n (/> — c/> r: CO — lan LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI NVINOSHi^S S3ldVygn LIBRARI 2 ^ ^ 7H ■, CO 2 5 ■' V 2 0 CO 1 3> 2 5 > CO 2 Jo •• " Z W z t/> TION NOlifUliSNI NVINOSHil/JS S3iavyan LIBRARIES SMITHSONIAN INSTITUTION NOliflili CO 2 -J lan LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI NVINOSHimS SBILIVyan LIBRAR r~ ^ >~ 2 r- 2 r- O :o > 33 — xt;" m XgN pcjx ^ m — x^oiayix m CO ± CO £ — _ TION NOliniiiSNI NVINOSHillNS SBiavaail LIBRARIES SMITHSONIAN INSTITUTION NOlinili CO 2 CO 2 CO 2 W CO (/) lan LIBRARIES SMITHSONIAN INSTITUTION NOlifUliSNI NVINOSHliWS S3iavaan LIBRAR ^ ^ ~ cO ^ oc /c J'^l < O' \o —1 \-t_ dcJ/ — O TION NOliniliSNl'^NVINOSHillNS S3IBVMan LIBRARIES SMITHSONIAN INSTITUTION NOlinili 2 ^ I" 2: “ o mV 9 9 m xT 03 33 > 33 ~ rn ' xv>- ^ (T) CO ■ _ CO an LIBRARIES SMITHSONIAN INSTITUTION NOliOiliSNI NVINOSHilWS S3IHVaan LIBRARI Z A ^ ^ 2 > CO 2 ^ ^ ■X V ^ < \j. S -Jfe. < 2 S 4;!. Z -C 2 H ^ ^ ^ J, ^ ...- 2 XOyosvli^ > CO 2 CO •*’" 2 CO Z CO TION NOliniiiSNI NVINOSHimS S3iaVHai1 LIBRARIES SMITHSONIAN INSTITUTION NOlinili CO — N. CO — CO 5 CO CO V ^ m icj m xToovoX ^ Jan LIBRARIES SMITHSONIAN INSTITUTION NOliniiiSNI NVINOSHillNS S3iavaan LIBRAR m ^ rn w rn CO ± CO £ — CO TION NOliniiiSNI NVINOSHimS S3iaVMai1 LIBRARIES SMITHSONIAN INSTITUTION NOlinili CO 2 CO 2 CO 2 ^ C/^ X “ ^ ~ . , ITHSONIAN INSTITUTION NOliniliSNI NVINOSHillMS S3ldVyan LI -y . in z. z > 2 >■ 5 C/5 2 CO 1 LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI _ NVINOSHimS S 3 I ~ 5 ^ z \ ^ z /#%\ - < m O XL\5T'^^ _ O _ Z _l z IITHSONIAN INSTITUTION NOliniliSNI NVINOSHilWS SBIdVaan LIBRARIES SMITHSONIAN_ INS7 z r- z <” 5 m ^;c2LHi^ 5; m in — c/> iNOSHims s3iHvaan libraries Smithsonian institution NoiiniiiSNi nvinoshhins s3 in z t/5 z y.- ^ ^ ^ < ax\ . s 2 .< > _ Z c/7 * Z CO Z cn ITHSONIAN INSTITUTION NOliniliSNI NVINOSHilWS S3iyvaail LIBRARIES SMITHSONIAN_ INS! Z CA) 5 V z ^ ^ in ^ . CO < CC Q " ^<^gg-J2>^ O X^M. pc^^ ^ O IN0SHimS^S3 iavyan~LIBRARI ES^SMITHSONIAN~'lNSTITUTION^NOIini!iSNl“’NVINOSHimS S3 >Z r- Z ~ “ O ~ O ™ /ToTvoX LaTv'^ m ^ ' m XL^osvcjy xk;a^ m CO ? CO — CO — IITHSONIAN INSTITUTION NOliniliSNI NVINOSHillMS S3iavaan L I B R A R I E S SMITHSONIAN INSl in X CO o ? » / — M-^MW.3^^7 =1/ > ■ S x^osv^vx > 2 CO 2 CO ■ ' CO 3 LIBRARIES SMITHSONIAN INSTITUTION NOlinillSN I _ NVINOSHIIINS^^ S 3 CO = c/7 2 C / ■■' ■' O _ o Z _J Z — ' ' VllTHSONIAN INSTITUTION NOliniliSNI NVINOSHIIWS S3ldVyan LIBRARIES SMITHSONIAN INS m xc/v o<-^ m m nq\. dc^ ^ cn CO ± CO ? CO _ CO viNosHims S3iavaan libraries Smithsonian institution NoiiniiiSNi nvinoshiiiais s3 CO Z CO Z V- ^ ^ ^ ^ VOLUME 6 • PART 3 Palaeontology OCTOBER 1963 PUBLISHED BY THE PALAEONTOLOGICAL ASSOCIATION LONDON Price £3 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations, and pubUshes the quarterly journal Palaeontology. Membership is open to individuals, institutions, libraries, &c., on payment of the appropriate annual subscription: Institute membership . . . . £5. 5s. (U.S. $15.50) Ordinary membership . . . . £3. 3.y. (U.S. $9.50) Student membership . . . . £2. 25'. (U.S. $6.50) There is no admission fee. Student members will be regarded as persons receiving full-time instruction at educational institutions recognized by the Council; on first applying for membership, they should obtain an application form from the Secretary or the Treasurer. Subscriptions are due each January, and should be sent to the Treasurer, Dr. T. D. Ford, Department of Geology, The University, Leicester, England. Palaeontology is devoted to the publication of papers (preferably illustrated) on all aspects of palaeontology and stratigraphical palaeontology. Four parts are published each year and are sent free to aU members of the Association. Members who join for 1963 will receive Volume 6, Parts 1 to 4. All back numbers are still in print and may be ordered from B. H. Blackwell, Broad Street, Oxford, England, at the prices shown below (post free) : Vol. 1 (for 1957-8) in 4 parts at £2 or U.S. $6.00 per part. Vol. 2 (for 1959) in 2 parts at £2 or U.S. $6.00 per part. Vol. 3 (for 1960) in 4 parts at £2 or U.S. $6.00 per part. Vol. 4 (for 1961) in 4 parts at £2 or U.S. $6.00 per part. Vol. 5 (for 1962) in 4 parts at £3 or U.S. $9.00 per part. A complete set. Volumes 1-5, consists of 18 parts and costs £40 or U.S. $120. Manuscripts on all aspects of palaeontology and stratigraphical palaeontology are invited. They should conform in style to those already pubhshed in this journal, and should be sent to Mr. N. F. Hughes, Sedgwick Museum, Cambridge, England. A sheet of detailed instructions for authors wfil be suppUed on request. CAMBRIAN TRILOBITES FROM THE PURLEY SHALES OF WARWICKSHIRE Abstract. Trilobites from the Parley Shales of Warwickshire are described and illustrated, including those from a new locality. It is concluded that the lowest 40 ft. of the Parley Shales belong to the Lower Cambrian, and at least 450 ft. of higher beds are probably of Paradoxides oelmdicus age (basal Middle Cambrian). The Cambrian rocks which crop out between Atherstone and Bedworth, Warwickshire, were subdivided by Lapworth (1898, pp. 338-50) as follows: Lapworth suggested that the Parley Shales were about 600 ft. thick, and from these Shales he recorded a number of horny brachiopods together with a single trilobite pygi- dium associated with a few thoracic segments. Of this trilobite he wrote: Tt is probably referable to Conocoryphe, and allied to C. cownata Barrande, and C. ex{s)idans Lin- narsson’ (Lapworth, 1898, p. 346). Pringle (1913, p. 71) and Illing (1913, p. 452) recorded Callavia sp. from red calcareous nodules near the base of the Purley Shales and from purple shales 40 ft. above the base, respectively. Illing (1916, p. 436) concluded that at least the lowest 40 ft. of the Purley Shales belong to The middle zone of the Lower Cambrian’. Illing (1916, p. 436) referred to the Conocoryphe mentioned by Lapworth and also to ‘a head of Paradoxides, a very convex type with well-developed though rather small eyes, and faint glabellar furrows corresponding to P. sjo{e)greni, which is found in the P. oelandicus Zone of Sweden’. Both these trilobites were found loose in a heap of shales by the roadside in Purley Park Lane, one mile south of Atherstone. Illing considered that they came from the Purley Shales at a position about 100 ft. below the base of the Oldbury Shales, and he tentatively equated this horizon with the P. oelandicus Stage of Sweden, i.e. the lowest stage of the Middle Cambrian. No records of other trilobites from the Purley Shales have been published except for a brief note concerning some of the finds described below (Smith, 1962, p. 54). The trilobites which form the basis of the present paper were collected in 1961 from temporary exposures of the Purley Shales at the site of a new factory on the Camp Hill Industrial Estate, one mile west-north-west of the centre of Nuneaton, Warwick- shire. The locality is 1,790 yds. N. 61° E. of the north-east corner of St. Paul’s Church, [Palaeontology, Vol. 6, Part 3, 1963, pp. 397-407, pis. 57-58.] C 1456 D d by J. D. D. SMITH and d. e. white lies Stockingford Shales es Hartshill Quartzite PRESENT RESEARCH 398 PALAEONTOLOGY, VOLUME 6 Stockingford : (N.G.R. SP/34819219). A specimen of Paradoxides was collected by Mr. J. E. Wright of the Geological Survey during an examination of the excavations. Trilobites, horny brachiopods, and sponge spicules were collected by the authors, who made subsequent visits. The Paradoxides was collected from a 2-inch band of pale brown, sandy siltstone exposed in a trench. Other trilobites were collected 67 ft. to the east-south-east in an adjacent trench, from a 2-inch band of similar lithology thought to represent the same horizon. Near this trench, two trilobites were collected in 1962 from loose material also of similar lithology by Mr. A. W. A. Rushton, who first informed the Geological Survey of these excavations. The authors found no other bands of pale brown, sandy siltstone in the exposure. Numerous specimens of LinguleUa and Acrotreta s.l. and sponge spicules of Protospongia fenestrata Salter were collected, these being conhned to strata immediately overlying the band in which the Paradoxides was found. The excavations exposed a thickness of about 130 ft. of purple shales, and three thin sills, striking at 110° and with a dip varying between 70° and vertical. The trilobite band cropped out 18 ft. below the highest beds seen in the excavations. It is not possible to establish exactly the stratigraphical position of the present ex- posure within the succession of the Purley Shales. However, from a consideration of the six inches to one mile Geological Survey maps of the Coventry area, it is estimated that the Purley Shales, with included sills, are at least 900 ft. thick, and that the trilobite-band occurs at least 350 ft. from the base. These estimates are based on the assumption that there are no major unmapped faults or thick sills in the Purley Shales. Mr. Rushton has since found trilobites at other temporary exposures in the Purley Shales. The trilobites from the locality on the Camp Hill Industrial Estate have been identified as: Condylopyge carinata Westergard ? C. regia (Sjogren) Pleuroctenium cf. gramdatum (Barrande) P. sp. Eodiscid Paradoxides cf. sedgwickii (Hicks) Bailiella emarginata (Linnarsson) SYSTEMATIC PALAEONTOLOGY The classification and morphological terms used are those adopted in Part O (Arthro- poda 1) of the Treatise on Invertebrate Palaeontology, Moore 1959. The lateral glabellar furrows are numbered from the posterior end forwards, and the pygidial furrows from the anterior end backwards. The localities of specimens described are given in the explanation of plates. The zonal classification of the Middle Cam.brian is indicated on pp. 405-6. Specimens in the Geological Survey and Museum collections bear the prefix GSM, those in the University of Birmingham collections the prefix BU, and those in the Sedgwick Museum. Cambridge, collections the prefix SM. J. D. D. SMITH AND D. E. WHITE: CAMBRIAN TRILOBITES 399 Family condylopygidae Raymond, 1913 Genus condylopyge Hawle and Corda, 1847 Condylopyge carinata Westergard Plate 57, fig. 2 1936 Condylopyge carinata Westergard, p. 27, pi. 1, figs. 4-8. Material. Internal (GSM 102117) and external (GSM 102118) moulds of one pygidial axis. Description. Pygidial axis convex, outline incompletely preserved. Trace of first pair of furrows at anterior end of axis ; second pair sliort and transverse ; third pair transverse, furrows meeting across axis just behind its centre. Median keel prominent in third axial ring, fading out in second ring. Faint, shallow, median furrow (or crack) on posterior part of axis. Discussion. The pygidial axis agrees with that illustrated for Condylopyge carinata. It dilfers from C. regia (Sjogren) in having a median keel rather than a slightly elongate tubercle, and from C. rex (Barrande) in that the median keel does not extend into the first ring. Age. Westergard (1936, p. 27) recorded C. carinata from the Zone of Paradoxides pimis, P. oelandicus Stage. Condylopyge sp. Plate 57, figs. 3, 4 Material. Internal (GSM 102119) and external (GSM 102120) moulds of one cephalon; external (GSM 1021 14) mould of one cephalon. Description. Cephalon parallel sided, anterior margin rounded. Border narrow, convex. Border furrow broad, shallow. Axial furrow deep, sharp. Glabella convex, consisting of two lobes separated by deep, transverse furrow. Anterior glabellar lobe semicircular in outline, extended postero-laterally. Posterior glabellar lobe subquadrate, longer than broad, narrower than anterior lobe, with slight median ridge towards posterior, and prominent median tubercle at posterior end. Genal region narrow around anterior glabellar lobe, broadens adjacent to posterior lobe. Short, paired, lateral spine extending backwards from genal region, with apex near mid-breadth. Discussion. The two cephala agree closely with Condylopyge carinata and C. regia, and are identified as one or other of these species. Westergard (1936, p. 27) has pointed out that it is only in the pygidium that these two species can be differentiated. The cranidium of C. carinata and C. regia can be distinguished from C. rex by the absence of a tubercle just in front of the centre of the posterior glabellar lobe; the specimens described above have no such tubercle. Age. Westergard (1936, p. 27) recorded C. carinata and C. regia from the Zones of Paradoxides pinus and P. insularis respectively, P. oelandicus Stage. 400 PALAEONTOLOGY, VOLUME 6 Genus pleuroctenium Hawle and Corda, 1847 Pleurocteniinn granulatum (Barrande) Plate 57, figs. 5-9 1846 Battiis giamilatus Barrande, p. 15. 1847 Pleuroctenium granulatum Hawle and Corda, p. 117, pi. 6, fig. 63. 1916 Agnostus granulatus Illing, p. 419, pi. 32, figs. 11-13. 1958 Pleuroctenium granulatum Snajdr, p. 56, pi. 58, figs. 5, 7-15. Material. Internal (GSM 102121, GSM 102123, and SM A53379a) and external (GSM 102122, GSM 102124, and SM A53379b) moulds of three pygidia; GSM 102123-4 with two thoracic segments attached. Description. Two thoracic segments. Paired lateral tubercles on each axial ring; large median tubercle on posterior axial ring. Pygidium rounded posteriorly, slightly tapering forwards. Border slightly convex. Paired lateral spines directed backwards, half as long as pygidium. Pleural field slightly broader than border, slightly convex and fiat behind axis. Axis strongly convex, con- tracted at middle, subacuminate posteriorly; three pairs of very short, weak transverse furrows; first and second pairs directed slightly forwards; third pair slightly backwards. Median tubercle at posterior end of each of first three axial rings; first tubercle very small, second and third tubercles prominent; third tubercle slightly elongated and directed backwards and upwards. Large terminal axial piece. EXPLANATION OF PLATE 57 All specimens were coated with ammonium chloride before photography by Mr. J. D. Thompson, Geological Survey and Museum. Fig. 1. Eodiscid. Internal mould of pygidium (GSM 1021 15). Fig. 2. Condylopyge carinata Westergard. Internal mould of pygidial axis (GSM 102117). Figs. 3, 4. Condylopyge sp. 3, Latex cast from external mould of cephalon (GSM 1021 14). 4, Internal mould of cephalon (GSM 102119). Figs. 5-9. Pleuroctenium granulatum (Barrande). 5, 6, Internal (GSM 102121) and external (GSM 102122) moulds of pygidium. 7, Internal mould of pygidium (SM A53379a). 8, 9, Internal (GSM 102123) and external (GSM 102124) moulds of pygidium with two thoracic segments. Figs. 10, 11. Pleuroctenium sp. Internal (GSM 102125) and external (GSM 102126) moulds of pygidium. Figs. 12, 13. Bailiella emarginata (Linnarsson). 12, Latex cast from external mould of cranidium (GSM 102113). 13, Part of fixigena of same to show surface markings. Fig. 14. Conocoryphid. Internal mould of pygidium and part of thorax (BU 546a). Fig. 15. Callavia? Internal mould of a left thoracic pleura (GSM 102127). Locality of figs. 1-13. Purley Shales, at least 350 ft. above base. Camp Hill Industrial Estate, 1,790 yds. at 61° from north-east corner of St. Paul’s Church, Stockingford, Warwickshire. N.G.R. — SP/34819219. Middle Cambrian — Paradoxides oelandicus Stage. Locality of fig. 14. Purley Shales, estimated at 100 ft. below top. Found loose, believed to be from Purley Park Lane, 1,210 yds. at 237° from south-west corner of Church, Mancetter. N.G.R. — SP/3 1029604. Middle Cambrian — Paradoxides oelandicus Stage. Locality of fig. 15. Purley Shales, calcareous nodules near base. Camp Hill Grange Quarry, 2,450 yds. at from north-east corner of St. Paul’s Church, Stockingford. N.G.R. — SP/33349364. Lower Cambrian — ? middle part. Palaeontology, Vol. 6 PLATE 57 SMITH and WHITE, Cambrian trilobites J. D. D. SMITH AND D. E. WHITE: CAMBRIAN TRILOBITES 401 Discussion. The three pygidia agree closely with Pleuroctenium granulation. No granu- lation can be seen, but this is considered to be due to poor preservation. Age. Pleuroctenium granulation has not been recorded from Sweden. Snajdr (1958, p. 59) recorded it from the Skryje Beds of Bohemia, Zone of Eccaparacloxides pusillus, which also contain Condylopyge rex and so are probably equivalent to the Zone of Hypagnostus parvifrons {Paradoxides paradoxissimus Stage) of Sweden, filing (1916, p. 419) recorded P. granulation from his horizons C2-F3 with one specimen from A4, i.e. from the Zone of Paradoxides aurora and above. Pleuroctenium sp. Plate 57, figs. 10, 11 Material. Internal (GSM 102125) and external (GSM 102126) moulds of one pygidium. Description. As for Pleuroctenium granulation, but pygidial outline more rounded, pleural field twice width of border, and lateral spines diverge more markedly from border. Discussion. This specimen is similar in the above respects to Pleuroctenium tuberculatum (filing) (1916, pp. 421-2, pi. 33, figs. 4-8), but differs significantly from that species in having a large terminal axial piece. Age. fn Sweden, the genus Pleuroctenium has the same range as Condylopyge, i.e. from the lowest zone of the Middle Cambrian to the highest zone of the Paradoxides para- doxissimus Stage (Westergard, 1946, p. 35). Family eodiscidae Raymond, 1913 EODISCID Plate 57, fig. 1 Material. Internal (GSM 102115) and external (GSM 102116) moulds of one incomplete pygidium. Description. Margin entire. Border narrow, convex. Border furrow deep. Pleural field convex, with three, or perhaps four, furrows. Middle part of axis narrow, convex with two furrows. Age. The family Eodiscidae is of Lower and Middle Cambrian age. The genus Eodiscus is confined to the Middle Cambrian. Westergard (1946, pp. 22-23) recorded one species, E. oelandicus (Westergard), from the Zone of Paradoxides pinus; this species has four pairs of pygidial furrows, as does E. sculpt us (Hicks) from the P. liarknessi Zone of South Wales. Family olenellidae Vogdes, 1893 Subfamily callaviinae Poulsen, 1959 Genus callavia Matthew, 1897 Call avia? Plate 57, fig. 15 1913 Callavia Pringle, p. 71. Material. Left-hand part of a thoracic pleura (GSM 102128) and its internal mould (GSM 102127); ten other thoracic fragments. 402 PALAEONTOLOGY, VOLUME 6 Description. Pleura gently curved with strong, falcate extremity. Proximal part of pleura deeply grooved, groove dying out less than half-way towards tip. Discussion. One of Pringle’s fossils from that part of the Parley Shales referred to the middle zone of the Lower Cambrian by Illing (1916, p. 436) is illustrated for the first time. Dr. J. W. Cowie has re-examined this specimen first referred to Callavia by P. Lake and comments that the fragmentary nature of the material makes a firmer determination than Callavia? hazardous. Nevertheless, he accepts a Lower Cambrian age as probable. Prof. V. C. Illing in a letter has assured the authors that his specimens identified as Callavia sp. from the Purley Shales of Worthington Farm (Illing, 1913, p. 452) were no more complete than the fragmentary pleura found by J. Pringle and described above. Family paradoxididae Hawle and Corda, 1847 Subfamily paradoxidinae Hawle and Corda, 1847 Genus paradoxides Brongniart, 1822 Paradoxides cf. sedgwickii (Hicks) Plate 58, figs. 1-8 1871 Plutonia Sedgwickii Hicks, p. 399, pi. 15, figs. 1-8. 1935 Paradoxides sedgwicki Lake, p. 221, pi. 31, figs. 7-10. Material. Internal ( BU 545a) and external ( BU 545b) moulds of one cephalon; internal (GSM 102108) and external (GSM 102109) moulds of one cranidium; internal (SM A53378a) and external (SM A53378b) moulds of part of the thorax comprising eight articulated axial rings of which four have incomplete pleurae attached; internal (GSM 102110) and external (GSM 102111) moulds of one incomplete thoracic segment showing the axial ring and the right pleura. Description. Cephalon semicircular, slightly pointed anteriorly. Border faintly striated, broad at facial suture, narrowing rapidly towards glabella. Palpebral lobes large, gently curved, convex outwards, touching glabella at its broadest point, extending obliquely backwards to line of first lateral glabellar furrow. Anterior branch of facial suture runs forwards and outwards to margin of cephalon; posterior branch directed backwards EXPLANATION OF PLATE 58 All specimens were coated with ammonium chloride before photography by Mr. J. D. Thompson, Geological Survey and Museum. Figs. 1-8. Paradoxides cf. sedgwickii (Hicks). 1, 2, Internal (GSM 102108) and external (GSM 102109) moulds of cranidium. 3, 4, Internal mould of cephalon (BU 545a), 3 to show impression left by supposed genal spine. 5, Internal mould of one axial ring and pleura (GSM 1021 10). 6, Internal mould of eight axial rings, four with pleurae attached (SM A53378a). 7, Occipital ring and part of glabella (BU 545a). 8, Occipital ring and part of glabella (GSM 102108). Fig. 9. Paradoxides sedgwickii (Hicks). Occipital ring and part of glabella of syntype (SM A1086), figured by Hicks (1871, pi. 15, fig. 1). Locality of figs. 1, 2, 5, 6, 8. Purley Shales. As for Plate 57, figs. 1-13. Locality of figs. 3, 4, 7. Purley Shales. As for Plate 57, fig. 14. Locality of fig. 9. Solva Beds. Near Nun’s Well, St. David's area, Pembrokeshire. Palaeontology, Vol. 6 PLATE 58 SMITH and WHITE, Cambrian trilobites fiiV* ■T*: J. D. D. SMITH AND D. E. WHITE: CAMBRIAN TRILOBITES 403 and outwards, junction with margin obscure. Axial furrow shallow, deepening pos- teriorly. Glabella strongly convex in BU 545, crushed flat in GSM 102108, pyriform, tapering posteriorly and slightly pointed anteriorly where it almost meets border. Four pairs of lateral glabellar furrows. First pair directed inwards and slightly backwards, with furrows united across glabella by shallow groove. Second pair directed slightly backwards with furrows united across glabella by very shallow groove in BU 545, transverse with furrows not united in GSM 102108. Third and fourth pairs short, shallow, directed slightly forwards. Fourth pair slightly behind line of maximum breadth of glabella. Occipital ring strongly convex with small median tubercle. Occipital furrow very deep laterally, shallow medially. Cephalon surface covered by close-set granules about OT mm. diameter. Thoracic pleurae weakly convex, smooth, with prominent pleural furrow. Axis weakly convex, surface bears close-set granules, some united by short ridges. Granules ap- proximately equal in size, rather coarser than on cephalon. In the matrix of BU 545a there is an impression which may have been caused by a thin, smooth genal spine; the impression, though in a different plane, follows the line of the outer margin of the cephalon but is not, however, continuous with it. It may even be a thoracic pleural spine. Discussion. Illing (1916, p. 436) compared BU 545 with Paradoxides sjoegreni. Dr. M. Lindstrom, of Lund University, Sweden, has sent the authors examples of this species from Oland, Sweden. BU 545 and GSM 102108 differ from the Swedish speci- mens in the following features: 1. The glabella tapers backwards much more. 2. The maximum width of the glabella occurs farther back. 3. The granulation is much coarser than in P. sjoegreni, only small specimens of which are granulate. 4. The median portion of the second pair of lateral glabellar furrows is not as deep and wide as in P. sjoegreni. Dr. Lindstrom has examined the Paradoxides from the Purley Shales, and in his opinion they do not belong to P. sjeogreni or to any other species occurring in Sweden. The only species of Paradoxides with granulation as coarse and prominent as those from the Purley Shales is P. sedgwiekii. Indeed, the surface markings of this species are so distinctive that Hicks (1871, p. 399) erected a new genus, Pbitonia, to contain it. Hicks (1871, pi. 15, fig. 1) figured a cephalon of P. sedgwiekii from the Solva Group of the St. David’s area, Pembrokeshire. Examination of this syntype (SM A1086) indicates that the surface markings are similar to those of BU 545 and GSM 102108, although slightly coarser, due in part to distortion. (Compare fig. 9 with figs. 7 and 8 of Plate 58.) In P. sedgwiekii the lateral glabellar furrows of the second pair are united across the glabella by a deeper groove than BU 545 and the palpebral lobe is relatively smaller than in the two specimens from the Purley Shales. Hicks (1871, pi. 15, figs. 6-8) also figured parts of the thorax of P. sedgwiekii illustrat- ing its coarsely granulate axial segments and pleurae. The axial segments of GSM 102110 and SM A53378 have a more coarsely granulate surface than the cranidia GSM 102108 and BU 545, although the pleurae are smooth. 