20088/1 (Fig. 2C; figured as Gavin Young sample localities from 1977 Nostolepis sp. in Turner et al., 1981: fig. 14D) and DEVONIAN ACANTHODIANS 7 CPC20088/2 (Fig. 2D), from SP799 (Ipesavis/sulcatus, ?late Lochkovian/early Pragian). DESCRIPTION. Scales amber-coloured; CPC20088/1 1.0mm wide; CPC20088/2 0.7mm wide; crowns flat, horizontal. CPC20088/2 with about eight short weakly-developed ridges along the anterior crown margin; most of this area on CPC20088/1 is broken off. Posterior crowns broken off on both scales. Necks deep, concave and about the same depth all round. Short vertical slits score neck just above the base/neck junction anteriorly and high on the neck posteriorly. Bases strongly convex forward of centre, protruding markedly in front of crown, tapering towards the posterior comer. REMARKS. Although diagnostically important posterior crown regions have broken off the FIG. 2. Acanthodian scales from SP799. A-B. scales, they compare closely in all other respects Nostolepis sp. cf. N. striata scale CPC 20088/3. with Radioporacanthodes sp. from Martins Well A, crown view; B, lateral view. C-D. Limestone (Ipesavis/sulcatus CZ; Turner et al., Radioporacanthodes sp. scales. C, CPC 2000; Burrow, 2002). Some Lochkovian scales 20088/2 crown view; D, CPC 20088/3 crown of Gomphonchoporus hoppei (Gross), 1971 and view. Anterior is to left, scale bar = 0.2mm. R. porosus (Brotzen), 1934, and scales from the Pridoli and Lochkovian of the East Baltic and but also have synapomorphies including Byelorussia which Valiukevicius (1998: pi. 7.4) blade-like hyoidean gill covers and a mandibular assigned to “Gomphonchus sandelensis or splint (Hanke & Wilson, 2004: app. 1,2). All Poracanthodes punctatus'\ are also similar. cheiracanthids have ornamented scales and lack intermediate spines; and all acanthodids have Order ACANTHODIFORMES Berg, 1937 smooth scales, erectile pectoral spines, either unpaired or no pelvic spines and no intermediate REMARKS. Berg (1940) listed seven orders spines. Zajic (1995) redefined Acanthodidae to within the Acanthodii: Climatiiformes, exclude Howittacanthus, and erected Ischnacanthi formes, Gyracanthiformes, Howittacanthidae, diagnosed by paired pelvic Dip lac ant hi formes, A c ant hodi formes, spines, smooth scales and lack of intermediate Cheiracanthiformes, and Mesacanthiformes. spines. However, paired pelvic spines is a Miles (1966) and Novitskaya & Obruchev (1967) plesiomorphy, and the Acanthodidae also have demoted the latter three monofamilial groups to smooth scales and lack intermediate spines. families within one order. Presumably to avoid confusion with Berg’s groups, Novitskaya & Family MESACANTHIDAE Moy-Thomas, Obruchev (1967) used the ‘-ida’ suffix rather 1939 than ‘-iformes’ for acanthodian orders, but present consensus favours reverting to the Teneracanthus gen. nov. ‘-iformes’ suffix. Acanthodiformes thus TYPE SPECIES. Teneracanthus toombaensis sp. nov. comprises Mesacanthidae, Cheiracanthidae and Acanthodidae. According to Denison (1979), this ETYMOLOGY. Latin: tener, delicate; and the common suffix used for acanthodian taxa: acanthus, thorn or spine. familial division was based on gradual transition from ‘primitive’ to specialised states, without DIAGNOSIS. Acanthodiform acanthodian; clearcut diagnostic characters. Denison placed scapulocoracoid with a long, slender scapular all relevant species in Acanthodidae, because of shaft having a circular cross-section, a lateral this perceived lack of clearcut boundaries wide-based, triangular scapular blade which is between the groups. However, some characters about one-quarter