404 PALAEONTOLOGY, VOLUME 6 Despite these diflferences it is considered that the specimens of Paradoxides from the Parley Shales may be compared with P. sedgwickii. Age. Paradoxides sedgwickii has been recorded only from the Solva Group of the St. David’s area, Pembrokeshire (Lake, 1935, p. 222). Hicks recorded it from the same zone as P. harknessi, i.e. in the lowest division of the Solva Beds. Dr. M. Z. Farshori has permitted the authors to state that he has found specimens of P. sjoegreui together with librigenae of P. sedgwickii 50 ft. above the base of the Solva Beds. This association suggests that P. sedgwickii occurs within the Stage of P. oelandicus. Family conocoryphidae Angelin, 1854 Genus bailiella Matthew, 1885 BailieUa emarginata (Linnarsson) Plate 57, figs. 12, 13 1877 Coiwconphe emarginata Linnarsson, p. 15, pi. 2, figs. 2-4. 1936 BailieUa emarginata Westergard, p. 58, pi. 11, figs. 1-3. 1950 BailieUa emarginata Westergard, p. 25, pi. 5, figs. 1-4. Material. Internal (GSM 102112) and external (GSM 102113) moulds of one cranidium. Description. Cranidium subtrapezoidal. Anterior margin slightly curved. Anterior border convex, almost uniform in breadth, although expanding slightly in front of glabella. Anterior border furrow deep, except in front of glabella where narrow pre- glabellar field rises gently to border. Fixigenae convex, broad. Eyes absent. Facial sutures gently sinuous. Glabella convex, tapering forwards and rounded anteriorly, defined by deep axial furrows; posterior margin of glabella one-third breadth of crani- dium. Two pairs of short, weak lateral glabellar furrows, directed slightly backwards. Occipital furrow deep, shallower medially. Posterior border furrow deep. Blunt median tubercle on occipital ring. Outer surface granulate, rather coarser on glabella than on fixigenae. Discussion. Westergard (1950, pi. 5, fig. 2) illustrated the flattened, shale-preserved lectotype of BailieUa emarginata (Linnarsson). GSM 102113 agrees closely in size and other characteristics with Westergard’s description and photograph, except that the anterior margin is slightly less rounded. Westergard (1950, pi. 5, fig 4) illustrated a small part of a fixigena of B. emarginata, showing tubercles approximately 0-07 mm. diameter. These compare closely with the surface markings of GSM 102113 (PI. 57, fig. 13), although some tubercles reach 0T2 mm. diameter. The cranidium resembles the stratigraphically later B. aequalis (Linnarsson) in general outline and in the uniformly convergent nature of the axial furrows. However, it differs from that species in having finer granulations, a slightly expanding border in front of the glabella, a border and preglabellar field of approximately equal axial length, and a less convex preglabellar field. Age. Westergard (1950, p. 25) recorded B. emarginata from the Zone of Paradoxides insidaris, P. oelandicus Stage. J. D. D. SMITH AND D. E. WHITE: CAMBRIAN TRILOBITES 405 ? CONOCORYPHID Plate 57, fig. 14 Material. Internal (BU 546a) and external (BU 546b) moulds of one pygidium with posterior thoracic segment attached and four other disturbed but articulated thoracic segments. Description. Thorax tapers backwards, the breadth (transverse) of the axial ring about half that of the pleura. Axis convex; pleural region less so. Axial and pleural furrows deep. Axial rings granulate, each with three to five small tubercles in line; finer granula- tion on pleurae, with scattered small tubercles. Pygidium semi-elliptical, posterior margin straight behind axis. Pleural field slightly convex, axis more so. Four pleurae; interpleural furrows weak, pleural furrows promi- nent. Axial furrow deep, sharp. Three axial rings, and short terminal axial piece rounded posteriorly. Doublure smooth, concave, narrow. Pleural fields granulate, axis more coarsely granulate except for smooth terminal axial piece. Three to five small tubercles in line on each axial ring and each pleura. Discussion. Lapworth (1898, p. 346) considered this specimen to be allied to Cteno- cephalus [Conocoryphe] coronata (Barrande) and C. [C.] exsuJans (Linnarsson). The pygidium of the genus Ctenocephalus is defined in the Treatise (Moore 1959), Part O, p. 242, as ‘very small, with extremely wide pleural fields, axis consisting of 1 or 2 rings and short terminal portion’. The pygidium of C. coronata is illustrated by Barrande (1852, pi. 13, fig. 23) and Snajdr (1958, pi. 35, fig. 1 1). BU 546 differs markedly from the description and illustrations. It can probably be referred to the family Conocoryphidae, but is not sufficiently complete to determine generically. Age. The family Conocoryphidae, as defined in the Treatise (Moore 1959), extends from the Lower Cambrian to the Ordovician. The genus Conocoryphe is of Middle Cambrian age. AGE OF THE PURLEY SHALES Dr. J. W. Cowie has examined the pleura (GSM 102127) collected by J. Pringle from calcareous nodules near the base of the Purley Shales. As indicated above. Dr. Cowie suggests that the horizon represented may be within the middle part of the Lower Cambrian. Tiling (1916, p. 436) compared the trilobite (BU 545) found at about 100 ft. from the top of the Purley Shales with Paradoxides sjoegreni, and tentatively assigned that horizon to the P. oelandicus Stage of Sweden. This must be reconsidered in the light of the reidentification of the trilobite as P. cf. sedgwickii and the evidence provided by the new trilobite fauna described in this paper. Westergard (1946, p. 8) divided the Middle Cambrian thus: C. Paradoxides forchhammeri Stage B. Paradoxides paradoxissimus Stage 4. Ptychagnostus pimctuosus Zone 3. Hypagnostus parvifrons Zone 2. Tomagnostus fissus and Ptychagnostus atavus Zone 1 . Ptychagnostus gibbus Zone 406 PALAEONTOLOGY, VOLUME 6 A. Paradoxides oeJandicus Stage 2. Paradoxides pinus Zone 1 . Paradoxides insularis Zone Illing (1916, p. 442) divided the Abbey Shales into twenty-two horizons whieh he grouped into the following faunal sequence; G3— G1 F3— FI E3 — El (pars) El (pars), D3 — D1 C3— Cl, B3— B1 A4— A1 Upper Paradoxides davidis Fauna Lower Paradoxides davidis Fauna Hartshillia injiata (Passage) Fauna Upper Paradoxides hicksii Fauna Lower Paradoxides hicksii Fauna Paradoxides aurora Fauna Of the species collected from the new exposures in the Purley Shales, Condylopyge carinata and Bailiella emarginata have been recorded from Sweden. There they are confined to the Paradoxides oelandicus Stage, the former occurring in the P. pinus Zone, the latter in the P. insularis Zone. Pleuroctenium granulatum has not been recorded from Sweden. In Bohemia it occurs in the probable equivalent of the Zone of Hypagnostus parvifrons. Illing has recorded it from the Zones of Paradoxides davidis and P. hicksii and from his horizon A4 in the Zone of P. aurora. A4 is the lowest horizon in the Abbey Shales from which Illing recorded determinate trilobites, and it is here considered that the range of Pleuroctenium granulatum may extend down into the Paradoxides oelandicus Stage. Paradoxides sedgwickii has not been recorded outside Pembrokeshire. Hicks (1871, p. 399) described this species from the lowest division of the Solva Group in the St. David’s area. He recorded it from the same Zone as P. harknessi, i.e. below the Zone of P. aurora. This is accepted by Lake (1935, p. 222). As stated above, P. sedgwickii is considered to occur within the Stage of P. oelandicus. The lower limit of the occurrence of P. sedgwickii is not known, since the Solva Group lies unconformably on the Caerfai Group, of supposed Lower Cambrian age. Conclusion. On the trilobite evidence it is concluded that the horizon about 350 ft. from the base of the Purley Shales is probably of P. oelandicus age. Unless the species de- scribed as P. cf. sedgwickii is long- ranging, the horizon about 100 ft. below the top of the Purley Shales can be assigned to the same stage, as was tentatively suggested by Illing (1916, p. 436). On the assumption that the total thickness of the Purley Shales is 900 ft. (see p. 398), a thickness of at least 450 ft. for the P. oelandicus Stage in the Purley Shales is indicated. Acknowledgements. We wish to record our thanks to the following: Prof. O. M. B. Bulman, F.R.S., and Mr. A. W. A. Rushton for informing the Geological Survey of the temporary exposures, and Mr. Rushton for permitting us to illustrate his specimens (SM A53378 and A53379); Mr. A. G. Brighton (Sedgwick Museum) and Dr. I. Strachan (University of Birmingham) for sending us specimens in their care; Dr. J. W. Cowie (University of Bristol), Dr. M. Z. Farshori (formerly King’s College, London), Prof. V. C. Illing, and Dr. M. Lindstrom (University of Lund, Sweden) for valuable assistance as indicated in the text. We are particularly indebted to Dr. C. J. Stubblefield, F.R.S., who has given us much assistance in the examination of the specimens and the preparation of this paper; it is published by permission of the Director, Geological Survey and Museum. J. D. D. SMITH AND D. E. WHITE: CAMBRIAN TRILOBITES 407 REFERENCES BARRANDE, J. 1846. Noiiveaiix trilobites, supplement a la notice preliminaire siir le systeme siliirien et les trilobites de Boheme. Prague. 1852. Systeme silurien dii centre de la Boheme. lire, partie: Recherclies paleontologiques. Vol. I. Prague, Paris. HAWLE, I., and CORDA, A. J. c. 1847. Prodrom einer Monographie der bohmischen Trilobiten. Abh. k. Bohm. Gesell. Wiss. 5. HICKS. H. 1871. Descriptions of new species of fossils from the Longmynd rocks of St. David’s. Quart. J. Geol. Soc. London, 27, 399-402, pi. 15-16. iLLiNG,v.c. 1913. Recent discoveries in the Stockingford Shales near Nuneaton. Geol. Mag. Dqc. 5, 10, 452-3. 1916. The Paradoxidian fauna of a part of the Stockingford Shales. Quart. J. Geol. Soc. London, 71, 386-448, pi. 28-38, tables 1-2. LAKE, p. 1935. A monograph of the British Cambrian trilobites. Pt. 9. Palaeontogr. Soc. 197-224, pi. 26-31. LAPWORTH, c. 1898. Sketch of the geology of the Birmingham district. Proc. Geol. Assoc. 15, 313-416, pi. 10-12. LiNNARSSON, G. 1877. Om faunan i lagren med Paradoxides olandicus. Sverig. geol. Unders., ser. C, no. 22, 1-24, pi. 1-2. MOORE, R. c. 1959. Treatise on Invertebrate Paleontology. Part O. New York. PRINGLE, J. 1913. In Summ. Progr. Geol. Surv. for 1912. 71. SMITH, J. D. D. 1962. In Summ. Progr. Geol. Surv. for 1961. 54. SNAJDR, M. 1958. Trilobiti ceskeho stfedniho Kambria. Rozpr. Ustfedniho Ust. geol. 24, 1-280, pi. 1-46. The trilobites of the Middle Cambrian of Bohemia. In Czech, with English summary. WESTERGARD, A. H. 1936. Paradoxides oelandicus beds of Oland. Sverig. geol. Unders., ser. C, no. 394, 17-66, pi. 1-12. 1946. Agnostidea of the Middle Cambrian of Sweden. Ibid., no. 477, 1-140, pi. 1-16. 1950. Non-agnostidean trilobites of the Middle Cambrian of Sweden II. Ibid., no. 511, 1-56, pi. 1-8. J. D. D. SMITH and D. E. WHITE Geological Survey and Museum, Exhibition Road, London, S.W. 7 Manuscript received 23 October 1962 SOME CALAMITEAN PLANTS FROM THE LOWER CARBONIFEROUS OF SCOTLAND by M. CHAPHEKAR Abstract. Material of the plant previously known as Protocalamites pettycurensis Scott is described, and evidence is given for its identity with Archaeocalamites goeppertii Solms. New information has been obtained concerning the nodal structure, the root-bearing stems, and root system of the plant. A new petrified cone species {Protocalamostachys pettycurensis sp. nov.), probably borne on Archaeocalamites goeppertii stems, is described. The name Protocalamites pettycurensis was proposed by Scott (1920) for some petrified stems which he had previously (1901) named Calamites pettycurensis. The stems were from the famous Pettycur Limestone of the Calciferous Sandstone Series (Lower Carboniferous). The stems are characterized by the presence of centripetal wood on the side of the carinal canal towards the pith. Lotsy (1909) distinguished the genus Protocalamites Goebel on this character, although Goebel (1906) had originally used it in a wider sense and included in it a number of species based on compressions and casts. Unpublished investigations on P. pettycurensis by the late Dr. M. Benson have been quoted (cf. Scott 1920, Walton 1940) as indicating that the vascular bundles alternate at the node just as they do typically in Calamites. On the other hand, however. Dr. J. A. R. Wilson after preliminary examination (unpublished) of stems identified as P. pettycurensis from the famous Arran tree-stumps came to the conclusion that there was no alternation at the node. Walton (1940, 1949) states that if both Dr. Benson and Dr. Wilson are right, then P. pettyeurensis is of the Mesocalamites type of Hirmer (1927) and distinct from Archaeocalamites. This presupposes that the Pettycur and Arran tree specimens belong to the same plant. As nothing was known about the morphology or detailed anatomy of P. pettycurensis, it remained a problematical plant. Closely associated with some vegetative remains of this plant in a block from Pettycur have been discovered some fertile parts; these are described in the second part of this paper. Material and methods. The material used in the present investigation is from two localities, viz. the Pettycur Limestone from the Calciferous Sandstone Series of the Lower Carboniferous at Pettycur, Fife, and Arran tree-stumps of the Isle of Arran, again of Lower Carboniferous age. A large number of serial peel sections were prepared from a number of blocks of both Pettycur and Arran tree material by the well-known peel method of Joy, Willis, and Lacey (1956). The serial peel sections were about 33 sections to 1 mm. For the study of cone specimens the peels were taken at about 40 to 1 mm. The peels were mounted temporarily in xylol for examination. [Palaeontology, Vol. 6, Part 3, 1963, pp. 408-29, pis. 59-60.] M. CHAPHEKAR: CALAMITEAN PLANTS 409 DESCRIPTIONS Archaeocalamites goeppertii Solms Plate 59, figs. 1-10, Plate 60, figs. 1-2; text-figs. 1-5 Comparison. There is no doubt that the material previously identified as Protocalamites petlycurensis from the two localities is closely similar except possibly in certain minor details. This will be discussed at the end of the paper. General structure of the internodes. A transverse section of the internode typically shows a ring of primary xylem strands surrounding a pith cavity and a ring of secondary wood developed to the outside of the primary strands. No phloem has been observed. The primary strands are mesarch and the protoxylem is usually represented by carinal canals. The primary medullary rays between the primary strands are continuous with the pith parenchyma; this tissue usually forms a narrow lining around a broad pith cavity. The cortex contains large air spaces. Altogether some thirty specimens with a more or less complete xylem system have been seen; a much larger number of fragmentary specimens have been observed. The specimens from both Pettycur and Arran show great differences among themselves particularly in the size of the pith and the number of primary strands, in the thickness of secondary wood and in the size of the carinal canals ; there is further variation in the size of the primary xylem strands and width of the primary rays. The number of primary xylem strands varies from three in the smallest specimens (text-fig. 1a) to thirty in the largest ones. A number of mainly slender stems (about 12) showing primary structure with little or no secondary thickening have been seen. Most of these have only a small number of xylem strands, commonly three to seven. These slender stems probably represent the ultimate branchlets of the plant. What were prob- ably the main axes of the plant, however, generally contain more than fifteen primary vascular strands and these nearly always have secondary thickening. In some of these specimens the secondary wood is as thick as 1-5 mm. Only one specimen with as many as thirteen xylem strands has been observed without secondary thickening (PI. 59, fig. 1). Most specimens are compressed, but the diameter of the pith has been worked out from the size of the pith circumference; the diameter ranges from OT mm. to 6 mm. The pith is generally hollow, except in some of the slender branchlets where it is solid (text-fig. 1a). In the basal portion of a branch also the pith is typically solid (PI. 59, fig. 2). In one specimen containing eight primary xylem strands (text-fig. 1b) the pith appears at first sight to be continuous, but it seems likely that there was a small space in the centre, but that this space became closed by compression. The thickness of paren- chyma between the primary xylem strands and the edge of the pith cavity varies consider- ably; sometimes there is none and sometimes this tissue is up to about six cells thick. The variation may be due largely to differences in preservation, but there is some indication that the smaller branchlets had relatively more pith parenchyma than the larger stems. There is no evidence that nodal diaphragms occurred, although in one specimen a slight thickening of the pith parenchyma was observed at the node. Most of the variation in the size of the primary strands is due to differences in the amount of centripetal metaxylem. Study of series of transverse and longitudinal sections has shown that the amount of centripetal metaxylem which in the internode typically TEXT-FIG. 1 . Archaeocalamites goeppertii Solms; Pettycur. a. Transverse section of a very young slender twig showing solid pith and three primary xylem strands; H.C., V35645-3, peel 127, x80. b, Trans- verse section of a young twig with a nearly solid pith which is slightly disorganized in the centre; peel 5, X 80. c, One primary xylem strand in the internode showing a carinal canal (cc) and very little centri- petal metaxylem (xi); peel 82, x 210. d. The same strand as in c at the node showing increase in the amount of centripetal metaxylem {xi) and emission of a leaf trace (7/); peel 139, X 210. (Figures b-d from block V35495, H.C.) M. CHAPHEKAR: CALAMITEAN PLANTS 411 consists of one or two layers of tracheids (text-fig. Ic, PI. 60, fig. 2) always increases rapidly as a node is approached from either above or below (text-fig. Id and PI. 60, fig. 1). The carinal canal in longitudinal section is sometimes seen to contain disrupted protoxylem elements (PI. 60, fig. 2). The centripetal metaxylem tracheids show scalari- form-thickening. In a few of the older, more woody specimens some of the bundles apparently lack centripetal metaxylem ; and the carinal canal is sometimes open to the pith cavity. The appearance, however, suggests that this condition is due to poor preservation. The structure of the primary and secondary xylem is identical in almost every detail to that described in Archaeocalamites goeppertii (Solms-Laubach 1897, Walton 1949). The centripetal and centrifugal metaxylem are composed of scalariform elements and the secondary wood tracheids have uni- to multiseriate pitting on their radial walls. The outer part of the wood in larger (woodier) specimens typically has multiseriately pitted tracheids (PI. 59, fig. 7). The ray cells are generally markedly elongated longitudinally as in Archaeocalamites and Calamites. The rays are generally uniseriate and evenly distributed except in the interfascicular wood where they are more abundant and grade into the primary medullary rays. Within a distance of about 0-3 mm. from the pith, however, this distinctive structure of the interfascicular wood disappears. Only a few specimens have been seen in which the cortex is preserved. In the specimen shown in Plate 59, fig. 4, can be seen the epidermis and two to three layers of compact parenchyma. This is followed by the zone of lacunar cortex. The lacunae are traversed by trabeculae which appear to consist of two rows of cells. Some of the small thin walled cells outside the wood (seen near the centre of the photograph) may represent cambial and phloem elements, but it is impossible to prove them as such. There is no evidence of fibrous tissue in the cortex. In Calamites the cortex consists of an inner zone of thin walled cells and an outer zone of thick walled cells. Nodal organization. The leaf traces arise in whorls at the nodes. The number of leaf traces usually equals the number of primary xylem strands in the lower internode. This has been demonstrated by observing close series of transverse sections. Each leaf trace arises from a primary strand, in fact from a core of rather irregularly orientated tracheids which at the level of the node occupy the position of the carinal canal (text-fig. Id). There is thus no carinal canal for a short distance at the node. As the leaf trace passes through the secondary wood it ascends slightly and is surrounded by a sheath of thin walled tissue. Within the leaf trace xylem there is usually discernible a space which somewhat resembles a carinal canal. As seen in cross-section of the trace (i.e. in tangential section through the stem) this space is seen to be situated nearer to the upper side of the xylem strand (text-fig. 2c). The tracheidal nature of the cells bordering the space is not clear from such sections, but is so from sections passing longitudinally through the space (text-fig. 2b). No definite protoxylem has been observed in the leaf trace, but in one or two well-preserved examples seen in cross-section (text-fig. 2c) a group of small tracheids is present at the lower edge of the canal. Similar canals have been described in the leaf traces of Calamites (Williamson and Scott 1894). Although the leaf traces are usually given out singly from the primary xylem strands, in a few instances they have been observed to arise in pairs (text-fig. 2a). The occurrence of such paired leaf traces is more frequent in slender branchlets (text-fig. 2d) although it TEXT-FIG. 2. Archaeocalamites goeppertii Solms; Pettycur. a, Transverse section through a primary strand to show paired leaf traces; B.C. 213, X 50. b, Longitudinal section through a portion of a leaf trace to show the presence of a space surrounded by tracheids and parenchyma; I.C., block C2, peel 7, X 80. c, Tangential longitudinal section of a node near pith region to show transversely cut leaf trace and associated space, px, protoxylem of the leaf trace; (Arran) F.S.C. 856 1/1, X80. d. Transverse section of a slender branchlet to show one single and one paired leaf trace; H.C., V35645-3, peel 67, X 80. E and f. Diagrammatic figures of tangential longitudinal sections through two nodes near the pith showing course of the vascular bundles. If, leaf traces; f, shows forking of one of the bundles, E, B.C. 299-9; f, B.C. 299-15. g. Part of a transverse section through node to show forking of one j of the bundles (on left); I.C. Slide 19, peel 1, x47. h. Transverse section through node to show linking of the nodal ring strand (c) between two bundles; H.C. V35495, peel 177, X47. M. CHAPHEKAR: CALAMITEAN PLANTS 413 occasionally occurs in larger stems. This is not associated with any increase in the number of cauline strands in the succeeding internode. The centripetal metaxylem which, as has already been stated, is considerably more massive at the node than in the internode, is undisturbed by the emission of leaf traces. At the node, strands of primary xylem run across the primary medullary rays linking one primary xylem strand with the next. This linking of the primary xylem strands, constituting a nodal ring, is best seen in tangential longitudinal sections (text-fig. 2 e, f), but it can also be observed by careful study of transverse sections (text-fig. 2h). As can be seen from these figures the linking strand of tracheids often does not run truly horizontally, but somewhat obliquely. These linking or nodal ring strands run from the sides of the vertical strands and connect with both centripetal and centrifugal metaxylem. The course of the primary strands through the node from one internode to the next is generally straight; that is, there is no regular alternation as there is typically in Catamites and Equisetum. In the example represented in text-fig. 2e the bundles of the lower internode are continuous with the corresponding bundles of the upper internode across the node. These bundles are interrupted only by leaf traces which are passing out- wards from the bundles of the lower internode. However, in some nodes observed, local alternation of the primary strands takes place as seen in text-fig. 2f. In this example due to forking of the bundle number 5 of the lower internode there is an increase in the number of bundles in the upper internode. The ‘new’ bundle causes, as it were, displacement of the neighbouring bundles in the upper internode. Thus the bundles 4 and 6 of the lower internode, instead of running straight like bundles 1, 2, and 3, take a slight turn to one side or the other, so accom- modating the additional strand. Text-fig. 2g shows a transverse section of part of a node. The left-hand and middle bundles here have arisen by division of one bundle after the emission of a single leaf trace. This figure shows a difference in size, which is quite typical, between the two bundles resulting from the forking. This interpretation of an occasional alternation by the forking of one of the strands at the node depends upon the correct orientation of the specimen. At first it was rather difficult to determine the correct orientation of sections passing through nodes but careful study has shown that it may be determined from the following characters: (i) The leaf trace is derived from the protoxylem of the internodal bundle below it and ascends slightly as it passes out into the secondary wood. It almost always lies in the median radial plane through this bundle. (ii) The leaf trace attachments are always just below the level of the continuous metaxylem ring at the node. Only four examples (two of which are represented in text-fig. 2f, g) out of twelve studied showed forking of one of the bundles at the node. Thus, the course of vascular bundles at the node in Pwlocalamites pettycurensis is generally straight, although in a minority of nodes local alternation does occur and is essentially associated with an increase in the number of strands from one internode to the next. Branching. The branches arise at the node but above the level at which the leaf traces depart. The branch is inserted in the region of the nodal ring. In the three attached branches investigated the insertion has been associated with two or three of the vascular E e C 1456 414 PALAEONTOLOGY, VOLUME 6 strands of the lower internode of the main stem and has resulted in a loss of one strand from the upper internode. Thus the specimen illustrated in text-fig. 3f, g had seventeen strands in the internode below the branch and only sixteen above. Text-fig. 3f, g shows two sections of a transverse series passing through the junction of the stem and branch. The pith of the branch tapers towards the point of insertion and is continuous with the stem pith by a narrow neck of tissue. The primary xylem of the branch is directly continuous with that of two bundles of the stem. A series of small strands of scalariform elements depart close together and enter the branch base as seen in text-fig. 3f. In both tangential and transverse sections of the base of a branch the secondary wood of the branch is seen to be continuous with that of the parent stem. The pith of the branch at the base is probably typically solid. This is seen in one probably charred specimen (PI. 59, fig. 2) where the delicate cells happen to have been preserved. The pith consists of parenchymatous cells. More usually, however, the base of the branch contains a small cavity in the centre due probably to decomposition of the pith cells. This pith cavity enlarges as the branch emerges and is seen to be surrounded by a ring of primary xylem groups which are separated by medullary rays (text-fig. 3j). The medullary rays are distinct but narrow, consisting of uniseriate or biseriate parenchy- matous cells. The branch from which text-fig. 3j was drawn was represented by a short series of rock sections and could not be traced far enough to demonstrate the appearance of carinal canals. This transition has, however, been demonstrated in an isolated branch base. It may not have been genuinely isolated for it was exposed in transverse section on one side of a block, its appearance at this level (text-fig. 3h) suggesting that it was very near its attachment to a parent stem. As the specimen was traced through a series of peels the originally very small pith expanded and gave way to a large pith cavity. The primary xylem strands enlarged and developed carinal canals, thus giving the typical stem structure (text-fig. 3i). The mode of branching in one observed example of a slender branchlet from Pettycur was somewhat different. The branchlet has five primary xylem strands and bears two laterals. The first lateral arises in the normal way at the node between two of the primary xylem strands (text-fig. 3a). The second branch is given off at a slightly higher level (text-fig. 3c, d). It appears to arise by division of the nodal ring by constriction EXPLANATION OF PLATE 59 Figs. 1-10. Archaeocalamites goeppertii Solms. 1, Transverse section of a young stem with thirteen primary xylem strands showing primary structure only; Arran, F.S.C. 824, x45. 2, Transverse section through basal portion of a branch showing solid pith; Arran, F.S.C. 808 2/2, X 20. 3, Trans- verse section of a rootlet showing dichotomy; Arran, F.S.C. 1702, x45. 4, Part of a transverse section of stem showing lacunar cortex; Arran, F.S.C. 1704, X 30. 5, Transverse section of central part of the main root to show the absence of pith; Arran, F.S.C. 1700, x45. 6, Part of transverse section of the root-bearing stem showing bases of three attached roots (r); Arran, F.S.C. 814 3/1, X 15. 7, Radial longitudinal section through a part of the secondary wood to show multiseriately pitted tracheids; Pettycur, I.C. Slide 3, peel 10, X 240. 8, Longitudinal section of root showing scalari- form tracheids, and towards the outside, imperfectly reticulate tracheids; Arran, F.S.C. 840, x60. 9, Transverse section of a leaf; Arran, F.S.C. 808 2/2, X 20. 10, Transverse section of a part of the root-bearing stem to show stage in an apparent dichotomy in an attached root; Arran, F.S.C. 815 1/3, x30. Note. Plate title: for ‘Asterocalamites' read ''Archaeocalamites'. Palaeontology, Vol. 6 PLATE 59 CHAPHEKAR, Lower Carboniferous Asterocalamites F J TEXT-FIG. 3. Archaeocalamites goeppertii Solms; Pettycur. a-e, Series of transverse sections through a slender stem to show stages in branching, e is orientated at an angle of 90° in relation to a-d, p, parent stem and /, the laterals; H.C., V35645-3; a, peel 82; b, peel 92; c, peel 93; d, peel 99; e, peel 104. All X 29. F, Transverse section through the junction of the stem and branch showing continuity of pith and primary xylem {px), b, branch; If, leaf traces; H.C., V35495, peel 125. g, Section of the same specimen as in F to show the departing branch (6); peel 161, both X 29. h, i, Transverse sections through base of an isolated branch to show stages in its development into a typical branch structure; H.C., V35594; h, peel 55; i, peel 121, x 29. J, Portion of the base of a branch showing primary xylem groups (px) and medullary rays (mr)-, B.C. 199-6, X 80. 416 PALAEONTOLOGY, VOLUME 6 (text-fig. 3b, c) and is not obviously associated with any two of the primary xylem strands although these at the level of the nodal ring where the carinal canals are closed are not very distinct. This mode of branching might be interpreted as a sort of dichotomy. In text-fig. 3e the two departing laterals can be seen. The parent branchlet {p) now has only four primary xylem strands. Alternatively one might regard the three-bundled axis as the continuation of the main axis and p as the second of the two laterals. Leaves. In the majority of specimens the cortex is not preserved and only xylem elements of the leaf traces are seen as they pass through and sometimes just beyond the zone of secondary wood. Just below the node the cortex of the stem shows a number of ribs corresponding to the bases of the leaves which are given off from the node. In no specimen have the in- dividual leaves been traced far from the point of attachment before fading out due to poor preservation ; thus the form of the leaf could not be determined. There is no evidence of a leaf sheath. The preserved basal portions of leaves are apparently simple in anatomy. Plate 59, fig. 9, shows a transverse section of a leaf. It is oval in outline ; in the centre there is a small bundle which consists of xylem surrounded by a thin-walled tissue, probably phloem. Outside the bundle there is parenchyma which becomes smaller-celled and presumably photosynthetic towards the outside. The epidermis is not clear and no stomata have been observed. Root-bearing stems. In general anatomy these stems are like the typical stems except for a few minor differences. The pith is hollow and surrounded by a ring of vascular bundles. The carinal canals, however, are comparatively small. Further, the amount of centripetal wood in the internode is generally no more than a single row of tracheids. In some places it disappears completely, though this is likely to be due to poor preservation, since the carinal canal is often open to the pith cavity. There is a thick zone of secondary wood through the greater part of the specimen, but this diminishes considerably towards the top. Altogether three specimens of root-bearing stems, two from Arran and one from Pettycur, have been seen. The Pettycur specimen was fragmentary and poorly preserved. The Arran specimen represented in the reconstruction (text-fig. 4a) shows four nodes (1, 2, 3, and 4). This reconstruction was drawn from a model based on camera lucida drawings. The sections passing through the nodal regions of this specimen were studied in detail and the data were recorded as shown in Table 1. At nodes 1 and 2 the leaf traces are considerably smaller than in typical stems and are in fact difficult to detect in the secondary wood. Text-fig. 5c shows a leaf trace cut nearly transversely from node 2. At node 3 the leaf traces are of more normal size (text- fig. 5d). At node 4 the specimen is somewhat squashed and the preservation is very poor, but there is no doubt that the diameter here is genuinely less and the amount of wood much smaller. Only one root is clearly seen (text-fig. 4b), but it is possible that there was a second feebly developed root (text-fig. 4c). The leaf traces are unfortunately indistinct, the two traces to the right in text-fig. 4c probably represent leaves. From this it appears that nodes 1 and 2 may probably represent the underground portion of the rhizome as in them the leaf traces are very small and the roots are well developed. Nodes 3 and 4, where the roots are poorly developed and the leaf traces (at PEEL MO'S. 1 FIG. no's. BLOCK TEXT-FIG. 4. Archaeocalamites goeppertii Solms; Arran, a, Reconstruction of the root-bearing stem showing four nodes (1, 2, 3, and 4) and root e at node 2. This figure also shows parts of the block peeled and position of the sections figured (text-fig. 4b-e and text-fig. 5a-q). b, c, Transverse sections through part of node 4, to show one root base in b and three small traces in c, one of which probably represents a root, the others, leaves; b, H.M.Pb., 3504-r/ (upper), peel 171; c, H.M.Pb., 3504-rf (upper), peel 41. d. Transverse section through a part of node 3 showing two large leaf traces; H.M.Pb., 3504-rf (lower), peel 1 12. e. Transverse section through part of node 2 showing one root trace and five very small leaf traces; H.M.Pb., 3504-c/ (upper), peel 112. All x9. 418 PALAEONTOLOGY, VOLUME 6 least in node 3) are larger, probably represent the basal portion of an aerial stem. The curvature of the whole specimen is probably to some extent genuine and not merely an accident of preservation. The woody cyhnder is thickest in the middle region. This specimen is comparable with the curved pith casts of branching rhizomes of Calamites (Seward 1898, p. 323, fig. 82), which probably grew underground giving aerial shoots and adventitious roots. It is also comparable with the curved, root-bearing stem base with its thick zone of secondary wood depicted in Eggert’s (1962) reconstruc- tion of a calamite based on American material. TABLE 1 Nodes Internodal length Approx, diameter of stem {woody cylinder) at node Approx, thickness of wood in the internode No. of primary xylem strands in the internode No of roots Node 1 \ 5-5 mm. \ 2 1 1-6 cm. j 1-4 mm. 15 ( Node 2 7-9 mm. > ( 5 1-95 cm. 1 IT mm. 15 Node 3 1 0-85 cm. 4-9 mm. 0-5 mm. I15(?) ( 1 (?) Node 4 1 [ 3 0 mm. 1 15 (.9 \ 2(?) Roots. The roots described here are all from the Arran material. Only a few poorly preserved specimens have been observed in Pettycur material. The roots are given off in irregular whorls from the nodes of certain stems (PL 59, fig. 6). The root traces arise at approximately the same level as leaf traces and are given off from the centrifugal metaxylem either in between the two primary xylem strands (text-fig. 5b) or opposite the primary xylem strand (text-fig. 4e). The primary xylem of the root can be traced through the zone of secondary wood to a point near the primary xylem of the stem, but it does not always join up with it; rather it tends to get lost in the innermost part of the secondary wood. This indicates that the roots were sometimes formed after secondary thickening had started in the stem. The roots are mostly decorticated and only one or two examples have been seen in which the cortex appears to be lacunar in structure. In the specimen represented in text-fig. 4a the root marked e was traced for a distance of about 7 mm. and the following description is based mainly on this specimen. Other roots on this stem could be traced for only comparatively short distances and some hardly beyond the secondary wood. The root e measures in its decorticated condition about T8 mm. in diameter, where it emerges from the wood of the stem. There is no evidence of pith, the centre being occu- pied by apparently tracheidal (metaxylem) cells. At some levels a small cavity is present in the middle which is probably due to the breaking down of tracheidal cells. In thick peels these cells often show a clear indication of scalariform thickening on their longitudinal walls. The tracheids of the metaxylem core are surrounded by a ring of probably five to six protoxylem groups and a thick zone of wood (PI. 59, fig. 5). TEXT-FIG. 5. Archaeocalamites goeppertii Solms; Arran, a-q, Continuation of the series of sections from the specimen shown in text-fig. 4a. a, b, Transverse sections through a portion of node 1 showing two root traces and a few leaf traces; A, H.M.Pb., 3504-r/ (upper), peel 27; b, H.M.Pb., 3504-r/ (upper), peel 3, X 9. c. Leaf trace at node 2 cut transversely; H.M.Pb., 3504-r/ (lower), peel 1 12, x 50. d. Leaf trace at node 3 cut transversely; H.M.Pb., 3504-r/ (lower), peel 206, X 50. e. Portion of the main root e showing medullary rays (mr) in the peripheral region of the root; H.M.Pb., 3504-t/l, peel 146. X 80. F-Q, Series of transverse sections showing stages in the branching of the root e. The numbers 1-5 represent the rootlets; H.M.Pb., 3504-r/l ; f, peel 7; g, peel 36; h, peel 68; i, peel 128; J, peel 137; K, peel 146; l, peel 152; m, peel 165; n, peel 168; o, peel 181; p, peel 189; q, peel 206. All x9. 420 PALAEONTOLOGY, VOLUME 6 Unfortunately the protoxylem groups are not at all distinct and the number could not be determined exactly. Text-fig. 5g-q shows the stages in the branching of the root e. It gives off five rootlets or secondary roots, which arise in succession and come off at five different points apparently opposite protoxylem groups in the main root. The rootlet number 1 in text-fig. 5g is given off while the main root is still attached to the stem. The rootlets are smaller in size than the main root. The protoxylem groups, which again are usually rather indistinct, are few in number varying from two to four. The amount of wood is less. Two cases of apparent dichotomous branching in roots have been observed; one in an attached main root and the other in an isolated rootlet. Plate 59, fig. 10, shows a main root still attached to the stem just before the dichotomy takes place. At this level the primary xylem (p) has divided into two nearly equal portions. A little higher, the secon- dary wood becomes organized around the two primary xylems and the two roots are eventually separated. The true nature of the branching, however, is not perfectly clear because of the plane of section. Plate 59, fig. 3, shows a stage in the dichotomy of an isolated rootlet which, on the evidence of structure (the size and pitting of the tracheids and the form of the primary xylem) belongs to the same plant. The secondary wood of the root consists of tracheids which are arranged in regular radial series. Rays have only been observed in main roots. Text-fig. 5e shows a portion of the transverse section of the main root e. The rays are present in the peripheral zone of the wood and can be distinguished from the tracheids by their larger size and thinner walls. No longitudinal sections were obtained from this particular root, but the tracheids of the outer zone of wood when these are somewhat obliquely sectioned, sometimes show a clear indication of multiseriate pitting on their radial walls. The roots seen in longitudinal section were all more slender than root e and only niore or less scalariform tracheids were observed (PI. 59, fig. 8). As with the stem, the first formed secondary wood apparently consisted of uniseriately pitted or somewhat reticulate tracheids, multiseriate pitting occurring in the later formed wood. This has been observed by close examination of slightly obliquely sectioned portions of the wood in attached root e. Discussion. The name Archaeocalamites (or Asterocalamites) is now generally employed for calamite-like fossils mainly of Lower Carboniferous age which differ from the typically Upper Carboniferous Calamites chiefly in the lack of alternation of the vascular bundles at the node. Other characters that appear to differentiate Archaeocalamites from Calamites are the dichotomous forking of the leaves, and the presence of centripetal metaxylem. Leistikow (1959) has given reason for regarding Archaeocalamites Stur as the legitimate name, rather than Asterocalamites Schimper or Asterocalamites Zeiller. Some of the species are based on pith casts or compressions, e.g. A. radiatus (Bgt.); others are based on petrified material, e.g. A. goeppertii Solms. Petrified material referred to A. goeppertii has been studied by a number of authors. In 1852 Goeppert described some details about the secondary wood. Solms-Laubach (1897) described the internodal structure of the primary and secondary xylem and Walton (1949) described the nodal structure. Renault (1893-6) described two species under the names Bornia esnostensis and B. latixylon from the Lower Carboniferous of Esnost in France. These species were M. CHAPHEKAR: CALAMITEAN PLANTS 421 later referred to Asterocalamites and grouped under the family Asterocalamitaceae (Hirmer 1927). These species of Renault’s appear to be genuinely different from Archaeocalamites goeppertii, for A. esnostensis had much shorter ray cells and wood tracheids with only one or two rows of bordered pits; A. latixylon was a larger plant and had wider wood rays. Walton (1949) in his investigations into a Scottish specimen of A. goeppertii, briefly compared this plant with Protocalamites pettycurensis and drew the following distinc- tions; (i) In Archaeocalamites the vascular strands always run straight through the node whereas, according to earlier (unpublished) investigations, there was sometimes alterna- tion in Protocalamites. (ii) In Archaeocalamites the centripetal metaxylem diminished (sometimes to nothing) in the internode and increased at the node, whereas in Protocalamites the centripetal metaxylem was always massive. (iii) Rays were evenly distributed in the wood of Archaeocalamites, whereas in Proto- calamites the interfascicular wood contained relatively more rays. (iv) The pitting of the wood tracheids was regularly multiseriate in Archaeocalamites and often uniseriate in Protocalamites. The present investigations into P. pettycurensis have shown that the structure of the xylem system of the internode including the structure of the secondary wood is essen- tially similar to that of A. goeppertii. The nodal structure of P. pettycurensis shows that the vascular strands pass straight through without alternation, but that in about one-quarter of the nodes investigated in detail, local alternation does occur and is always associated with the interpolation of an additional strand in the upper internode. Such a local alternation has not been pre- viously reported in A. goeppertii, but then few nodes have been studied in detail. It is of interest to note that a few specimens of Lower Carboniferous pith casts have been reported, one in the Goeppert Collection of the Breslau Museum mentioned by Seward (1898), and Feistmantel’s specimens from Rothwaltersdorf (1873, p. 491, pi. xiv, figs. 3, 4) which do show a local alternation at the node associated with an increase in the number of bundles from one internode to the next. The same thing has been observed in a specimen labelled A. scrobiculatus from the Calciferous Sandstone Series in the Kidston Collection in the Geol. Survey Museum (specimen no. 3190). With regard to other supposed distinctions between A. goeppertii and P. pettycurensis mentioned by Walton, the structural distinction between the fascicular and interfascicular wood has been found to disappear in a radial distance of about 0-3 mm. in Protocala- mites. Thus only in small specimens (i.e. with little secondary wood) is there any general distinction between fascicular and interfascicular wood. Many of the specimens in both Pettycur and Arran tree material are small. Similarly, in small specimens the wood tracheids may not have attained their characteristic multiseriate pitting since there is gradation from the scalariform elements of the centrifugal metaxylem. The outer part of the secondary wood in all larger specimens characteristically has multiseriately pitted tracheids. The remaining supposed distinction is the marked decrease in the amount of centri- petal metaxylem in the internode of Archaeocalamites. This has now been demonstrated 422 PALAEONTOLOGY, VOLUME 6 in Protocalamites. It is perhaps especially marked in one root-bearing specimen, and it is therefore noteworthy that Walton’s petrified specimen of A. goeppertii was probably a root-bearing stem. There is thus no good case for continuing to separate goeppertii and P. petty curensis. The present investigations have added to our knowledge of A. goeppertii, chiefly with regard to the root-bearing stems and root system. The root system is simpler than that of Calamites, for the main roots are protostelic and probably no more than hexarch, the lateral rootlets have two to four protoxylem groups. There is some evidence of occasional dichotomies in the roots. They compare somewhat with the roots described by Renault (1893-6) in A. esuostensis; here the roots were apparently regularly dichotomous, but there was a pith and the number of primary xylem groups was more (about 8). Protoealamostachys pettycuremis sp. nov. Plate 60, figs. 3-7 ; text-figs. 6, 7 Diagnosis. Cone axis slender, with three mesarch xylem strands and six longitudinal series of sporangiophores; three pairs of sporangiophores in each whorl, each pair lying opposite an axial xylem strand. Sporangiophores approx. 1-5 mm. long and 0T5 mm. in diameter. Sporangia approx. 