the height of the scapular shaft, of the families (cf. Miles, 1966) are not and a medial coracoid blade of similar shape transitional. Mesacanthids have one pair of which diverges at ca. 45° to the axis of the shaft; intermediate spines and smooth scales, scapulocoracoid articulates with the trailing and characters which are plesiomorphic for the order, lateral sides of the proximal end of the pectoral 8 MEMOIRS OF THE QUEENSLAND MUSEUM TABLE 1. Comparison between five mesacanthid taxa. Lodeacanthus Mesacanthus Melanoacanthus Teneracanthus Triazeugacanthus toombaensis gen. et gaujicus mitchelli minutus affinis sp. nov. longitudinal ribs on fin deep anterior groove, deep anterior groove, deep anterior groove, spines fine post, grooves fine post, grooves none some with fin post, 2-3 shallow grooves grooves juveniles- distal half of leading edge; distal half of leading denticulations on fin spine adults- none except none none edge on pectoral none on intermediate spines spines plornogxiitmudali npaelc ctolerfatl, spine yes no ? yes 7 fin spine insert base short short short short short scapulocoracoid scapulocoracoid spceacptourlaolc foirna cspoiinde/ blades dorsal and pectoral spine lateral ? blades dorsomedial blsacdaep u?dloocrosroalcaoteirda l articulation lateral stpoi npee ctoral to coracoid blade paencdt olaratel rsapl intoe to pectoral fin spine cscroaspsu-lsaerc stihoanfat l shape of jaudvueltnsi-l ecsir-c fulalat;r circular ? circular circular height of expanded scapulocoracoid blade: total scapulocoracoid 1:5 2:3 1:2 1:5 1:4 height shape of scapulocoracoid blade short-based triangle ? triangle short-based triangle long-based triangle short-based triangle procoracoid none none none none none palatoquadratc single growth centre; two growth centres; ? ? 7 fenestrated no fenestra mid-sized, curved, branchiostegal rays outer face with long, ornamented long ? short, some forked single fine longitud. and some pointed ridge ceratotrichia none ?none ?none ? round, forked tectal tesserae edgsinengs ionrfyr aloinrbe ital cover head cover head ?head alninde s ensory cover head size of flank scales 7/mm 16/mm ‘small’ 4-5/mm 5/mm scale crown smooth smooth smooth smooth smooth scale base juvaednuilltess -- c coonniccaalv e; rounded convex ‘tumid’ pyramidal relatively flat scale histology ? opruthlpo dcaennatilnse, Acanthodes-iype? Ancoa ninthteordceosn-ntyepceti nbgu t 7 network fin spine, astride its longitudinal cleft; small square crown, concave neck and a base shaped rounded 'glenoid’ process posteromedial to base like an inverted pyramid, and probable caudal of scapular shaft; no procoracoid; pectoral fin scales with an elongated crown, negligible neck, spines bear fine denticulations on the distal half and a shallow, diamond-shaped base; scales have of the leading edge; fin spines are slightly curved, Acanthodes-type histology but lack a fine with one deep longitudinal groove separating the network of processes between centripetal dentine rounded ridge which forms the leading edge from tubules in the crown. the main body of the spine; the sides of some non-denticulated (?median) spines bear fine Teneracanthus toombaensis gen. et sp. nov. longitudinal ridges proximally; some (Figs 3A-E,4A-T,5A-K,6D-F; Table 1) symmetrical fin spines with basal cartilages; fin \9%\'Gomphonchus? sp.' Turner et al.: fig. 14A-C spines have a relatively wide central pulp cavity, 1981 ‘acanthodian spines’ Turner et al.: 60 a network of fine dentine tubules leading into 1987‘smooth ischnacanthid scales' Young et al.: 239 vascular canals which run more or less longi¬ 1995‘climatiid