0-9 mm. long and 0-4 mm.wide. Spores about 38 p in diameter. Sporangium wall cells with peg-like thickenings on the anticlinal walls. Type specimens. The description is based on twelve specimens clustered in one small portion of the block H.C. V35645. Horizon. Calciferous Sandstone Series, Lower Carboniferous. Locality. Pettycur, Fife, Scotland. Description. The cones described here are all from Pettycur material. Altogether about a dozen specimens have been seen. They are closely associated with vegetative remains of Archaeo calamites goeppertii. All the cones were found clustered in one small portion of a block. The preservation of the material is variable; some specimens are quite well preserved. The cones were probably borne at the extremities of slender branchlets (text -fig. 6m). These contain three mesarch primary xylem strands without secondary thickening. EXPLANATION OF PLATE 60 Figs. 1-2. Archaeocalamites goeppertii Solms; Pettycur. 1, Radial longitudinal section of a primary xylem strand at node showing massive centripetal metaxylem (cpx), centrifugal metaxylem {cfx), protoxylem ipx), and a leaf trace {If). 2, The same strand as in 1 in the internode to show one or two layers of centripetal metaxylem {cpx) towards the pith {p), carinal canal (cc) with disrupted protoxylem elements and the centrifugal metaxylem {cfx) \ Pettycur, I.C. Slide 21 ; 1, x 160; 2 X 200. Figs. 3-7. Protoealamostachys pettycurensis sp. nov.; Pettycur. 3, Oblique longitudinal section of part of cone axis passing through a sporangiophore to show attachment of two of its four sporangia; H.C., V35645-3, peel 22, x 40. 4, Transverse section of cone axis with three mesarch xylem strands and two of its six sporangiophores; several sporangia are seen near it; H.C., V35645-2, peel 18, x40. 5, Spore, x900. 6, Transverse section of a sporangiophore; H.C., V35645-2, peel 47, X 160. 7, Part of the sporangial wall in transverse and surface sections; H.C., V35645-3, peel 66, X 200. Note. Plate title: for ^ Asterocalamites’’ read 'Archaeocalamites'. Palaeontology, Vol. 6 PLATE 60 CHAPHEKAR, Lower Carboniferous Asterocalamites and Protocalaniostachys M. CHAPHEKAR: CALAMITEAN PLANTS 423 Several branchlets of this kind, mostly decorticated, and differentiated into nodes and long internodes, are closely associated with the cones. The continuity between one such stem and a typical cone axis has been demonstrated by tracing the specimen through serial peels. The stem is decorticated but as the cone axis is approached (about T5 mm. above an observed node) the cortex becomes preserved. The preservation of the cortex in the cone axis may be due to the fibrous nature of some of the cortical cells as seen in one or two of the better-preserved specimens, or perhaps only to the presence of rather closely arranged sporangiophores which might protect the cortical tissues. The longest portion of cone axis observed was 9 mm. but no definite cone apex has been seen. The diameter of the fertile region of the cone is approx. 4 mm. The cone consists of a slender axis bearing superposed whorls of sporangiophores. There are six sporangiophores in each whorl and each sporangiophore carries four sporangia. Ax/s. The diameter of the cone axis based on six specimens is approx. 0-8 mm. As seen in transverse section (PL 60, fig. 4) the cone axis is somewhat triangular in shape and consists of a solid pith surrounded by three mesarch primary strands without secondary thickening. Unfortunately the cellular details of the cortex cannot be satisfactorily distinguished. Study of a few transverse and longitudinal sections through the cone axis, however, has shown that in small patches where the cortical cells are preserved, they appear to be fibrous in nature. It is not clear whether the rest of the cortical cells are fibrous or not. The three xylem strands follow a straight course through the axis without any division or anastomoses, except that at the ‘nodes’ where the sporangiophores are attached, each strand gives off two weak strands which supply the sporangiophores. There is no nodal ring. Each xylem strand possesses a protoxylem canal surrounded by centripetal metaxylem towards the pith side and centrifugal metaxylem on the outside. The tracheids all show scalariform thickening. Arrangement of sporangiophores. The sporangiophores are the only appendages on the cone axis. They are arranged in longitudinal series and are attached in pairs. The pairs correspond in position to the three xylem strands of the axis. Text-fig. 6l shows a diagrammatic reconstruction of a transverse section of a cone axis with three pairs of sporangiophores in a whorl. The distal ends of the sporangiophores are shown with two of their four pedicels bearing sporangia. The distance between successive whorls of sporangiophores is constant and about 1 mm. This is demonstrated well in the series of slightly oblique longitudinal sections passing through a cone represented in text-fig. 6a-k. Text-fig. 6a shows five (I-V) of the eight superposed whorls of sporangiophores observed in this specimen. Text-fig. 6b-k shows only whorls I and II and demonstrates the attachment of the sporangiophores. Whorl II may be taken as a typical example. Text-fig. 6b shows three sporangiophores (1,2, and 3) of whorl II, out of which 1 and 2 join together (text-fig. 6c) and thus represent the bases of one pair. Next the sporangiophore 3 joins into the axis (text-fig. 6d) and just beyond (text-fig. 6e-g) departs the other member of its pair 4. Lastly the bases of the third pair of sporangiophores (5 and 6) are seen (text-fig. 6h-k). The sporangiophores. The sporangiophores are long (1-5 mm.) slender, stalk-like struc- tures about 0T5 mm. in diameter. The stalk of the sporangiophore consists of fairly L TEXT-FIG. 6. Pwtocalamostachys pettycurensis sp. nov. A, Oblique longitudinal section of portion of ! cone to show five (I-V) superposed whorls of sporangiophores. The attachment of sporangiophores in | the first two whorls is illustrated in figures b-g. In these figures only sporanglophore bases are shown; | H.C., V35645-2; a, peel 79; b, peel 85; c, peel 90; d, peel 91 ; E, peel 95; F, peel 97; G, peel 100; H, peel li 105; I, peel 106; J, peel 108; k, peel 110. All x 18. l. Diagrammatic reconstruction of the cone axis to j show three pairs of sporangiophores opposite three xylem strands. The distal ends of sporangiophores are shown with two of its four pedicels bearing sporangia, X 25 approx, m. Diagrammatic reconstruc- I tion of a cone to show continuity between the slender stem and cone axis; ;< 10 approx. j M. CHAPHEKAR: CALAMITEAN PLANTS 425 thick-walled cells (PI. 60, fig. 6) within which is generally present a small space and a strand of tracheidal cells. In some specimens the xylem strand appears to have the form of a ring; the space it contains may be equivalent to the space present in the leaf traee strands of Archaeocalamites. At the distal end of the sporangiophore there are attached four branches or sporangial pedicels. There is some indieation (text-fig. 7a) that these may represent two successive dichotomies of the sporangiophore stalk, but the preservation is such that it has not been possible to determine this with certainty. The four pedieels curve towards the axis of the cone and terminate in single sporangia. This general form has been demonstrated from both transverse and longitudinal series of sections through the sporangiophores. An obliquely transverse series is represented in text-fig. 7a-f. Text-fig. 7a shows the distal end of the sporangiophore with its four sporangial pedicels a, b, c, and d which are arranged in a cross-wise manner. In text-fig. 7b the stalk of the sporangiophore is seen in the centre surrounded by the four pedicels {a, b, c, and d) and parts of two spor- angia. Text-fig. 7c shows the attachment of pedicel a to its sporangium and text-fig. Id, f the attachment of pedicels b and c respectively. Pedicel d is seen to continue into sporangium d in text-fig. 7b, c. Text-fig. 7e shows the stalk of the sporangiophore with four sporangia and pedicel c. In text-fig. 7f the attachment of the sporangiophore to the cone axis is seen. In the longitudinal series of seetions, again the pedicels a, b, c, and d with their re- speetive sporangia show the cross-wise arrangement (text-fig. 7j-s). Text-fig. 7j shows pedieel a in attachment to its sporangium ; on the right is seen part of the cone axis and on the left the base of pedicel a near its attachment to the end of the sporangiophore stalk. The attachments of sporangia b and c are seen in the more medial sections illustrated in text-fig. 7k-n. The sporangia appear to be somewhat tueked in by the pedicels which are attached to the sporangia on the side away from the main shaft of the sporangiophore (PI. 60, fig. 3). The remaining sporangium <7 and its attachment is seen in text-fig. 7o-s. The pedicel as seen in transverse section (text-fig. 7h) shows a distinct epidermis consisting of rather large, thick-walled cells. In the centre there is present a minute strand of tracheids. The other tissues are not clear. Sporangia. The sporangia are about 0-9 mm. long and 0-4 mm. wide. The sporangial wall consists of cells slightly elongated in a direetion parallel to the sporangial axis. As seen in transverse seetion (PI. 60, fig. 7) the inner walls of these cells are uniformly thickened and the outer walls are thin. The thickness on the anticlinal walls appears to vary considerably; this is mainly due to the presenee of projections on the walls which are seen in tangential seetions through the cells (text-fig. 7i and PL 60, fig. 7). The pro- jections on the walls are swollen at the tips, sometimes with two lobes. In some a light spot or a small hole can be seen in the ‘head’ which suggests that the projections are really folds of the sporangial wall. Spores. The spores appear to be very uniform. They vary in size only from 34 p to 44 p. The average worked out from fifty individually measured spores is 38 p. The spore wall appears to be plain or possibly very finely punctate. There is a well-marked triradiate ridge (text-fig. 7g and PI. 60, fig. 5) the rays of which are 9 pin length. TEXT-FIG. 7. Protocalamostachys pettyciireiisis sp. nov. a-f, Oblique transverse series of sections through sporangiophore to show the attachment and form of sporangia; H.C., V35645-2; a, peel 39; B, peel 50; c, peel 51; d, peel 70; e, peel 75; f, peel 84. All x30. g. Spore, x810. h. Transverse section of a pedicel showing epidermis and minute strand in the centre; H.C., V35645-3; peel 61, X 80. I, Part of sporangial wall in surface section to show ridges of thickening; H.C., V35645-3; peel 63, x 510. j-r. Longitudinal series of sections through a sporangiophore to show the attachment j| of sporangia (a, b, c, and d); H.C., V35645-3; J, peel 5; k, peel 17; l, peel 21 ; m, peel 26; N, peel 30; ■ o, peel 40; p, peel 51 ; q, peel 56; r, peel 61 ; s, peel 64. All x 30. i M. CHAPHEKAR: CALAMITEAN PLANTS 427 Discussion. This new cone possesses the general characteristics of the genus Proto- calamostacliys which was erected for a single cone from Arran tree material by Walton (1949). The general features of the Pettycur cones fit very well the generic diagnosis published by Walton except for spore size; but this is a character which should probably not be regarded as an important generic criterion. The Pettycur cones however differ in a number of respects from the type species of the genus, Protocalamostachys arranensis. They are therefore made the basis of a new species P. pettycurensis sp. nov. The follow- ing are the main points of contrast between the two species. The axis of the new cone contains only three mesarch primary xylem strands without secondary thickening, whereas in P. arranensis the cone axis contains six mesarch xylem strands with a little secondary xylem on the outside. The distance between successive sporangiophore whorls in the Pettycur cone is 1 mm. whereas in P. arranensis this distance is stated to be only 0-8 mm., although in view of the size of the sporangia, it is difficult to see how the sporangiophores could have been quite so closely arranged. In the Pettycur cones there are six longitudinal series of sporangiophores which are attached in pairs, whereas in the Arran cone there are ten to eleven longitudinal series of sporangiophores which are usually attached in pairs but sometimes singly. The sporangiophores in P. pettycurensis are 1-5 mm. long and 0T5 mm. in diameter. The corresponding measurements in P. arranensis are 2-7 mm. long and 0-22 mm. in diameter. In both species the sporangia are borne at the extremities of four pedicels which are arranged in a cross-wise manner. The sporangia in P. pettycurensis are 0-9 mm. long and 0-4 mm. wide whereas in P. arranensis they are 1-2 mm. long and 0-65 mm. wide. The structure of the wall of the sporangium in both species is similar. The spores in P. pettycurensis are 38 ju. in diameter while those in P. arranensis are given as 66 p. The cones of Archaeocalamites from compressed material have in the past been referred by different authors to various genera, e.g. Pothocites grantonii (Paterson 1841), AsterophyUites spaniophyllus (Feistmantel 1873) and Archaeocalamites radiatus (Stur 1875-7). Kidston (1883o, b) employed the name Pothocites for all these cones. Walton (1949) compared Protocalamostachys arranensis with Pothocites grantonii and concluded that the petrifaction and the compression cones were probably closely similar in general structure and in size. Protocalamostachys pettycurensis probably differs from Pothocites in the same ways as it does from Protocalamostachys arranensis, i.e. mainly in being smaller, having fewer sporangiophores in a whorl, and fewer vascular strands in the axis. The cone named Bornia radiata (Renault 1893-6) differs from Protocalamostachys in having its sporangia sessile on what appear to be peltate sporangiophores. It is possible that Renault’s figures may be misleading in this, for it is rather surprising that a Lower Carboniferous cone should have sporangiophores almost exactly like those of a modern Equisetum. The evidence that the petrified form of the cone of Archaeocalamites was Protocala- mostachys is now considerable. Both Protocalamostachys arranensis and P. pettycurensis are probably similar in general organization to Pothocites grantonii which is a fructifica- tion borne on leafy stems of the Archaeocalamites type. Moreover, it is now known that vegetative remains indistinguishable from A. goeppertii are frequent at both the Arran and Pettycur localities. The cone axis in both species of Protocalamostachys is similar in general structure to 428 PALAEONTOLOGY, VOLUME 6 slender axes of Archaeocalamites. There is thus good reason to believe that cones of the Pwtocalamostachys kind were borne on Archaeocalamites stems. The association of stems identified as A. goeppertii with one kind of cone at the Arran locality and with a specifically different, though generically similar, cone at Pettycur raises the question as to whether the vegetative remains at the two localities do in fact represent the same natural species. Careful comparison has been made between the material from the two localities and no differences worthy of consideration as criteria for specific separation have been discovered. The following two differences should be noted for what they are worth. (i) The pith cells appear to be somewhat larger in Arran specimens as shown in Table 2. However, the size of the pith cells varies enormously, both in individual specimens and also from one specimen to another. In the latter case the variation apparently depends to some extent on the size of the specimen as determined by the number of primary xylem strands. TABLE 2 Pettycur specimens Arran specimens No. of Size of No. of Size of primary largest primary largest strands pith cells strands pith cells 6 45 (X 1 50 p 1 50 p. 9 75 p 8 30 p 10 15 p 8 35 p 11 90 p 9 65 p 12 80 p 12 60 p 13 15 p 12 65 p 14 145 p 18 75 p 26 195 p 23 15 p 23 90 p (ii) The smallest stems seen in Arran material have 7 vascular strands, whereas several examples with 6, 5, 4, and 3 have been noted from Pettycur, several specimens with 4 and 3 strands having been seen in close association with the cones at Pettycur. It is known that amongst living plants different species and sometimes even different genera may be virtually indistinguishable in anatomy. Thus the genera Qitercus and Lithocarpus belonging to Fagaceae are quite distinct on floral characters but the wood of a number of tropical species of Quercus is virtually indistinguishable from that of Lithocarpus. Acknowledgements. I am indebted to Professor J. Walton for the loan of slides and cut blocks, to the Trustees of the British Museum (Natural History) for the loan of material, and to the Principal, Royal Holloway College, for the loan of slides from Benson’s Collection. I am grateful for the inspiration of Dr. K. L. Alvin who supervised my work. I should also like to thank Dr. K. I. M. Chesters for showing me slides in the British Museum (Natural History). Location of collections. The slides and peels are preserved at the following Institutions: Hemingway Collection (H.C.), British Museum (Natural History). Benson Collection (B.C.), Royal Holloway College, London. Figured Slide Collection (F.S.C.), Botany Department, University of Glasgow. M. CHAPHEKAR: CALAMITEAN PLANTS 429 Hunterian Museum Palaeobotanical Collection (H.M.Pb.), University of Glasgow. Departmental Collection (I.C.), Botany Department, Imperial College, London. EGGERT, D. A. 1962. The ontogeny of Carboniferous arborescent Sphenopsida. Palaeontographica, B 110, 99. FEiSTMANTEL, o. 1873. Die Kohlenkalkvorkommen bei Rothwaltersdorf. Z. Deutsch. Geol. Ges. 25, GOEBEL, F. 1906. Protocalamariaceae Potonie 1899. Centralbl. Miiieralogie, 241. GOEPPERT, H. R. 1852. Foss. Flora des Uebergangsgebirges. Nov. Act. Leop. Carol. 64, Suppl. (1852), HiRMER, M. 1927. Handbuch der Paldobotaiiik, 1. Miinchen und Berlin. KiDSTON, R. 1883fl. On the affinities of the genus Pothocites, &c. Trans. Bot. Soc. Edin. 16. 28. 1883Z>. On the affinities of the genus Pothocites, &c. Ann. Mag. Nat. Hist., ser 5, 11, 297. LEISTIKOW, K. u. 1959 Archaeocalamitcs und Archaeocalamitaceae. Taxon, 8, 48. LOTSY, J. p. 1909. Vortrdge iiber botanische Stammesgeschichte, b. 2, Cormophyte Zoidogamia. PATERSON, R. 1841. Description of Pothocites grantoni, a new fossil vegetable, &c. Trans. Bot. Soc. Edin. 1, 45. RENAULT, B. 1893-6. Bassin Houiller et Permien d’Epinac. Etudes des Gites Min. de la France, Paris. SCOTT, D. H. 1901. Note of a primitive type of structure in Calamites. Ann. Bot., Lond. 15, 773. 1920. Studies in Fossil Botany, 1. London. SEWARD, A. c. 1898. Fossil Plants, I. Cambridge. SOLMS-LAUBACH. 1897. Uber die in den Kalkstein des Kulm von Glatzisch Falkenberg, &c. Bot. Zeits. 219. WALTON, J. 1940. Introduction to the Study of Fossil Plants. London. 1949. On some Lower Carboniferous Equisetineae from the Clyde area. Trans. Roy. Soc. Edin. WILLIAMSON, w. c. and SCOTT, D. H. 1894. Eurther observations on the organization of the fossil plants of the Coal Measures. Pt. I. Calamites, Calamostachys, and Sphenophyllum. Phil. Trans. Roy. Soc. Lond. 185, 863. 1895. Ibid. Pt. II. The roots of Calamites. Ibid. 186, 703. REFERENCES 463. 108. 61. Manuscript received 18 September 1962 M. CHAPHEKAR Department of Botany, Imperial College, London, S.W. 7 C 1456 Ff THE OSTRACOD SPECIES O RTHONOTACYTHERE INVERSA (CORNUEL) AND ITS ALLIES IN THE SPEETON CLAY OF YORKSHIRE by P. KAYE Abstract. The ostracod species Orthonotacythere inversa (Cornuel) is described and split into three subspecies which form an evolutionary sequence with O. diglypta Triebel. Other allied species are described from the Speeton Clay. The genus Orthonotacythere was proposed in 1933 by Alexander for one Cretaceous species. In 1934 he compared the shell characters and hinge structure of the genera Monoceratina and Orthonotacythere and expressed the opinion that the latter was a derivative of the former. In this later paper he described a further three species of the genus. Though in his initial paper he mentions the likeness to Cytheropteron, in his 1934 paper Alexander dwells at length on the similarity between Orthonotaeythere and Monoceratina. He was also of the opinion that the hinge in Orthonotacythere had developed from that of Monoceratina by the development of crenulations along the median bar of the left valve, and the formation of terminal teeth in the right valve, and he thought that early forms of Orthonotacythere might be found in the Jurassic and Lower Cretaceous with a more primitive hinge than the type species, and a greater similarity to Monoceratina. Though similarities to Monoceratina cannot be denied, it is found that the earlier members of this genus are in fact less akin to Monoceratina than are the later ones, and that the species of Orthonotacythere found in the Lower Cretaceous and Upper Jurassic have a shorter margin and a costate ornament. This costate ornament gradually changes to the characteristic tuberculate ornament of the type species as the genus is traced from the Upper Jurassic to the Upper Cretaceous. Upper Jurassic species of Orthonotaeythere still have a hinge with strong terminal teeth and crenulate elements. Instead of the dominant tuberculate ornament charac- teristic of Orthonotacythere s.s. these forms show a pattern of longitudinal ridges and reticulations. The posterior margin is angled at mid-height rather than at the dorsal margin as in the type species of Orthonotacythere. Some species, which tend to be con- fined to the Western European Lower Cretaceous, were grouped together by Neale (1960) into the subgenus Acrocythere, with Orthonotacythere hauteriviana Bartenstein 1956 as type species. On further examination of this species it was found that the median sulcus, typical of the genus Orthonotacythere s.s. is either absent or only weakly developed, and that the shape differs markedly from other species of this genus, being rather subrectangular and not ovate. On these further grounds it is thought advisable to raise the subgenus Aerocythere to full generic status, as suggested by Malz in 1961. In the Hauterivian and Barremian beds at Speeton there are a number of forms of Orthonotacythere s.s. constituting a group differing in ornamentation from the type [Palaeontology, Vol. 6, Part 3, 1963, pp. 430-9, pi. 61.] P. KAYE: ORTHONOTACYTHERE INVERSA 431 species. These species have an ornament composed of longitudinal and vertical ridges and reticulations on the lateral surface. The median sulcus, however, is smooth and very prominent, and the shape is more akin to Orthonotacythere s.s. than to Acrocythere. The group is centred around O. inversa (Cornuel) and O. diglypta Triebel. It has been found necessary to subdivide O. inversa into three subspecies which form a chronological and evolutionary sequence. O. diglypta from the Hauterivian and Lower Barremian shows a complex pattern of vertical and longitudinal ridges and a marked reticulation. The earlier subspecies of O. inversa show a relationship to O. diglypta but there is a tendency towards the loss of the vertical ridges, and reticulation and an increase of tuberculation as one ascends the sequence. In the later subspecies the ornament is more dominantly tuberculate than costate and tends towards that of the type species. The genus Orthonotacythere can therefore be subdivided into two major groups. One group, which has a dominant costate and reticulate ornament, seems to be confined to Upper Jurassic and Lower Cretaceous strata, whilst the other group, which has a dominantly tuberculate ornament and differing shape, is confined to the Upper Cre- taceous and Tertiary, having evolved from the earlier one during the Lower Cretaceous with O. inversa as an intermediate form. This hypothesis, however, is not as simple as it first appears. Certain species of Orthonotacythere from the Valanginian, Berriasan, Middle Purbeckian, and Kim- meridgian either form part of the sequence or are related to the O. diglypta group. The costation and reticulation of these forms is increasingly complex as the sequence is traced into older strata. Other species, however, show a divergence from the sequence. O. ramulosa (Sharapova) from the Upper Hauterivian and Lower Barremian possesses a tuberculate ornament with strong reticulation but poorly developed costation. In this respect, and in its shape, it is very close to the type species. O. anglica Neale from the Lower Hauterivian has the intercostal areas smooth, and shows subdued tuberculation. The ribs are rather poorly differentiated when compared with the other species, and it also differs in the absence of a marked antero-dorsal tubercle. It is oval in shape and very globose. The pattern of ribbing is more complex than in O. inversa and is similar to that of O. diglypta. It is possible, therefore, that it shows the same kind of relationship to O. diglypta as O. blanda sp. nov. does to O. inversa s.s. The subspecies of O. inversa and the other related species can, in view of their re- stricted range, be used for correlation, at least over short distances. A chart of the postulated evolutionary sequence is reproduced as text-fig. 1, whilst the variation in the ornament is shown on text-fig. 2. SYSTEMATIC DESCRIPTIONS Orthonotacythere inversa (Cornuel) 1848 ?Cythere inversa Cornuel 1848, p. 244, pi. 1, figs. 12-14. Orthonotacythere inversa (Cornuel); Stchepinsky 1954, p. 496, pi. 22, fig. l.text pi. 4, figs. 26 a-d. Orthonotacythere inversa (Cornuel); Deroo 1956, p. 1516, pi. 3, figs. 46-48. Orthonotacythere inversa (Cornuel); Neale 1960, p. 121, pi. 3. figs, la, b, 5a, b, pi. 4, fig. 11. Description. This species is here split into three subspecies which differ only in the degree and distribution of the ornament. The internal features are identical and are therefore not repeated after the initial description. 432 PALAEONTOLOGY, VOLUME 6 U q: cr < CD q: ui CL CL 3 I/) Q LU CD LU LU U CD CL LU $ o u C\J o n < > q: 1x1 I— 3 < X o u If) u (0 u u 00 u ^ >> h- h- H o D. G. STEPHENSON; SPINES AND DIFFUSE FASCIOLES 467 The function of the diffuse fasciole. This probably depended on the secondary spines carried on each tubercle. We can assume that at least part of each secondary spine was ciliated as in recent echinoids, so they probably produced currents in much the same way as the clavulae of the fascioles of more advanced spatangoids. If this were so, the secondary spines would not need a large range of movement. The clavulae of the recent Echinocardium cordatwn have cilia in two bands on opposite sides (Nichols 1959/?), but the secondary spines of Echinocorys scutata are circular in cross-section and give no evidence about the arrangement of cilia on them. On the assumption that the diffuse fascioles have the same function as true fascioles in living echinoids, an attempt is made in text-fig. 5b to reconstruct the currents produced by the diffuse fasciole. It suggests that the diffuse fascioles were situated where the currents produced would be effective. In particular, it emphasizes the importance of the ‘anal diffuse fasciole’. This runs round the slope of a raised part of interambulacrum 5, which was named by Lambert (1898) the ‘anal plastron’. The periproct is at the far end, which slopes so that the periproct faces downwards and backwards. It seems likely that the inward flowing currents produced by the diffuse fasciole would have combined as a strong current away from the anus, as indicated in text-fig. 5b. It would have been par- ticularly important to have carried ofl' waste products in this way when the animal was stationary or moving slowly. This arrangement resembles that described by Nichols (1959fl) for Spatangus raschi of which the mode of life resembles that postulated for Echinocorys scutata. Apparently the density of spines needed to produce a current strong enough for this sanitary function is more than can be obtained merely by packing normal-sized secondary tubercles as close as possible. Hence the very small tubercles in this area which first attracted attention as the ‘diffuse fasciole’. The occurrence and position of fascioles has been extensively used in the classification of spatangoids, although it has recently been shown that even in the Amphisternata this is not always justified (Nichols 1959c). From the phylogenies given by Mortensen (1951) and Durham and Melville (1957), it seems that fascioles were evolved indepen- dently several times in different spatangoid stocks. A search for diffuse fascioles or similar structures on primitive spatangoids where true fascioles are absent might throw some light on this process. Pedicellariae. Some calcite fragments on E. 663 may represent parts of the valves of pedicellariae. The fragment shown in text-fig. Ad may be the basal part of the valve of a tridentate pedicellaria, of which the blade is missing. A restoration is shown in text- fig. 4c. The upper figure in text-fig. 4<7is interpreted as being of the inner side of the valve, showing the prominent ridge of the muscle insertion. The serrated edges of the upper part suggest the ‘ irregularly dentate ’ edges of the tridentate pedicellariae of Plexiechinus spectabilis figured by Mortensen (1950). As this species is one of the few living Merido- sternata, the likeness may support the reconstruction. The lower figure shows the base, with ridges and indentations which may be part of the articulations of the valves. Another fragment (text-fig. 4/) may be part of another type of pedicellaria, seen from the inner surface, but it is so damaged as to be very difiicult of interpretation. However, under strong oblique illumination it is translucent, like fragments of echinoid spine. This strongly suggests that it is a single crystal of calcite and hence if organic, an echinoid fragment. 468 PALAEONTOLOGY, VOLUME 6 Ecology. Kongiel (1949) discusses the ecology of Echinocorys. He concludes that it buried itself in the soft sediment of the sea-floor, but not to any great depth. Willcox (1953) considers that E. scutata at times showed a trend towards a partial burrowing mode of life, but he implies that most forms lived on the surface of the sediment. The present work confirms this. E. scutata did not burrow, but lived on the surface of the sea-bed, at most sunk into the sediment to just above the level of the ambitus. It pro- pelled itself by means of the spatulate spines, and its upper surface was protected by the thorny spines. Other evidence can be found to support this. The shape of the test with flat adoral sur- face and pyramidal adapical surface separated by a sharp ambitus is consistent with it. TEXT-FIG. 6. Pore-pairs of a ‘feeding tube-foot’ adjacent to the peristome, showing ornament. In this example the central column of calcite is triangular, but this is variable. Camera lucida drawing. As Kongiel (1949) pointed out, a less elevated form of test would be expected in a burrowing form. Examination of the pore-pairs of E. scutata and the ornament surrounding them suggests a marked difference between the tube-feet of oral and adoral surfaces. But there are no pore-pairs which may be interpreted as supporting burrow-building tube-feet anywhere on the test. On the underside of the test is a conspicuous group of pore-pairs round the peristome. Each is set on an oval depression, and out of this, between the pores, rises an irregular column of calcite (text-fig. 6). This ornament is similar to, though more pronounced than that around similarly placed pore-pairs in Micr aster, which Nichols (1959a) considers carried the feeding tube-feet. But Echinocorys had more feeding tube-feet than Micraster: between 35 and 43, against an average of 22 in Micraster. The difference in the size of the ornament suggests that the feeding tube-feet of Eehinocorys were also more muscular. This contrast with Micraster is to be expected from the ecology of Echinocorys. Both echinoids were microphagous, but Echinocorys, living on the sea-floor, must have had to pick up all its food with its tube-feet, while Micraster, living in a burrow, could have trapped food particles falling on to the surface of the test, and carried them to the mouth D. G. STEPHENSON: SPINES AND DIFFUSE FASCIOLES 469 by ciliary action. Nichols (1959n) describes an apparatus possessed by the urehin to remove from the tube-feet those particles which Micraster picked up. It consisted of a series of spines around and inside the mouth, by whieh particles adhering to the tube- feet were scraped off, and then caught by a projecting labrum. There is no trace of any comparable apparatus in Echinocorys scutata. The area round the peristome is relatively clear of spines. A projecting labrum is absent and the mouth is sub-central within the peristome (Hawkins 1912). The remaining pore-pairs of the undersurface are a much smaller version of the feed- ing tube-feet. Eaeh pore-pair is set in a pit as deep as those of the feeding tube-feet, but much narrower, with only a small bump of calcite between the pores. This type of pore- pair extends to the ambitus only, except in E. scutata var. ciiicta, where, as Willcox (1953) described, the first two pore-pairs above the ambitus are of this type. The orna- ment surrounding these pore-pairs suggests that they carried a feebly muscular tube- foot with a sensory function. Willcox (1953) studied the pore-pairs of the aboral surface. He coneluded that they supported respiratory tube-feet, and demonstrated changes in form which were corre- lated with changes in environment. One small point does not appear to have been remarked, but may be significant : in the petals of many irregular echinoids, the ribbon- like tube-feet and the corresponding pore-pairs are not perpendicular to the axis of the ambulacrum, but converge adapically. The ciliary current in the living animal is de- scribed as flowing down the axis of the ambulacrum and out between the tube-feet. A typical example is Spatangus purpureas (Nichols 1959u). The pore-pairs of the respira- tory tube-feet of E. scutata converge adorally (see Willeox 1953, fig. 6). This is not always marked, but I have seen no specimen with the opposite convergence. It would have been an inefficient arrangement for the current system described. Possibly the ciliary currents in Echinocorys flowed in between the tube-feet and continued down the axis of the ambulacrum. Kongiel (1949) suggests that because of the strueture of the apical disk, with relatively small genital plates and large ocular plates, the Holasteridae were unable to develop petaloid ambulacra. The development of the ambulacral pattern seen in Echinocorys may provide increased aeration of the tube-feet in another way. Acknowledgements. This work was done while I held a Research Assistantship at the University of Reading. I am grateful to Professor P. Allen and Dr. R. Goldring for their advice and encourage- ment, to Mr. A. G. Brighton for the loan of material and for reading the manuscript, and to Dr. D. Nichols for several helpful discussions. REFERENCES BRYDONE, R. M. 1914. The Zone of Offaster piliilo in the South English Chalk. Geol. Mag. 41, 359-69. COTTEAU, G. H. 1849. Etudes snr les Echinides Eossiles du Depavtement de I'Yonne, 2, Terrain Cretace. Paris. DURHAM, J. w. and melville, r. V. 1957. A classification of echinoids. J. Paleont. 31, 242-72. GASTER, c. T. A., 1929. Chalk zones in the neighbourhood of Shoreham, Brighton and Newhaven, Sussex. Proc. Geol. Ass. 40, 328-40. HAWKINS, H. L. 1912. The plates of the buccal membrane in Echinocorys. Geol. Mag. 59, 222-5. HESSE, E. 1901. Mikrostruktur der fossilen Echinoidenstacheln und deren systematische Bedeutung. N. Jb. Min. Geol. Paldont. 13, 185-262. HOFFMAN, B. 1914. Uber die allmahliche Entwicklung der verschieden differenzierten Stachelgruppen und der Easciolen bei den fossilen Spatangoiden. Paldont. Z. 1, 216-72. 470 PALAEONTOLOGY, VOLUME 6 HYMAN, L. H. 1955. The Invertebrala, 4. New York. KONGiEL, R. 1949. Les Echinocorys du Danien de Danemark, de Suede et de Pologne. Trav. Serv. geol. Pologne, 5, 1-89. LAMBERT, J. 1898. Note sur les Echinides de la Craie de Ciply. Bull. Soc. belg. Geol. Pal. Hvdr. 11, 141-90. 1903. Description des Echinides Cretacees de la Belgique: I. Etude monographique sur le genre Ecliincorys. Menu Mas. Hist. nat. Belg. 2, 1-151. MORTENSEN, o. T. 1940. A Monograph of the Echinoidea, III, 1. Aidodonta. 1950. Ibid. V, 1. Spatangoida I. 1951. Ibid. V, 2. Spatangoida \l. NICHOLS, D. 1959n. Changes in the heart-urchin Micraster interpreted in relation to living forms. Phil. Trans. B, 242, 347-437. 1959b. The histology of the tube-feet and clavulae of Echinocardiiim cordatian. Quart. J. niicr. Sci. 100, 73-87. 1959c. Mode of life and taxonomy in irregular sea-urchins. Systeniatics Assoc. Piibl. 3, 61-80. UEXKULL, J. von. 1907. Studien tiber den Tonus: IV. Die Herzigel. Z. Biol. 49, 307-32. wiLLCOX, N. R. 1953. Zonal variations in selected morphological features of Echinocorys scutata Leske. Geol. Mag. 90, 83-96. D. G. STEPHENSON The Royal College, P.O. Box 30197, Nairobi, Manuscript received 3 September 1962 Kenya THE ANATOMICAL STRUCTURE AND SYSTEMATIC POSITION OF PENTABLASTUS (BLASTOIDEA) FROM THE CARBONIFEROUS OF SPAIN by K. A. JOYSEY and A. breimer Abstract. Pentablastiis siipracarboniciis Sieverts-Doreck was described originally from only two specimens. The collection of over twenty additional specimens has now made possible a more detailed description, including an account of the internal structure. Optical discontinuities in the calcite have been used to detect boundaries between the component plates, and serial sections have been prepared by the new annular sawing technique at 0-35 mm. intervals. These methods have revealed several unexpected features, including an optical discontinuity between the body and the limbs of each radial plate. The posterior deltoid is divided, and all the deltoids have a concealed vertical limb lying internally to the radials. There are basal pits on the floor of the thecal cavity. Although the under-lancet plate is a distinct structural unit, it is divided by an optical discontinuity along the mid-line of the ambulacrum. Each half of the under-lancet plate is continuous adorally with a hydrospire-plate, so providing direct evidence in favour of a compromise solution to the long standing controversy on the nature of the under-lancet plate in blastoids. While making comparisons with other forms it has been found that the posterior deltoid of Acentrotremites is divided, and that the subradial plate of this genus is merely a concealed internal limb of the deltoid. Penta- blastus and Owphocrimis show a close similarity in the form of the under-lancet plate, and in the relationship between the hydrospire-folds and the hydrospire -clefts. A distinction is made between hydrospire-slits and hydrospire-clefts, the latter being present in Pentablastiis and some species of Owphocrinus, including the type species. It is suggested that those species of Orophocriniis which do not possess hydrospire-clefts should be removed from this genus. Orophocriniis is itself removed from the Phaenoschismidae and placed in the same family as Pentablastiis, for which family the name Orophocrinidae Jaekel has priority over Pentablastidae Sieverts-Doreck. Amended diagnoses are given. The blastoid species Pentablastus supracarbonicus Sieverts-Doreck was described origi- nally on the basis of two specimens from the Carboniferous limestone of Rabanal, in the province of Paleneia, Northern Spain. Sieverts-Doreek (1951) recognized this speeies as the only known representative of a new family, the Pentablastidae, within the subclass Fissiculata. More specimens of Pentablastus have now been collected and the material at present available consists of twenty-two individuals and several additional fragments. All the specimens figured in the present paper have been deposited in the Rijkmuseum voor Geologie en Mineralogie, Leiden, Holland. The Leiden catalogue numbers 102451-76 correspond serially to our reference numbers 1-26, whieh are marked on the specimens. In the present paper, these numbers are given in brackets to faeilitate reference to indi- vidual speeimens, and to indicate those specimens which well display particular characters. Most of the new material was collected at Rabanal, near Cervera de Pisuerga, but three speeimens came from other loealities, near Orbo (17), the valley of El Ves (18) and Perapertu (22) respeetively, all within the provinee of Paleneia. The loeal geology has been described by de Sitter (1955). It has been suggested that the Rabanal limestone is of Westfalian age (Quiring 1939), but investigation of the fusulinid fauna suggests that it is probably of Namurian age (A. C. van Ginkel, personal communication). [Palaeontology, Vol. 6, Part 3, 1963, pp. 471-90, pis. 66-9.] 472 PALAEONTOLOGY, VOLUME 6 The present description of Pentablaslus is supplementary to that given by Sieverts- Doreck (1951). The relatively large number of specimens at our disposal has enabled us to make some additional observations on the anatomical structure, and we have found it necessary to revise the systematic diagnosis. Comparisons between the hydrospire system of PentabJastus and that of a number of other blastoid genera has led us to make a reappraisal of its systematic position. TECHNIQUES (fl) Optical properties of the component plates It is well known that the calcite in each component plate of an echinoderm skeleton is usually orientated as a single crystal. The literature on this subject has been reviewed recently by Raup (1959). With fossil echinoderms it is often possible to detect plate boundaries by differences in the reflecting properties of the surfaces of the adjacent plates (PI. 67, figs. 4, 6). Before we had thin sections available, we coarsely ground some plane surfaces trans- versely to the ambulacra, on some of the broken specimens. It was then possible to detect differences in the optical orientation of the component plates by means of their different reflecting properties, when viewed in ordinary light. This effect can be accentuated by increasing contrast during photography (PI. 68, figs. 5, 6). Lucas (1953) used a polarizing microscope to distinguish plate boundaries in thin sections of blastoids, and determined the orientation of the crystallographic axis of each of the component plates. In his work on Cryptosehisma, Lucas found that the radial, deltoid, and lancet plates all contribute to the formation of the hydrospire-folds. In Pentablastus, we have found (i) an optical discontinuity between the body of the radial and its two limbs (p. 475); (ii) the posterior deltoid is divided (p. 475); (iii) an EXPLANATION OF PLATE 66 Pentablastus supracarbonicus Sieverts-Doreck. Fig. 1. Ambulacral view of adult, x 1-5. Spec. 4, Leiden No. 102454. (Cambridge photo.) Fig. 2. Ambulacral view of young individual. X 3. Spec. 1, Leiden No. 102451. (Cambridge photo.) Fig. 3. Adoral region of ambulacrum of young individual, showing parallel rows of brachiole pits on widening series of side-plates in region of hydrospire-clefts. X 6. Spec. 2, Leiden No. 102452. (Leiden photo.) Fig. 4. Summit view of young individual, showing large hypodeltoid (subanal) in posterior interradius, and narrow limb of epideltoid (lateral) on each side of hypodeltoid. X 5. Spec. 1, Leiden No. 102451. (Leiden photo.) Fig. 5. Adoral region of ambulacrum of weathered specimen, showing broken lancet lying in groove on under-lancet. On right side, in region of hydrospire-cleft, the widened lancet covers the adoral spur of the under-lancet (now hydrospire-plate), x 7. Spec. 24, Leiden No. 102474. (Cambridge photo.) Fig. 6. Transverse section of an interambulacrum near the adoral limit of the hydrospire-folds. Note the vertical limb of the deltoid (reflecting) lying within the radial plates. The alternating strips of radial and deltoid material indicate a crenulated surface of contact between the radial and the horizontal limb of the deltoid. Each of the hydrospire-sacs lies within a strip of radial material. X 11. Spec. 5, section 5 (top), Leiden No. 102455. (Leiden photo.) Fig. 7. Naturally weathered transverse section of ambulacrum near equatorial region, showing narrow lancet lying in groove on wide top of under-lancet, and side-plates resting partly on each. Note median division of under-lancet. X 14. Spec. 11, Leiden No. 102461. (Leiden photo.) Palaeontology , Vol. 6 PLATE 66 JOYSEY and BREIMER, Pentablastus K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 473 optical discontinuity between the two halves of the under-lancet plate (p. 479). All of these observations were made using the ‘reflection’ method in ordinary light, and subsequently confirmed in thin section with a polarizing microscope (e.g. PI. 67, fig. 5). While the value of optical discontinuities in the elucidation of blastoid anatomy is well established, we have found it necessary to use the method with caution on the present material. Although most of the optical properties of the component plates represent their primary structure, there can be no doubt that both optical continuities and discon- tinuities can be the product of secondary recrystallization. It has not always been possible to make consistent observations between the several ambulacra of the same individual, and so it has proved necessary to weigh the evidence provided by several specimens. {b) Serial sections Serial drawings and photographs prepared by serial grinding are commonplace in palaeontology (e.g. PI. 68, figs. 1-4) but their preparation usually involves the total loss of the specimen. This method clearly has the disadvantage that it is impossible to check back to critical sections. On the other hand, thin sections prepared from rock slices cut by the conventional diamond-armed circular-saw cannot be obtained sufficiently close together for detailed anatomical studies. This is because the original rock slice must be fairly thick in order that it should not be shattered by lateral movements of the saw blade during the cutting process. Thwaites and Sayers (1955) have described a new saw which has overcome this diffi- culty. It consists of a thin annulus of metal which is clamped round its outer edge to tension the blade against deflexion. The inner edge is armed with diamond dust as the cutting agent, and the material to be cut is mounted within the annulus, and can be traversed into contact with the cutting edge. In suitable material it is possible to cut a series of slices, each only 0-3 mm. thick, separated by a series of parallel saw-cuts of the same width as the slices, and so only half of the material is destroyed in the cutting pro- cess. This machine was originally designed to cut thin plates of quartz for use as piezo- electric vibrators in radio components, and it is now manufactured under patent by Caplin Engineering Co. Ltd., Ipswich. Credit is due to Mr. C. F. Sayers for recognizing the potentiality of this machine for cutting serial sections of fossils. Two specimens of Pentablastus (5, 22) have been cut by the annular-saw method. Prior to cutting they were embedded in an artificial matrix consisting of a mixture of Marco Resin and plaster of Paris. Some difficulty was encountered with one of the speci- mens, partly due to calcite cleavage and partly because it was hollow. These difficulties were largely overcome by impregnating with Marco Resin after the cavity had been exposed by cutting the first few sections. The sections of Pentablastus range in thickness from 0-3 mm. to 0-4 mm., and are separated by saw-cuts of comparable width, and so by examining both sides of each slice it has been possible to achieve permanent serial sections at approximately 0-35 mm. intervals. These have been examined under a high-power binocular microscope by shining an intense transmitted light through the thickness of the slice. By this method it is possible to trace sutures through the slice and observe the direct connexion between the two surfaces. Nevertheless, it is possible to obtain separate photographs of the two sides by using a short depth of focus, and partly relying on scattering of light within the thick- 474 PALAEONTOLOGY, VOLUME 6 ness of the slice to blur the image of the opposite side (PL 69, figs. 1-10). It will be appreciated that it has been necessary to photographically reverse (mirror image) alternate pictures in order to make a continuous series, so that some surfaces actually photographed from below have been printed