spines' Turner: 683 1995‘Acanthodes-typc scales' Turner: 683 tudinally; tri-basal pectoral fin; smooth-crowned 2000‘new mesacanthid' Young & Turner: 464 scales of two types: normal flank scales with a 2003‘Acanthodian remains’, in part Young & Goujet: 9 DEVONIAN ACANTHODIANS 9 ETYMOLOGY. For the Toomba Range. Denticulated fin spines (Fig. 4H,L-Q). ANUV 2939.10 (Fig. 4N) is an almost complete, but MATERIAL. HOLOTYPE: ANUV2940 (Figs fragmented, pectoral fin spine, 11mm long, 3A-C,4R-S), pectoral fin spine attached to a curving slightly, and is from the left side; the scapulocoracoid; Locality GY 11, Toomba Range, western flank of Toko Syncline, Georgina Basin, western main shaft has a circular cross-section and Queensland (Fig. 1A,B). PARATYPES: left smooth sides, with a longitudinal ridge along the scapulocoracoid ANUV2969.2 (Figs 3D-F,4C-D), leading edge. The distal half to two-thirds of this pectoral fin spine ANUV2969.1 (Fig. 4L-M), fin spine ridge bears fine backward-pointing ANU V2969.5 (Fig. 4K). fin spine ANUV2969.6 (Fig. denticulations. The proximal end of the spine has 4P-Q). ADDITIONALMATERIAL: One right a thin film of ossified cartilage across the scapulocoracoid articulated with pectoral fin spine, one left triangular cleft between the two sides of the pectoral fin spine, 13 fin spine fragments, one ground thin spine. This spine possibly articulated with the left section from fin spine fragment, one left and one right scapulocoracoid described above. ANU- scapulocoracoid, two scapulocoracoid shaft fragments, one ?pharyngobranchial, and 47 scales in samples ANU V2939.11 is a right pectoral fin spine, 11 mm long V2937-41 from GY11; one left and one right but missing the proximal part; it is otherwise scapulocoracoid, fin spines, branchial elements, scales, and comparable to the left pectoral fin spine, having a a left palatoquadrate in ANUV2969 from GY 11; circular cross-section and distal denticulations. seventeen scales including CPC20087/1 and CPC20087/2 These pectoral fin spines are not exactly and two ground thin sections, in BMR sample 74710577 symmetrical dorso-ventrally, having a smooth from locality GEO 65/28 (Turner et al., 1981); all from the transition from the leading edge ridge to groove Cravens Peak Beds. One scale in BRJ103A and 16 scales to main body of spine on one side, but with an in BRJ104B (Fig. 5K). Bractcata Formation, and ?one abrupt, ‘stepped-down’ transition from the ridge scale in BRJ133K, Chinaman Creek Limestone, Broken River Group. to the main bodv of the spine on the other side. Paratype ANU V2969.6 (Fig. 4P-Q) is a DIAGNOSIS. As for genus. complete, laterally flattened fin spine 14mm long, also showing slight longitudinal curvature. GEOGRAPHICAL & STRATIGRAPHIC The pulp cavity is open along the proximal half of DISTRIBUTION (AUSTRALIA). Cravens Peak the trailing edge, and the distal half of the leading Beds (?Iate Emsian/early Eifelian), Georgina Basin, edge is denticulated. Other fin spine material in western Queensland (Fig. 1A-C); Bracteata the samples includes short fragments of Formation and possibly Chinaman Creek denticulated spines, including one in ANUV2941 Limestone (late Emsian/early Eifelian), Broken and three in ANUV2939 (Fig. 4H). Paratype River Group, north Queensland. ANUV2969.1 (Fig. 4L-M) is the proximal 3mm of a ?pectoral fin spine which has a delicate sheet DESCRIPTION. Scapulocoracoids (Figs of perichondral ossification preserved across the 3D,4A,B). Samples from GY 11 include four longitudinal cleft. This sheet has three holes, each scapulocoracoids, two left and two right. 