as though viewed from above. We are indebted to both Mr. C. F. Sayers and Mr. R. Jones for embedding the specimens and cutting the serial sections and to the Post Office Engineering Department for permission to use an Annular Sawing machine for cutting this material. We should also like to thank Mr. B. F. M. Collet and Mr. J. Hoogendoorn of Leiden, and Mr. R. D. Norman of Cambridge, for their invaluable assistance with the photography and illustrations. DESCRIPTION Size and shape. The smallest complete specimen is 14 mm. high and 11 mm. in diameter, and the largest specimen is 30 mm. high and 26 mm. in diameter. The ratio diameter/height ranges from 0-70 to 0-96 (based on eight specimens), and diameter is less than height in all the other specimens in which their relationship can be observed. The general form of specimens of different size is shown in Plate 66, figs. 1 and 2, where different scales of magnification have been used to bring the two specimens to nearly the same apparent size. The younger individual (fig. 2) is ovoid in outline, whereas the adult (fig. 1) has a pear-shaped profile. Stalk. A small remnant of the stalk is preserved on one of the smaller specimens (dia- meter 13 mm.), where a single columnal plate lies within the basal concavity of the EXPLANATION OF PLATE 67 Pentablastiis supracarbouicus Sieverts-Doreck. Fig. 1. Basal region of weathered specimen, showing three basal plates lying within basal concavity, bounded by radial plates. x7. Spec. 2, Leiden No. 102452. (Leiden photo.) Fig. 2. Same region as previous photograph, on a more deeply weathered specimen, showing basal plates surrounded by the basal pits of thecal cavity, situated on interradial-basal sutures. x7. Spec. 23, Leiden No. 102473. (Cambridge photo.) Fig. 3. Transverse section through basal region of large incomplete specimen, showing basal plates (reflecting) surrounded by basal pits of thecal cavity, some partly filled with matrix. (The basal cavity has been artificially deepened, and the small cavity filled with matrix is part of a ‘worm’- burrow.) x2-5. Spec. 5, section 20 (bottom); Leiden No. 102455. (Cambridge photo.) Fig. 4. Oblique radial view, showing optical discontinuity between body of radial (reflecting) and limbs of radial. Note lappet of radial body in contact with tip of ambulacrum. X 3. Spec. 3, Leiden No. 102453. (Cambridge photo.) Fig. 5. Longitudinal section through basal region, viewed between crossed nicols, showing on left the optical discontinuity between body of radial (light) and limb of radial (dark). Note matrix in basal concavity, basal plates (black), and on right the basal pit of thecal cavity. X6. Spec. 16, Leiden No. 102466. (Leiden photo.) Fig. 6. Side view, showing same features as fig. 4 (above). x3. Spec. 3, Leiden No. 102453. (Cambridge photo.) Fig. 7. Transverse section of radial just beyond tip of ambulacrum, showing main part of radial body (lower dark), the radial limbs (both reflecting) and the lappet of the radial body (upper dark). Note tip of lancet, under-lancet, and hydrospires completely embedded in radial. (The uppermost dark area is a bubble in the embedding medium.) X 8. Spec. 5, section 19 (top); Leiden No. 102455. (Cambridge photo.) Fig. 8. Oblique radial view of small weathered specimen, showing line of division between radial body (infraradial) and radial limb (supraradial). Note difference in cleavage between the two regions. X 3-5. Spec. 2, Leiden No. 102452. (Cambridge photo.) Palaeontology , Vol. 6 PLATE 67 JOYSEY and BREIMER, Pentablastus K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 475 theca (1). It almost covers the basal plates, was probably circular prior to weathering, and has a maximum diameter of 3-1 mm. Basal plates. Three basal plates form a pentagonal area lying entirely within the basal concavity, the margin of which is formed by the radial plates (PI. 67, fig. 1). As is usual in blastoids, one of the basal plates is smaller than the other two, and it has been con- firmed, both on a whole specimen (23) and in serial sections (5, 22), that the small basal plate is situated in the normal position, the left anterior interradius, as seen in basal view. Radial plates. As described by Sieverts-Doreck the radial plates are large and form the greater part of the theca. In large specimens the ambulacra extend five-sixths of the total height, and so the body of the radial plate is mainly on the underside of the theca. In contrast, in small specimens the ambulacra extend only three-quarters of the total height, and so the convex body of the radial plate forms a relatively larger part of the side of the theca (PI. 66, figs. 1, 2). The margins of the radial sinus are swollen alongside the lower (aboral) part of each of the ambulacra. Serial sections show that this swelling of the external surface is associated with a general increase in thickness of the radial plates in this region (PI. 69, figs. 6-10). There is an optical discontinuity between the body of the radial and its limbs (PL 67, figs. 4-7). The line of demarcation between the two regions is not represented by a suture on the surface, except in weathered specimens (PI. 67, fig. 8). It is almost horizontal on the flanks of the theca, but is characteristically curved upwards near the tip of the ambulacrum, so that a small lappet of the radial body overlies the radial limbs in this region. The tip of the ambulacrum is always in contact with the body of the radial. This optical discontinuity suggests a composite origin of the radial plate in Penta- blastus. We propose to name the lower part (radial body) the infraradial region, and the upper parts (radial limbs) the supraradial regions. The supraradial regions of Pentablas- tus have a similar disposition to the bibracliial plates of Blastoidocrinus, but insufficient is known to conjecture on possible homologies. While it is possible that only the infra- radial region of Pentablastus is homologous with the radial plate of other blastoids, it seems more likely that the whole of the radial region of Pentablastus is homologous with the radial plate. Clearly these subdivisions of the radial plate must be sought in other genera. Weathered specimens (6, 18, 23) and serial sections reveal the presence of five deep pits on the floor of the thecal cavity, situated at the junction of the interradial sutures and the radial-basal suture, similar to the pits for the chambered-organ in crinoids (PI. 67, figs. 2, 3, 5). Deltoid plates. As described by Sieverts-Doreck, the deltoid plates are nearly horizontal, and form the summit of the theca. This description, however, applies only to that part of the deltoid which is visible in external view. Examination of weathered specimens (3, 7, 14, 15, 24) and serial sections (PI. 69, figs. 1-7) has revealed that the deltoid plate also has an almost vertical limb, which extends downwards internally to the radial plates (text-fig. 1). This concealed part is nearly equal in size to the exposed horizontal limb of the deltoid. In the posterior interradius the concealed limb of the deltoid extends more than one third of the total height of the theca, and in the other interradii it extends for one quarter of the total height. 476 PALAEONTOLOGY, VOLUME 6 The upper surface of the horizontal limb of the deltoid plates is characterized by a fairly deep, median, elongated depression, but in the posterior interradius this depression is separated from the aboral margin of the anus by a transverse ridge. We find that the posterior deltoid is divided, but this condition is not immediately obvious on some specimens. A few of the deeply weathered specimens have two sutures extending from the lower (aboral) margin of the anus to the lower border of the deltoid plate (23, 24). These sutures divide the area of the deltoid below the anus into three TEXT-FIG. 1 . Pentablastus siipmcarbonicus. Transverse section of young specimen, in the middle region of the hydrospire-clefts, just above the adoral limit of the hydrospire-plates. Note the vertical limb of the deltoid (not shaded) lying internally to the radial plates (stippled). Lancet plates (vertical hatch); side-plates (horizontal hatch). X9. Spec. 17, Leiden No. 102467. regions of approximately equal width — a subanal region and two lateral regions. In contrast, on an unweathered specimen (1), the central subanal plate occupies nearly the whole deltoid area, and it overlaps the lateral regions to such an extent that they are reduced to a very narrow strip on each side (PI. 66, fig. 4). The extension of these lateral regions under the subanal plate has been confirmed by serial sections which show oblique sutures between the subanal plate and the lateral regions of the deltoid (text-fig. 2a-c and PI. 69, figs. 1-3). No suture has been observed between the two lateral regions of the deltoid in the small area of deltoid plate above (oral to) the anus (text-fig. 2a), and so these two lateral regions are, in fact, two limbs of a single plate. Using Wanner’s terminology (1940), this plate can be referred to as an epideltoid, and the subanal plate is the hypodeltoid. Horizontal sections across the theca near the adoral tip of the hydrospire-folds (text- fig. 2, on the left side of each section, at a level between b and c) reveal a complicated K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE All pattern in the form of each deltoid plate. This may be aptly described as a ‘cedar-tree’ pattern (text-fig. 3 and PL 66, fig. 6). The crown of the ‘cedar’ is formed by the vertical limb of the deltoid and the trunk is the horizontal limb of the deltoid. At this level, extensions of the radial plates underlie each side of the horizontal limb of the deltoid, TEXT-FIG. 2. Pentablastus supracarbonicus. Nearly horizontal sections through an ambulacrum (V) and the adjacent interambulacra, showing the division of the posterior deltoid into hypodeltoid (h) and epideltoid (e), and the relationship between the radial (r), and deltoid (d) plates. Lancet (l); side- plates (s). Sections at irregular intervals selected from a series at approx. 0-35 mm. interval. All drawn as though viewed from above. X 5. Spec. 5, Leiden No. 102455. For additional detail see PI. 69, figs. 1-4, where 1 = a;2 = b;3 = c;4 = d. and the lower branches of the ‘cedar’ are formed by a crenulated surface of contact between these plates, resulting in alternate strips of radial and deltoid material. Each of the strips of radial material is associated with the adoral tip of a hydrospire-fold. (The hydrospire openings have the form of a single hydrospire-cleft on each side of the adoral region of the ambulacrum. They extend adorally beyond the limit of the hydro- spire-folds, and reach a level just above the radial-deltoid suture, as seen on the exterior of the theca (p. 481).) Side-plates. The ambulacra are long and narrow. Each ambulacrum, at its lower (aboral) end, is raised on a prominence relative to the interradial areas, but it passes into a I i C 1456 478 PALAEONTOLOGY, VOLUME 6 shallow groove when traced towards the mouth. Even in the raised region the side-plates lie in a furrow relative to the swollen margin of the radial sinus (PI. 66, figs. 1, 2). The adjacent rows of side-plates meet in the mid-line of each ambulacrum through- out its length, and so the lancet plate is not exposed in a perfect specimen. There are fifty side-plates in each row in a specimen 16 mm. high (1) and about ninety in a speci- men 30 mm. high (3). A small triangular outer side-plate is associated with each side- plate, overlying its top outer corner and separated from it by an oblique suture (8). There is a single brachiole pit on each of the side-plates. Although the ambulacra TEXT-FIG. 3. Pentablastus siipracarbomciis. Horizontal section through an interambulacrum near the adoral limit of the hydrospire-folds, showing the ‘cedar-tree’ pattern of the deltoid plate (d). Radial plates stippled. For explanation see text. X 15. Spec. 5, section 5 (top); Leiden No. 102455. (Com- posite drawing made from several photographs, one of which is illustrated on PL 66, fig. 6.) EXPLANATION OF PLATE 68 Pentablastus supracarbonicus Sieverts-Doreck Figs. 1-4. Series of nearly transverse sections of young specimen, obtained by parallel grinding and photographed at intervals. x7-5. Spec. 17, Leiden No. 102467. (Leiden photos.) Fig. 1 . Adoral region of hydrospire-clefts, showing adoral end of hydrospire-folds at slightly different levels of section. Fig. 2. Middle region of hydrospire-clefts, showing the hydrospire-folds opening independently into each hydrospire-cleft. Fig. 3. Aboral region of hydrospire-clefts, showing hydrospire-folds opening into a common hydro- spire-canal. Note slight thickening of thecal wall between hydrospire-folds. Fig. 4. Dorsal to hydrospire-clefts, showing thickened wall of theca enclosing bases of hydrospire- folds, and under-lancet lying between the two groups of hydrospires. Figs. 5-6. Obliquely transverse sections of ambulacra, obtained by coarse grinding, and photographed in ordinary light. Differences in reflecting properties accentuated by increasing contrast during photography. Spec. 25, Leiden No. 102475. (Cambridge photos.) Fig. 5. Equatorial region of ambulacrum, showing optical discontinuity between the two halves of the under-lancet. The secondary infilling of the hydrospire-folds has formed in optical continuity, on left, with the adjacent half of the under-lancet, and on right, with the lancet, x 14. Fig. 6. Aboral region of ambulacrum, showing lancet resting on under-lancet plate, and the almost complete enclosure of the hydrospire-folds within the wall of the theca. X 12. Palaeontology, Vol. 6 PLATE 68 JOYSEY and BREIMER, Pentahlastus r%2snssnni K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 479 change in width, the distance between the two rows of brachiole pits remains constant. In consequence, the outer margins of the ambulacral plates are unsculptured in the upper part of the theca (PL 66, figs. 3, 4). Lancet plate. In a specimen in which the side-plates and outer side-plates have been re- moved (PI. 66, fig. 5), it may be seen that the lancet plate extends along the entire length of the ambulacrum. Within the radial sinus the lancet plate is nearly parallel sided and does not occupy the whole width of the ambulacrum, but it gradually widens in the region of the hydrospire openings (7), and finally, in the region between the deltoids it occupies the whole width of the ambulacrum. A longitudinal canal perforates the lancet through- out its length, and at the adoral end connects with a ring canal lying within the deltoid plates (text-fig. 2a). The exterior surface of the lancet plate has a crest along its mid-line throughout its length, underlying the suture separating the adjacent rows of side-plates. These lie on each side of the crest, in a shallow groove, which is sculptured to receive the individual side-plates. The inner surface of the lancet plate shows several changes in form when traced from the lower (aboral) end of the radial sinus towards the mouth. These changes in the shape of the cross-section of the lancet plate can be observed in the serial sections illustrated (text-figs. 2, 4; PI. 69), but it must be noted that at the top of the theca the lancet plate curves over towards the mouth and so these horizontal sections cut the lancet at an oblique angle, and give an exaggerated impression of its thickness relative to the width. At the lower end, the inner surface of the lancet is slightly convex, but it gradually becomes more convex, and in the region just aboral to the hydrospire-clefts there is a median ridge with a shallow depression on each side. In the adoral region of the hydrospire- clefts the lancet has an almost flat inner surface, so that it has a pentagonal cross-section. Traced farther towards the mouth the pentagonal cross-section persists, but the thick- ness of the lancet increases relative to its width. Under-lancet plate. Within the radial sinus the lancet plate rests upon an under-lancet plate which is wider than the lancet and so is exposed in exterior view when the side- plates have been stripped off. A suture marks the line of contact between the under- lancet and the adjacent radial plate (PI. 66, figs. 5, 7). The surface of the under-lancet has a groove along its mid-line to receive the convex lower surface of the overlying lancet, and the side-plates rest partly on the under-lancet and partly on the lancet itself. The under-lancet occupies the whole space between the hydrospire groups, its lateral margin being determined by the position of the innermost hydrospire-fold in each group. The inner surface of the under-lancet is almost flush with the inner surface of the radial plate on each side, and both increase in thickness towards the lower (aboral) part of the ambulacra (PI. 69, fig. 10). Although the under-lancet plate forms a distinct structural unit, it consists of two parts which join along a suture in the mid-line of the ambulacrum, and they are not in optical continuity with one another (PI. 68, fig. 5). These two halves of the under-lancet are continuous with a pair of hydrospire-plates. Direct evidence of this interpretation has been obtained from serial sections (text-fig. 4n-c; PI. 69, figs. 6-10). At the level of the hydrospire-clefts, there is a discrete hydrospire-plate lying on each side of the lancet, in the depressions on each side of the median rib on its inner surface. Aborally to the 480 PALAEONTOLOGY, VOLUME 6 hydrospire-clefts these hydrospire-plates increase in size and come into contact with one another under the lancet plate, and other sections lower down the ambulacra show a progressive thickening to form an under-lancet plate. At first sight, on the assumption that the two halves of the under-lancet might be TEXT-FIG. 4. Pentablastiis supracarbonicus. Nearly horizontal sections through an ambulacrum (IV) and the adjacent interambulacra, showing the hydrospire-plates (stippled) coming into contact in the mid-line, and forming an under-lancet plate in the region aboral to the hydrospire-clefts. Deltoid (d); lancet (l); radial (r); side-plates (s). Sections at irregular intervals selected from a series at approx. 0-35 mm. interval. All drawn as though viewed from above. X 5. Spec. 5, Leiden No. 102455. For additional detail see PI. 69, figs. 6-10, where 6 = a;7 = b;8 = c;9 = d; 10 = e. derived ultimately from the same radial plate, the optical discontinuity between the two halves is unexpected. The situation is complicated, however, by the optical discontinui- ties within the radial plate itself. Apart from the optical difference between the body of the radial and the radial limbs, in some specimens the two limbs also differ from one another in their optical orientation. While optical discontinuity between the two limbs of the radial might be a product of mechanical distortion of the specimen, it is difficult to envisage that this could be the cause of optical discontinuity between the two closely contiguous halves of the under-lancet plate. K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 481 In a section of an apparently undistorted specimen the under-lancet halves situated on opposite sides of the same interradius are not in optical continuity with one another, and so it is unlikely that the under-lancet plates are derived ultimately from the deltoid plates. In many sections there is an optical dilference between one half of the under- lancet, the adjacent radial limb and the adjacent deltoid, and so it would appear that the component parts of the under-lancet have an optical orientation of their own. The compound under-lancet plate of Pentablastus occupies an analogous position to the under-lancet plate described by Etheridge and Carpenter (1886) in some species of Pentremites and Orophocrinus. The possible homology of these structures is discussed later. Hydrospires. The hydrospire system is entirely internal. There are ten hydrospire groups, each consisting of six hydrospire-folds in the larger specimens, but only five folds in smaller individuals (PI. 68, fig. 3; PI. 69, fig. 6). The hydrospire-folds are relatively long and narrow, but inflated at the tips. In the region of the hydrospire-clefts the individual hydrospire-folds open indepen- dently into a common longitudinal cavity which directly communicates with the exterior through the hydrospire-cleft itself (PI. 68, fig. 1). Aborally to the level of the hydrospire- cleft there is a tendency for the hydrospire-folds to merge to form a common longitudinal canal which opens into the lower (aboral) end of the cavity within the hydrospire-cleft (PI. 68, fig. 3). Throughout the rest of its length the longitudinal canal does not com- municate directly with the exterior, its roof being sealed by the contact of the under- lancet and radial plate. The hydrospire-clefts themselves are long narrow openings, one lying on each side of each ambulacrum, extending aborally from a position just above (oral to) the radial- deltoid suture, but confined to the oral quarter of the theca (PI. 66, figs. 2-5). At its upper end the hydrospire-cleft widens slightly between the deltoid and the row of side- plates (1), but in this region it does not communicate directly with the interior cavity (text-fig. 2a, b\ PL 69, figs. 1, 2). At the level of the hydrospire-cleft the hydrospires are entirely free within the body cavity (PI. 68, fig. 2), but below (aboral to) the hydrospire-cleft the radial plate thickens (PI. 68, fig. 4), the under-lancet thickens, the spaces between the individual hydrospire- folds are progressively filled in, and so the hydrospire-folds become progressively en- closed within the wall of the theca (PI. 68, fig. 5). At the lower (aboral) end of the radial sinus only the tips of the innermost (medial to the ambulacrum) fold of each group project into the body cavity (PI. 68, fig. 6). Sections reveal that the tip of the lancet plate, the much reduced hydrospire groups and the under-lancet project into the body of the radial plate, beyond the externally visible end of the ambulacrum (PI. 67, fig. 7). SYSTEMATIC POSITION Comparison with Acentrotremites Sieverts-Doreck (1951) rightly drew attention to the superficial resemblance between Acentrotremites and Pentablastus, in the form of the theca, in the relative size, shape, and position of the deltoids, radials, and basals, and in the similar hydrospires. She also recognized the fundamental difference between the two forms, Acentrotremites having 482 PALAEONTOLOGY, VOLUME 6 spiracles and marginal pores, whereas Pentablastus has elongate hydrospire-clefts and no marginal pores. On this basis she retained Acentrotremites within the subclass Spira- culata and placed Pentablastus within the subclass Fissiculata. Having re-examined the two specimens of Acentrotremites in the British Museum (Natural History) (E 782 and E 8256) and directly compared them with specimens of Pentablastus, it is possible to confirm all the similarities and differences described by Sieverts-Doreck, and also to note several additional characters in which the two genera may be compared and contrasted. The median furrow on the deltoid of Pentablastus is remarkably similar to that de- scribed by Bather (1912) on Acentrotremites. Eurthermore, in the deeply weathered specimen of Acentrotremites (E 782) a vertical limb of the deltoid may be seen to extend for a considerable distance beneath the radial plates, in a similar manner to that of Pentablastus. Optical discontinuities in the posterior deltoid of one specimen of Acentrotremites (E 782) suggest that it is divided in a similar manner to that of Pentablastus. On each side of the anus, and extending slightly below its aboral margin, there are two small EXPLANATION OF PLATE 69 Pentablastus sitpracarboniciis Sieverts-Doreck. Figs. 1-10. Nearly horizontal sections of an ambulacrum and the adjacent interambulacra. 1-5 centred on Amb. V, and 6-10 centred on Amb. IV. Sections selected from a series at approx. 0-35 mm. interval, cut by the annular sawing technique. Photographed by transmitted light, and all printed as though viewed from above. x7. Spec. 5, Leiden No. 102455. (Cambridge photos.) Fig. 1. Section 2 (top) near summit, with matrix in the grooves on the surface of the deltoids, and in the upper end of the hydrospire-clefts. Note (left) thin leaf of radial overlying outer corners of deltoid; (centre) matrix filled brachiole pits on nearly horizontal region of ambulacrum; (right) anus lying between descending limbs of epideltoid, with the hypodeltoid forming the main outer surface of the posterior deltoid region. Fig. 2. Section 3 (bottom), 105 mm. below previous. Note (left) massive deltoid completely overlain by thin radials; (centre) matrix infilling canal within lancet; (right) epideltoid limbs completely internal, and outer corner of hypodeltoid overlain by radial. Fig. 3. Section 4 (bottom). 0-70 mm. below previous. Note (left) upper limit of hydrospire-folds lying adjacent to deltoid of ‘cedar-tree’ form; (right) epideltoid limbs still present, and hypodeltoid covered by radial, except in weathered region. Fig. 4. Section 5 (bottom), 0-70 mm. below previous, showing on each side the vertical limb of the deltoid completely covered by the radials. Note (left) hydrospire-folds suspended within thecal cavity. Fig. 5. Section 6 (top), 0-35 mm. below previous, showing further reduction of deltoids and enlarge- ment of hydrospire-folds. Fig. 6. Section 7 (top), 0 70 mm. below previous, showing (right) hydrospire-folds opening into the hydrospire-cleft, adjacent to the adoral tip of the hydrospire-plate. Fig. 7. Section 8 (bottom), 1-05 mm. below previous, showing a hydrospire-plate lying on each side of the lancet plate. Note (left) aboral termination of hydrospire-cleft. Fig. 8. Section 10 (top), L05 mm. below previous, showing the hydrospire-plates just in contact beneath the mid-line of the ambulacrum. Fig. 9. Section 10 (bottom), 0-35 mm. below previous, showing the under-lancet plate, formed by the union of the hydrospire-plates beneath the lancet plate. Note the proximal ends of the hydrospire- folds embedded within the wall of the theca. Fig. 10. Section 12 (bottom) 1-40 mm. below previous, showing the increase in thickness of the radial and under-lancet plates. Note the mid-line suture between the two halves of the under- lancet plate. Palaeontology, Vol. 6 PLATE 69 JOYSEY and BREIMER, Pentahlastus K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 483 ‘ plates ’ which have the same optical orientation as one another, but which are optically discontinuous with the main deltoid plate. It seems likely that they are descending limbs of an epideltoid plate, but the state of preservation is such that their continuity round the adoral side of the anus cannot be established. Hence, it is possible that these two ‘plates’ might be the equivalent of the two ‘unnamed plates’ recently described by Fay (1960) in Orbitremites. It is, however, impossible to decide between these two possibilities, and the critical region of the other specimen, E 8256, is obliterated by cracks. It is well known that the holotype is lost, and the posterior deltoid of the specimen described by Phillips (1936) was not preserved. TEXT-FIG. 5. Aceiitrotreinites ellipticiis. A series of transverse sections through an ambulacrum and parts of the adjacent interambulacra. Redrawn from Phillips (1936). No scale given. The internal structure of Acentrotremites is known from a single specimen (Phillips 1936) and shows a number of marked differences from Pentablastus. There are four hydrospire-folds in some hydrospire groups and five in others, and the hydrospires hang freely within the theca throughout their length (with the exception of a single loop in one ambulacrum). Near the tip of the ambulacra the inner surface of the lancet plate is flat (text-fig. 5g), but when followed towards the mouth it becomes progressively concave until the cross-section of the lancet has the form of an inverted V (text-fig. 5c). This internal structure is so different from that of Pentablastus that we cannot envisage any near relationship between the two genera, and we regard their resemblance as entirely superficial. According to Phillips (1936), there is only a single pentagonal basal plate in Acentro- tremites. Furthermore, Phillips’s illustrations suggest that the outer side-plates are con- tinuous with the plate which underlies the lancet (text-fig. 5c-c), and both of these structures were regarded as being extensions of the radial plate (text-fig. 5/). These unexpected interpretations clearly deserve further investigation. Phillips (1936) also described the presence of an internal subradial plate, lying beneath 484 PALAEONTOLOGY, VOLUME 6 the interradial sutures, apparently supporting the radial plates (text-fig. 5a-c). She suggested that Acentrotremites was sufficiently different from the Orbitremitidae to justify its being placed in a new family. Sieverts-Doreck (1951) accepted this opinion and established the new family Acentrotremitidae on the grounds that the subradial plate described by Phillips is a structure unique among blastoids. It appears to us that the subradial plate is, in fact, the concealed vertical limb of the deltoid, extending down behind the radial plates. This interpretation was suggested by comparison with our sections of Pentablastus, and as has already been described, the presence of an internal extension of the deltoid plate below the external level of the radial-deltoid suture has been confirmed on the deeply weathered specimen of Acentrotremites (E 782). It is, therefore, proposed that Sieverts-Doreck’s (1951) diagnosis of the family Acentrotremitidae be amended by deleting the sentence which refers to the subradial plate. The generic and family diagnoses require further amendment, as our observations indicate that the posterior deltoid is divided, but the available material does not allow us to be definite on the nature of this division. It is inopportune to discuss whether the other characters of Acentrotremites merit its being placed in a separate family, because we understand that another specimen of Acentrotremites has been found in Somerset. It is now in the Department of Geology, Bristol University, and it has been worked upon already by Professor W. F. Whittard. It is perhaps worth noting that Acentrotremites has been recorded from the Lower Carboniferous of England and Wales, but, contrary to Sieverts-Doreck (1951), not from Scotland. Comparison with Nymphaeoblastus Sieverts-Doreck (1951) suggested that the new family Pentablastidae was intermediate between the Phaenoschismidae and the Nymphaeoblastidae in Wanner’s (1940) classi- fication. She noted that the Pentablastidae and Nymphaeoblastidae have large radial plates and that in both families the small basal plates are situated in a basal concavity. She stated that all three families possess long ambulacra, ten hydrospire groups, and an undivided posterior deltoid. Our finding that the posterior deltoid of Pentablastus is divided does not weaken this comparison, because Fay (1961) has recently found that the posterior deltoid of Nymphaeoblastus is divided into epideltoid and hypodeltoid regions. However, in Nymphaeoblastus, Yakovlev (1926) and Fay (1961) have described the presence of up to fifteen hydrospire-slits in each hydrospire group, opening to the surface in the regions of the radial-deltoid sutures in a manner comparable to Cadaster. Nymphaeoblastus is not sufficiently well known to be sure of its affinities, but this condi- tion of the hydrospire openings appears to be so different from that of Pentablastus that any close relationship between these two forms seems unlikely. Comparison with Orophocrinus The presence of a divided posterior deltoid in Pentablastus does not detract from Sieverts-Doreck’s comparison with the Phaenoschismidae, because Reimann (1945, 1950) has found that in some species previously referred to Phaenoschisma the posterior deltoid is divided, and he has proposed the new genus Pleuroschisma for these forms. Furthermore, Wanner (1940) indicates that the posterior deltoid of Orophocrinus (Phaenoschismidae in his classification) is divided. K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 485 Sieverts-Doreck drew particular attention to other resemblances between Oropho- crinus a.nd Pentablastus. In both genera part of the hydrospire-folds are embedded within the thickness of the radial plates, and in Pentablastus she described an under-lancet structure in a similar position to the under-lancet plate known in Orophocrinus, but was undecided whether it was an under-lancet plate or part of the radial plate. The nature of the under-lancet plate in blastoids has been the subject of a long-standing controversy which was summarized, but not satisfactorily resolved, by Etheridge and Carpenter (1886). An under-lancet plate has been described previously only in Pentremites and Orophocrinus, but it has a completely different form in the two genera. In Pentremites it is a thin plate which invests the lower surface of the lancet. Etheridge and Carpenter found that it sometimes remained in place when the lancet was removed (1886, pi. 12, figs. 13, 14, 16), but they were unable to locate it in cross-sections through the ambulacra. In contrast, the under-lancet of Orophocrinus is a massive structure underlying the lancet plate, and visible in some cross-sections of the ambulacra, extend- ing down between the hydrospire groups (Etheridge and Carpenter 1886, pi. 17, fig. 12). Hambach (1884) denied the presence of an under-lancet in Pentremites and suggested that the structure observed by Etheridge and Carpenter was only the uppermost blades of the hydrospire-sacs. In contrast, Etheridge and Carpenter ( 1 886, pp. 47-50) maintained that the under-lancet was a mid-ambulacral structure, and stressed that the hydrospire- plates, as known in Mesoblastus, Orbitremites, and Crvptoblastus, ‘belong to the sides of the ambulacra and do not meet beneath its middle line, where a more or less wide gap is left between them, leading down into the interior of the calyx’. In support of this view they illustrated a specimen of Orophocrinus pentangularis in which the under-lancet is partly broken away and appears to be resting upon the truncated edges of the inner- most walls of the hydrospire-sacs (Etheridge and Carpenter 1886, pi. 15, fig. 10). Never- theless, their section through an ambulacrum of O. pentangularis (op. cit., pi. 17, fig. 14) shows no under-lancet, but in their section of O. stelliformis (pi. 17, fig. 12), the region between the innermost hydrospire-folds is occupied by the stem of the under-lancet. Despite this conflicting evidence drawn from different species of Orophocrinus, Bather (1900, pp. 84, 85) put forward the obvious compromise solution, and stated that in this genus: ‘The concentration of the hydrospire-slits causes the inner walls of the two [hydrospire-folds] nearest the median line of the ambulacrum to meet along that line, and so to form a new structure known as the under-lancet plate.’ It appears that our finding that the under-lancet plate of Pentablastus is continuous with a pair of hydro- spire-plates and yet forms a single mid-ambulacral structure, has provided the first direct evidence in support of this compromise solution. The under-lancet plate of Orophocrinus clearly deserves further investigation. Pentablastus is similar to Orophocrinus in the nature of the hydrospire openings. Within the subclass Fissiculata there are two different types of communication of the hydrospire-folds with the exterior. By common usage in the literature one type of opening has become known as ‘hydrospire-slits’, and we suggest that the other type is best described as the ‘hydrospire-cleft’ (Etheridge and Carpenter (1886) used hydrospire- cleft, spiracular-cleft, elongated spiracle, linear spiracle; Bather (1900) used spiracle- slit; Jaekel (1918, p. 109) used the singular form Faltenschlitz; Sieverts-Doreck (1951) used Faltenschlitze). In the first type (hydrospire-slits) each individual hydrospire-fold opens separately through the wall of the theca by its own hydrospire-slit. In some forms 486 PALAEONTOLOGY, VOLUME 6 these individual slits open directly to the exterior (e.g. Codaster) and in others they are partly concealed (e.g. Phaeiwschisma) or almost wholly concealed (e.g. Cryptoschisma) by the side-plates. The other type of hydrospire opening is found in Penlablastus and some species of Orophocriuus. In these forms the hydrospire-folds of each hydrospire group all open into a common cavity which communicates with the exterior through a single opening, the hydrospire-cleft, situated between the row of side-plates and the wall of the radial sinus. At first sight this second condition appears to be similar to that found in Crypto- schisma, but in this form the radial sinus is wide and the extremely broad side-plates roof over the area of hydrospire-slits, leaving an opening to the exterior along the margin of the side-plates. In contrast, the hydrospire-cleft in Pentablastiis and Orophocriuus is associated with a narrow radial sinus, and so the cleft leads inwards at a steep angle. Obviously the distinction between these two conditions is only one of degree, but it appears to be a useful distinction because the concealment of the hydrospire-slits has been effected in a different manner in the two forms. This distinction is further supported by the fact that in Cryptoschisma the aboral tips of the hydrospire-slits are sometimes exposed, projecting just beyond the margin of the side-plates, whereas in Pentablastiis and Orophocriuus the aboral ends of the hydrospire-slits are completely enclosed, and in some species the hydrospire-cleft itself is confined to the adoral region of the ambula- crum. Accordingly, we propose that the term ‘hydrospire-cleft’ should be restricted to those forms in which the hydrospire-slits are completely concealed, this condition being associated with a narrow radial sinus. There is a minor complication of structure in Pentablastiis which has already been mentioned (p. 481). Throughout the greater part of the length of the ambulacra the individual hydrospire-folds open into a common longitudinal canal which is roofed over by the contact between the under-lancet plate and the radial plate. The hydrospire-cleft breaks through this roof and so opens into the common canal. At the extreme aboral end of the hydrospire-cleft the hydrospire-slits practically merge because they all open into the floor of the common canal at about the same place. When traced towards the adoral end of the hydrospire-cleft they diverge and form a series of distinct hydrospire-slits open- ing separately into the floor of the hydrospire-cleft just in as most species of Orophocriuus. Etheridge and Carpenter (1886) distinguished between hydrospire-slits and hydrospire- clefts, and Jaekel (1918, p. 109) regarded them as being different characters worthy of systematic recognition, but subsequently this view has not been generally accepted. In some forms it is difficult to decide whether the hydrospire opening should be regarded as a hydrospire-cleft or a hydrospire-slit. For example, in the Permian blastoid genus Anthoblastus there is a single slit-like opening situated adorally alongside the ambulacra, communicating with a single hydrospire-fold which does not extend any farther along the ambulacrum than the slit (Wanner 1924, 1940). But such a case need not detract from the usefulness of the character in general application. Reviewing the species of Orophocrinus as illustrated by Etheridge and Carpenter (1886) it appears that in O. orbignyanus and O. puzos some of the hydrospire-slits are exposed, whereas O. gracilis, O. pentangularis, O. stelliformis (genotype), and O. verus all have hydrospire-clefts. Etheridge and Carpenter’s figures (1886, pi. 11, fig. 10; pi. 13, figs. 15, 16, 18) show that in O. orbignyanus and O. puzos some of the hydrospire-slits are visible at the sides of the ambulacra, just as in Phaenoschisma. They stated (p. 90) K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 487 that: ‘but for the form of the calyx and the appearance of the deltoid plates externally they might well be referred to Phaenoschisnia' . The present authors are of the opinion that the nature of the hydrospire openings is a more fundamental character than the extent of the deltoid plate. It seems to be of trivial anatomical importance whether the deltoid plate appears on the lateral wall of the calyx or is just confined to the summit (and in some forms it is not easy to define the boundary between the summit and the lateral wall of the calyx). We believe that those species in which some of the hydrospire- slits are exposed should be excluded from the genus Orophocrinus and placed in the genus Phaenoschisma, and it appears from the literature that O. orbignyamis and O. puzos belong to this category. Alternatively, it is possible that either or both of these species may belong to the genus Pleuroschisma, depending upon whether or not the posterior deltoid is divided. In addition to the six species discussed in the previous paragraph, Bassler and Moodey (1943) list another eight species of Orophocrinus. Of these, O. conicus Wachsmuth and Springer has been described (Whitfield 1895) as a synonym of O. whitei (Hall). The original description of O. whitei suggests the presence of several hydrospire-slits, but Whitfield’s (1895) illustration shows that these are only present on the internal cast. His illustration of an entire specimen shows no hydrospire openings, but Wachsmuth and Springer’s (1890) illustrations of O. conicus show a single hydrospire-cleft on each side of the ambulacra, and so this species is probably correctly referred to Orophocrinus rather than Phaenoschisma. Orophocrinus depressus (Cumberland) has been suppressed as a synonym of O. pen- tangularis (Miller) (Etheridge and Carpenter 1886). Orophocrinus fusiformis (Wachsmuth and Springer) has wide depressed ambulacra, and in the original description it was stated that four to six hydrospire-slits are visible in the hydrospire-clefts (Wachsmuth and Springer 1890). In the absence of an illustration of the hydrospire openings it is impossible to decide whether this species belongs to Phaenoschisma or Orophocrinus. Etheridge and Carpenter (1886) suggested that Orophocrinus hibernica (Cumberland) and O. humero-stellata (Cumberland) appear to belong to the genus Orbitremites (Granatocrinus), and judging from Cumberland’s (1826) figures (bearing in mind that plates A and B have been reversed), there appears to be no reason to doubt their opinion. The description of Orophocrinus praelongus Baily was originally based on two speci- mens, but in a postscript Baily (1886) stated that the specimens had been examined by Etheridge and Carpenter, who concluded that one of the specimens belonged to O. pentangularis, although they recognized the other as belonging to a distinct species. Nevertheless, apart from a difference in size between the two specimens, the original description does not provide any characters by which O. praelongus can be distinguished from O. pentangularis. Finally, although O. sirius (White) has long been regarded as congeneric with O. stelli- formis, Etheridge and Carpenter (1886) stated that ‘White’s description and figure alone do not convey this idea to our minds at all’. We cannot add to this comment, as no hydrospire openings are shown on the original illustration (White 1865). Hence, of the fourteen species of Orophocrinus listed by Bassler and Moodey (1943), it appears from the literature that gracilis, pentangularis, verus, and probably whitei, are congeneric with the genotype O. stelliformis. It further appears that conicus is a 488 PALAEONTOLOGY, VOLUME 6 synonym of whitei, and that both depressus and praelougus are synonyms of pentangu- laris. It appears that orbignyanus and puzos should be excluded from Orophocrinus and we suggest that they should be referred to Phaenoschisma (or Pleuroschisma). There is insufficient information to decide whether fiisiformis belongs to Orophocrinus or Phaenoschisma. We accept that hibernica and humero-stellata probably belong to the Orbitremitidae, and we do not wish to speculate on the relationship of siriiis. As here restricted the genus Orophocrinus becomes a more homogeneous group which shares the nature of the hydrospire openings and the internal structure of the ambulacra with Pentablastus and differs in these characters from Phaenoschisma. We therefore pro- pose that Orophocrinus should be removed from the family Phaenoschismidae and placed in the same family as Pentablastus. For this family the name Orophocrinidae Jaekel 1918 has priority over Pentablastidae Sieverts-Doreck 1951. Pentablastus supracarbonicus can be most closely compared with Orophocrinus verus from the Lower Carboniferous of Lancashire, England, this latter species being the most inflated form of Orophocrinus. The morphological differences between Orophocrinus and Pentablastus are relatively minor. Whereas Orophocrinus, according to the species, has fifteen to twenty side-plates in a row, the only known species of Pentablastus has up to ninety side-plates. It has been shown that the under-lancet plate of Pentablastus is derived from a pair of hydrospire-plates, whereas the relationship of the under-lancet of Orophoerinus is unknown. Finally, Orophocrinus has an elongate base ecwd Pentablastus has a basal concavity, but comparable differences in form can be cited within other blastoid families. In the present state of knowledge of blastoid anatomy it is impossible to judge the significance of our finding that the radial plate of Pentablastus is divided by an optical discontinuity. At present, we prefer to place Pentablastus and Orophocrinus in the same family. Future studies may reveal that the divided radial is of considerable phylogenetic and taxonomic significance and that the Pentablastidae deserve to stand apart from the Orophocrinidae. Even if this should prove to be the case, our present grouping of these two forms will have drawn attention to the common features of their hydrospire systems, which would then have to be regarded as a remarkable example of convergent evolution. AMENDED DIAGNOSES Order fissiculata Jaekel Family orophocrinidae Jaekel Amended diagnosis. A family of fissiculate blastoids in which the hydrospires communi- cate with the exterior through a single hydrospire-cleft on each side of the ambulacra, and in which an under-lancet plate is present. Ten hydrospire groups. The hydrospires are freely suspended in the thecal cavity in the adoral region but towards the aboral end of the ambulacra they become progressively embedded within the wall of the theca. Ambulacra narrow, linear to subpetaloid. The posterior deltoid is known to be divided in some forms, the anus passing between hypo- and epideltoid plates. Genus orophocrinus Von Seebach 1864 The diagnosis given by Etheridge and Carpenter (1886) requires only two small K. A. JOYSEY AND A. BREIMER: ANATOMICAL STRUCTURE 489 amendments, replacing the term ‘spiracles’ by ‘hydrospire-clefts’ and inserting ‘Hydro- spire-slits completely concealed by the side-plates and partly covered by an under- lancet plate’. Type species. Peutremites stelliformis Owen and Schumard 1850. Genus pentablastus Sieverts-Doreck 1951 Amended diagnosis of genus and only known species. An Orophocrinid with a Pentremites- like theca, and a depressed base. Basal plates three, unequal, the smaller one in the usual blastoid position; the basal plates are situated in a basal concavity and nearly covered by the stalk. Stalk circular in outline. Radials large, divided by a transverse optical discontinuity, the infraradial region (body) forming the floor of the theca, the supra- radial regions (limbs) long and forming the upflaring part of the theca. Deltoids with a subhorizontal upper limb forming the summit of the theca, and a subvertical lower limb extending under the radial plates. Posterior deltoid divided, the hypodeltoid lying between the descending limbs of the epideltoid, the former almost covering the latter. Anus situated between hypo- and epideltoid. Ambulacra long and narrow, entirely covered by the side-plates and outer side-plates. Up to ninety side-plates in each row. Brachiole pits present, but brachioles unknown. Lancet plate present, slightly longer than the ambulacrum, perforated by a longitudinal canal. Lancet occupies the space between the deltoids, but it becomes narrower aborally and does not occupy the full width of the radial sinus. Lancet underlain by a substantial under-lancet, lying in the region aboral to the hydrospire-clefts. Under-lancet continuous with a pair of hydrospire- plates, one on each side of the lancet in the region of the hydrospire-clefts. A single hydrospire-cleft on each side of the ambulacra confined to the adoral region. Except in the region of the hydrospire-clefts, the hydrospire-folds open into a common longitu- dinal hydrospire-canal which is covered by the contact between the under-lancet and the radial. Ten hydrospire groups each with five or six hydrospire-folds. Type species. Pentablastus siipracarbcmiciis Sieverts-Doreck 1 95 1 Addendum. Subsequent to the completion of this study another work by Fay has become available to us (fay, r. o. 1961. Blastoid Studies. Univ. Kans. Paleont. Contr. Echino- dermata. Art. 3. 