0.2-0.3mm long: a proximal pair and a more ANUV2939.5 (Fig. 4A) is an element from the distal single central hole, which probably right side, 2.5mm high, with a long thin scapular represent the articulation points of the fin radials, shaft having a circular cross-section of 0.4mm showing that these fin elements were arranged in diameter, expanding to a thin wide-based lateral a triangle rather than linearly. scapular blade with a stretched-triangular shape; the blade is 1.5mm long x 0.5mm high anteriorly. Non-denticulated fin spines (Figs 3G,4E-G,K). A rounded medial knob at the base of the scapular Paratype ANUV2969.5 (Fig. 4K) is a complete, shaft is possibly a glenoid process. ANUV2969.8 symmetrical? spine 12mm long, with a smooth (Fig. 4B) is a left scapulocoracoid, 3.6mm high. leading edge ridge and several (up to five) fine ANUV2939.6 is a left scapulocoracoid, 5mm longitudinal ridges on each side. A delicate sheet high, with only the upper part of the lateral of perichondral ossification is partly preserved scapular blade preserved. Paratype boxing over the open pulp cavity of the proximal ANUV2969.2 (Figs 3D-E,4C-D) is a right half of the spine, and is presumed to have covered scapulocoracoid, 2.7mm high, with the scapular the basal cartilage of the fin. As the spine is shaft, ?glenoid process, medial coracoid and symmetrical, and dorsal fins are the only median lateral scapular blade all preserved. Some fins for which basal cartilages have been delicate perichondral granular mineralisation is recorded in any acanthodians, the spine is preserved anteriorly, with no indication of a probably from the dorsal fin. Two fragments from coracoid process or procoracoid. sample ANUV2939 are proximal ends of spines, 10 MEMOIRS OF THE QUEENSLAND MUSEUM DEVONIAN ACANTHODIANS 11 FIG 4. Pectoral girdles, fin spines and palatoquadrate of Teneracanthus toombaensis gen. et sp. nov. from GY11. A, right scapulocoracoid, medial view ANUV2939.5; B, left scapulocoracoid, lateral view ANUV2969.8; C-D, right scapulocoracoid, medial and anterior views ANUV2969.2; E-G, non-denticulatcd spine fragment ANUV2939.8 E. lateral view. F, end-on view, and G, latero-basal view; H, lateral view denticulated fin spine fragment ANUV2939.7; J, transverse section ANUV2939.9, under cross nicols; K, median fin spine ANUV2969.5, showing perichondral ossification of basal cartilage; L-M, proximal cleft of pectoral fin spine ANUV2969.1, showing three holes in the perichondral ossification; N, ?left pectoral fin spine ANUV2939.10, dorsal view ; P-Q, denticulated fin spine ANUV2969.6 P, lateral view, and Q, trailing edge view, R-S, articulated right scapulocoracoid and pectoral fin spine ANUV2940, R, lateral view, and S, medial view; T, fractured perichondral ossification of left palatoquadrate ANUV2969.4. Scale bar=lmm in A-D, K, N-S, 0.5mm in L-M, T, and 0.1mm in E-J; arrow to anterior. 12 MEMOIRS OF THE QUEENSLAND MUSEUM FIG. 5. Scales of Teneracanthus toombaensis gen. et sp. nov. from GY 11 (A-J) and BRJ104B (K). A, antero-lateral view flank scale ANUV2939.12; B, vertical longitudinal section flank scale, ANU V2939.13, under cross nicols; C, horizontal section crown flank scale ANUV2939.14, under cross nicols; D, E, vertical transverse section flank scale ANUV2939.15, D under cross nicols; F, G, caudal scale (sciotoensis var.) ANUV2937.8, F, crown view, G, latero-crown view; H, caudal scale ANU V2939.16, basal view; J, ?cheek tessera ANUV2939.17, crown view; K, Vertical longitudinal section flank scale UQY9335. Scale bar=0.1 mm; arrow is anteriad; b = base, c =crown, cW =canals of Williamson, e = enameloid.