147 pp., 54 pi., 221 figs.). Fay’s work does not contain any description of Pentablastus, but he has independently reached the conclusion (p. 15) that the under- lancet plate is formed from the hydrospire-plates. In his provisional classification Fay has placed Pentablastus in the family Orophocrinidae, in which he includes all fissi- culate blastoids possessing ten spiracular-slits, with the hydrospire-slits opening into them. Fay’s term ‘ spiracular-slit’ includes the condition described by our term ‘hydro- spire-cleft’, but the two terms are not equivalent. Fay uses the term ‘spiracular-slit’ in a broad sense, to describe any excavation or aperture at the margin of the ambulacrum of a fissiculate blastoid. Hence, he uses the same term for the diverse conditions present in Cryptoschisma and Orophocrinus, and also uses this term to indicate excavations present in a form such as Pleuroscbisma, which also has exposed fields of hydrospire-slits. We note that Fay (p. 14) has regarded the terms ‘hydrospire-slit’ and ‘hydrospire-cleft’ as being synonymous, but we hope that the latter term will be adopted and used in the restricted sense suggested in our paper. 490 PALAEONTOLOGY, VOLUME 6 REFERENCES BAiLY, w. H. 1886. On a new species of Orophocrimis (Pentremites) in Carboniferous Limestone, County Dublin. Also remarks upon Cadaster trilobatiis (M‘Coy), from Carboniferous Limestone, County Kilkenny. Sci. Proc. R. Dublin Sac., n.s., 5, 31-3, 1 pi. BASSLER, R. s. and MOODEY, M. w. 1943. Bibliographic and faunal index of Paleozoic pelmatozoan echinoderms. Spec. Pap. geol. Sac. Amer. 45. BATHER, F. A. 1900. The Echinoderma; in A Treatise on Zooiogy, Pt. Ill (ed. E. Ray Lankester). London. 1912. Two blastoids (Orophocrimis and Acentrotremites) from Somerset. Proc. Bristoi Nat. Soc., ser. 4, 3, 47-50, 1 pi. CUMBERLAND, G. 1826. Reiiquiae Conservatae. 45 pp., 12 pi. Gutch, Bristol. ETHERIDGE, R. and CARPENTER, p. H. 1886. Cataiogue of Biastoidea in the Geoiogicai Department of the British Museum. London. FAY, R. o. 1960. The type-species of Orbitremites Austin & Austin 1842, and Eiiipticobiastus, a new Mississippian genus. Okia. Geoi. Notes, 20, 315-17, figs. 1-7. 1961. Nymphaeobiastus, a Mississippian blastoid from Japan. Ibid. 21, 150-3, pi. 1. HAMBACH, G. 1884. Notes about the structure and classification of the Pentremites. Trans. St. Louis Acad. Sci. 4. 537-47. JAEKEL, o. 1918. Phylogenie und System der Pelmatozoen. Paidont. Z. 3, 1-128. LUCAS, M. G. 1953. Etude, au microscope polarisant, des hydrospires des blastoides; in Traite de Paieontoiogie, 3, 635-7, figs. 10-13 (ed. J. Piveteau). PHILLIPS, w. 1936. The structure of Acentrotremites. Proc. Swansea sci. Fid. Nat. Soc. 1, 360-6. QUiRiNG, H. 1939. Die Ostasturischen Steinkohlenbecken. Arch. Lagerst Forsch. 69, 1-69, pi. 1-3. RAUP, D. M. 1959. Crystallography of echinoid calcite. J. Geoi. 67, 661-74. REiMANN, J. G. 1945. New Devonian blastoids. Buii. Biijfaio Soc. not. Sci. 19, 22-42, pi. 5-9. 1950. Possible phylogenetic relationships of some early eublastoids. /. Paieont. 24, 499-500. siEVERTS-DORECK, H. 1951. Echinodeimen aus dem spanischen Ober-Karbon. Paidont. Z. 24, 104-19, pi. 8, figs. 3, 4. SITTER, L. u. de 1955. Nota previa sobre la geologia de la cuenca Carboniferadel Rio Pisuerga (Palencia). Estud. Geoi. 11, 115-26, pi. 23. THWAiTES, J. E. and SAYERS, c. F. 1955. A novel type of saw for the economical cutting of quartz crystals or other materials. P.O. eiect. Engrs'. J. 47 (Pt. 4), 1-3. WACHSMUTH, c. and SPRINGER, E. 1890. New species of crinoids and blastoids from the Kinderhook Group of Lower Carboniferous rocks at Le-Grand, Iowa. Geoi. Surv. liiinois, 8, 155-208. WANNER, J. 1924. Die permischen Blastoiden von Timor. Jaarb. Mijnw. Ned.-Oost-Ind., Verb., jaarg. 1922, 161-233, 5 pi. 1940. Neue Blastoiden aus dem Perm von Timor. Geoiogicai Expedition to the Lesser Siinda Isiands under ieadership of H. A. Brouwer, 1, 215-77, 3 pi. WHITE, c. A. 1865. Descriptions of new species of fossils from the Devonian and Carboniferous rocks of the, Mississippi Valley. Proc. Boston Soc. nat. Hist. 9, 20-33. WHITFIELD, R. p. 1895. Republication of descriptions of Lower Carboniferous Crinoidea from the Hall Collection now in the American Museum of Natural History, with illustrations of the original type specimens not heretofore figured. Mem. Amer. Mas. nat. Hist. 1, 36. YAKOVLEV, N. N. 1926. Sur le Cystobiastus, Nymphaeobiastus et Acrocrinus. Buii. Com. geoi. 45, 43-49, pi. 1, fig. 4. K. A. JOYSEY University Museum of Zoology, Cambridge A. BREIMER Geologisch en Mineralogisch Instituut, Rijksuniversiteit te Leiden Manuscript received 5 October 1962 DEVONIAN GONIATITES AND STRATIG R APHIC AL CORRELATIONS IN WESTERN CANADA by M. R. HOUSE and a. e. h. redder Abstract. A systematic account of available Devonian ammonoids from western Canada is given and the evidence they provide concerning correlations within the Canadian Devonian is reviewed. Most of the material described has been provided by oil companies which have been exploring along the Mackenzie River and ad- jacent areas, but an attempt has been made in addition to study all goniatite material from this region previously mentioned in the literature. Goniatites probably representing Emsian faunas occur in beds below the Hume Formation on the Ogilvie River and Eifelian goniatites occur in the Funeral Formation of the Northwest Territories. Givetian goniatites occur above the Hume Formation on Francis Creek and the terebratum Zone is represented in the Fort Creek Shale of the Ogilvie and Anderson River areas; it is also present in dark shale immediately overlying the Hume Formation in the Norman Wells area. The basal Frasnian lumilicosta Zone is represented in the Maligne Forma- tion of Alberta and in dark shales overlying the Beavertail Formation at Carcajou Ridge; it is also probably represented in the upper part of the Fort Creek Shale in the Ogilvie and Thunder River areas. Middle and late Frasnian goniatites have been examined from the Perdrix, Mount Hawk, Escarpment, Redknife, and Carcajou Mountain Formations but have not enabled the precise recognition of zones. Famennian faunas are rare, but a fauna of the upper Cheiloceras major Zone is recognized in the lower D5 beds, and the Platyclymenia major Zone may be recognized in the upper D5 beds of the Northwest Territories. A specimen probably referable to Platyclymenia is recorded from the Palliser Formation. No evidence of upper Famennian ammonoids has been found. The following genera are recorded for the first time from Canada: Gyroceratites (Lamelloceras), Teicherticeras, Anarcestes (Latanarcestes), Cabrieroceras, Maenioceras, Ponticeras, Lobotornoceras, Cheiloceras, and Sporado- ceras. A new genus, Sellagoniatites, is erected with S. discoides Waldschmidt as type species. Two new species are described, Sellagoniatites jacksoni and Teicherticeras lenzi. Amongst other revisions it is shown that Manticoceras septentrionale should be regarded as a synonym of M. cordiforme and Timanites occidentalis a synonym of T. keyserlingi. Devonian goniatites have been known in western Canada since 1887, when McConnell discovered a specimen now known to be Manticoceras on Hay River. Many more Frasnian goniatites have been found subsequently, but it is only in recent years that Middle Devonian and Famennian ammonoids have been found in the Northwest Territories and the Yukon. The recent discoveries have largely been due to the impetus to exploration provided by the discovery of oil in western Canada. This account is based mainly on material collected by The California Standard Company, the Geological Sur- vey of Canada, Pan American Petroleum Corporation, Shell Oil Co. of Canada Ltd., and Triad Oil Co. Ltd. We are particularly indebted to these organizations for making their specimens available to us. In addition we have attempted to re-examine all Devonian ammonoids previously recorded in western Canada. The new ammonoid evidence is interesting and important for two reasons. Firstly, a nearly complete ammonoid sequence has been identified from equivalents of the Emsian to the lower Famennian, which shows considerable affinity with the well-known European faunas. Secondly, the correlations to which the ammonoid evidence contri- butes help in the elucidation of the complex facies changes known within the Devonian of western Canada. This account is given in two parts; the first, by A. E. H. P., is a synthesis of the [Palaeontology, Vol. 6, Part 3, 1963, pp. 491-539, pis. 70-77.] 492 PALAEONTOLOGY, VOLUME 6 Devonian stratigraphy in western Canada in the light of the new evidence; the second part is by M. R. H. and is a systematic description of the ammonoids at present known. STRATIGRAPHY by A. E. H. Redder As would be expected from a mere consideration of the huge areal extent and con- siderable thickness of the Devonian in western Canada, the relationships of its various facies are complex. This has naturally led to the erection of a large body of stratigraphic nomenclature, which is constantly being expanded and modified with the acquisition of new information. Text-fig. 2 is a simplified correlation chart of the ammonoid-bearing Devonian strati- graphic sequences in western Canada. In preparing it an attempt was made to reconcile all the available palaeontological evidence. The main part of this section is an alpha- betically arranged discussion of the nomenclature and fossil content of rock units appearing in the chart. Many of the standard world ammonoid zones are now known to be present in western Canada. Their correlation with the accepted local brachiopod zones is indicated in Table 1. The Canadian Famennian brachiopods are at present under study by P. Sartenaer and it is expected that McLaren’s (1954) Famennian zones will be modified at least in name. No attempt is made here to anticipate Sartenaer’s work in any way and therefore McLaren’s early generic assignments are retained in parenthesis. STAGES EUROPEAN AMMONOID ZONES CANADIAN BRACHIOPOD ZONES Famennian (pars) Platyclymeuia ‘Niidirostra' utahensis ventricosa ‘Nudirostra gibbosa seversotd Cheiloceras ‘’Nudirostra gibbosa walcotti Frasnian Crickites holzapfeli Theodossia scopidonim Manticoceras cordatum Calvinaria albertensis Calvinaria inscidpta Pharciceras IimuUcosta O Calvinaria athabascensis Ladogioides kakwaensis Givetian Maeiiioceras terebratum Leiorhynchiis castanea o Maenioceras molariiim Sclmchertella adoceta Cabrieroceras roiivillei Table 1. A comparison of the western Canadian brachiopod zones with the standard European ammonoid zones. Beavertail Formation The limestone forming the Ramparts of the Mackenzie River was considered by Bos- worth (1921a, p. 287) and by Kindle and Bosworth(1921, pp. 45-47) as being distinct and older than the limestones upstream at Beavertail Point and Carcajou Rock. The belief M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 493 TEXT-FIG. 1. Map of western Canada showing the position of the main localities referred to in the text. C 1456 K k 494 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 2. Correlation chart of the Devonian rock units in western Canada. Ammonoid occurrences indicated by solid circles. M. R. HOUSE AND A. E. H. PEDDER; DEVONIAN GONIATITES 495 led them to erect two names, Rampart(s) and Beavertail, for them, but subsequently it was recognized that these are approximately synonymous and the latter was suppressed (Hume and Link 1945, p. 21). Unfortunately Rampart had been pre-empted by Spurr (1898, pp. 155-69) for a Mississippian formation in Alaska and in view of this Bassett (1961, pp. 492-4) applied the name Kee Scarp to the Beavertail and Ramparts Limestones. However, the relationship of the type Kee Scarp lentil to other limestones with diagnostic Givetian fossils is so controversial, that at present it is considered wiser to confine the name Kee Scarp to the type locality and provisionally to reinstate the name Beavertail for the definitely Givetian limestones in the Norman Wells and Ram- parts regions. At Beavertail Mountain and the Ramparts only about 200 and 250 feet respectively of the formation are preserved, but according to Bassett (1961, p. 493) a maximum thickness of 800 feet has been measured about 55 miles south of the Ramparts. Warren and Stelck’s (1949, pp. 142, 143; 1950, p. 73; 1956, pi. 8, 9) 'Cyrtina panda' fauna, said to have come from the Beavertail Formation at Carcajou Ridge and Rock, may have been collected from both above and below the upper limit of the formation. There is also very good evidence, which was personally demonstrated to the writer by D. J. McLaren, that there is at least some mixing of the published localities of these specimens. The presumed younger species of the fauna, such as AUanaria sp. and Spinocyrtia euruteines (Owen), may have come from thin limestones at the base of the overlying dark shale unit, which, on the basis of its ammonoids, is very early Frasnian in age. Other species such as Leiorhynchus hippocastanea (Crickmay), Hadrorhynchia sandersoni (Warren), Warrenella timetea Crickmay, and Stringocephalus aleskamis Crickmay are normal elements of the Beavertail fauna. They are also typical Givetian forms and together with the overlying Frasnian goniatites, allow the Givetian-Frasnian boundary to be fixed, in the Carcajou Ridge area, with unusual precision. Canal Formation The shales immediately above the Beavertail, and possibly also the Kee Scarp Forma- tion in the Norman Wells area, are particularly bituminous. The geologists of the Canol Project referred to them informally as the bituminous shale member of the Fort Creek Formation (Stewart 1945; Hume and Link 1945); later Bassett (1961, pp. 494, 495) formalized their status by erecting the Canol Formation for them. The type section is nearly 55 miles due west of Norman Wells and is 75 feet thick, but the Canol Formation’s thickness in the Norman Wells area is generally greater and varies between 100 and 400 feet. Since it is thickest where the underlying Beavertail Limestone is thinnest, and vice versa, these formations are possibly at least partly correlatives, although it is equally possible that depositional draping over an uneven sea-floor was responsible for this. Bassett recognizes the formation over a very wide area; however, it is clear that to the north and south of the Norman Wells region, he includes in it beds which are here referred to the upper shale unit and Horn River Formation. Furthermore the reported goniatite impressions from the ‘Canol’ exposures at Carcajou Ridge (Bassett 1961, p. 495) may have come from the same horizon as those shown here to belong to the hmidicosta Zone. Carcajou Mountain Formation Kindle and Bosworth first recognized this unit and proposed (1921, p. 48) the name 496 PALAEONTOLOGY, VOLUME 6 Bosworth Sandstone and Shale for it. Independently, however, Bosworth (1921a, p. 287; 19216, p. 282) in papers published in London named it the Camp Creek Series. The first of these papers predates their joint publication by a few days, but this is irrelevant, since both names were preoccupied, Bosworth by a Cambrian formation in British Columbia (Walcott 1908, pp. 2, 3) and Camp Creek by a member of a Pennsylvanian formation in Texas (Drake 1893, pp. 402, 416, 417). Consequently Kindle (1936, pp. 14, 15) erected Carcajou Mountain Beds as a replacement. Hume and Link (1945, pp. 34, 35) rejected the new name on the grounds that its type section is incomplete and attempted to replace it by a new formation, which they called Imperial. Not only was this un- necessary, but their choice of name was unfortunate, as Imperial had been used by Hanna (1926, pp. 434-5) for a Miocene formation in California. In this paper, therefore, a return to the name Carcajou Mountain Formation is advocated. The formation is composed of dark shales interbedded with greenish-grey, partly calcareous and argillaceous, micaceous siltstones and sandstones. In the Root and Dahadinni River regions it includes prominent biostromal limestones. The thickness of the formation is extremely variable due to its diachronous lower contact and subsequent erosion in many areas, as well as to dilferences in depositional thickness. Incomplete sections about 60 miles west-north-west of Norman Wells are about 2,000 feet thick, whereas a complete section in the Root River area is only 1,360 feet thick. Incomplete sections at Carcajou Ridge and Norman Wells are much thinner. The unit can be traced as far south as the lower reaches of Liard River, where it passes laterally into silty lime- stones of the Grumbler Group. The uppermost beds in the Root River region contain early Famennian brachiopods of the Cyrtiopsis portae and C. prepta species group, but the main part of the formation contains middle and late Frasnian corals and brachiopods. Miller (1938) described Manticoceras cordiforme and M. septentrionale as new species from near the top of the formation at Norman Wells. Their types have been re-examined and are considered to be specimens of the same species, for which the name M. cordiforme is retained. Another species reported from the formation (Hume and Link 1945, p. 39; Hume 1954, p. 46) is M. intumescens (Beyrich). The specimen on which this record is based also came from near the top of the preserved formation at Norman Wells and is now deposited in the Walker Museum, Chicago. Through the kindness of M. H. Nitecki it has been loaned to the writers and proves to be another example of M. cordiforme. The species is considered to be indicative of either the upper cordaturn or the hohapfeli Zone. Chinchaga Formation Cameron (1918, p. 25) proposed the name Fitzgerald Formation for various, partly Devonian, carbonate and gypsum beds occurring below the Pine Point Formation in the southern Great Slave Lake area. Because its definition is in doubt, Douglas (1959, p. 9) recommended that Cameron’s formational name be allowed to lapse. Subsequently, officers of the Geological Survey of Canada have referred the Devonian part of the original Fitzgerald Formation to the Chinchaga Formation. This practice is followed here. The Chinchaga Formation was proposed by Law (1955, pp. 1963-4) for a sub- surface evaporitic unit in northern Alberta; its type section is in California Standard’s Steen River 2-22-1 17-5W.6M well and is 205 feet thick. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 497 D4 to D7 units Throughout most of the Northwest Territories Famennian beds are either poorly exposed, or have been removed by erosion. Good exposures are presently confined to the Yohin syncline, just north of the lower reaches of the North Nahanni River. Hume (1922, pp. 71-73) first described the Famennian succession in that area and recognized four units above what are now termed the Fort Simpson Shales. In ascending order, he designated these units D4 to D7. Except for D4 they are here retained without modification. The exception is necessary not only because Hume failed to recognize the Carcajou Moun- tain Formation as a separate entity, but also because he failed to assign its beds con- sistently to the same unit. For instance, on Root River, beds which include a biostrome and are now assigned to the Carcajou Mountain Formation were included in the D4 unit, whereas on North Nahanni River they were excluded from it. In this paper the D4 unit is used in the way Flume used it on North Nahanni River. Thus defined it consists there of 830 feet of soft light-grey argillaceous microgranular limestone containing abundant specimens of Basilicorhyncinis basilicum (Crickmay). The exceptional abundance of this brachiopod, which does not occur in the underlying Carcajou Mountain Formation, prompted Hume to designate the unit the " Leiorhynchiis Zone’. The D5 unit is an olive and maroon weathered shale interbedded, especially in the upper part, with thin calcareous siltstones. Hume also called it ‘Shale Zone No. 2’. In Yohin syncline it is 1,178 feet thick. Except for a few beds crowded with brachiopods, the unit is sparsely fossiliferous, but the lower part has yielded a particularly interesting suite of goniatites, including Lobotornoceras aff. bilobatum (Wedekind), Cheiloceras (C.) sacculum (G. and F. Sandberger), Sporadoceras cf. primaevuin Schindewolf, and Imito- ceras sp. These are indicative of the upper Cheiloceras major Zone. Higher in the unit, the presence of a Tornoceras comparable with T. crebriseptuni (Raymond) of the Three Forks Formation of Montana, probably indicates that the upper part of the unit is to be assigned to the Platyclymeuia major Zone. The D6 unit, or ' Athyris angelica Zone’, is a thick-bedded, light grey, buff-weathered, microgranular, and partly silty limestone. It is about 90 feet thick and contains a rich Famennian brachiopod fauna, but to date has not yielded ammonoids. The D7 unit, or ‘Shale Zone No. 3’, consists of dark grey, non-calcareous shale, of which only the lower 325 feet are preserved in the Yohin syncline. It is barren of macro- fossils and may be Devonian, or Mississippian, or both. Dark shale units Since the days of Canol Project drilling, dark, variably bituminous shales with minor siltstone and limestone interbeds have been known to lie above and below the oil- producing limestone of the Norman Wells area. These have most frequently been called the Upper and Lower Fort Creek Shales (Hume and Link 1945, p. 29; Stewart 1945, pp. 5, 6; Warren and Stelck 1949, p. 143; 1950, fig. 1; Hume 1954, p. 35; Storey 1961, p. 499), but Bassett (1961, pp. 490-2), who followed Crickmay (1957, p. 1 1) in disagree- ing with earlier correlations, applied the name Hare Indian to the lower unit and included the upper unit in the Canol Formation and lower part of the Imperial Formation. In this paper the shales are provisionally and informally referred to as the upper and lower 498 PALAEONTOLOGY, VOLUME 6 dark shale units. Drilling results in the norman Wells region (Hume and Link 1945, pp. 79, 82; Hume 1954, p. 35) indicate that the thickness of the lower unit varies between 385 and 780 feet and that of the upper between 660 and 840 feet. The upper may be an exaggerated tongue of the Fort Creek Formation extending southwards over Givetian limestones, but at the present time this possibility is difficult either to prove or disprove, because erosion has worked deep in critical areas. On Bosworth and Francis Creeks, near Norman Wells, the lower beds of the lower dark shale unit have yielded: Agoniatites cf. fidgiiraUs (Whidborne), A. cf. vamixemi (Hall), Cabrieroceras karpinskyi (Holzapfel), C. sp., Tornoceras {T.) cf. westfalicum (Holzapfel), Leiorhynchus castanea (Meek), WarreneUa sp., EmamieUa sp. and various coniconchines. This is a Givetian fauna and enables the lower dark shale unit to be correlated with the lower part at least of the Fort Creek Shale to the north and the Horn River Shale to the south. It also demonstrates a Givetian age for the castanea brachiopod Zone. An important goniatite assemblage, consisting of Ponticeras cf. tschernyschewi (Holzapfel), P. sp., Probeloceras sp. and anaptychi, has been collected by A. C. Lenz 10 to 40 feet above the base of the upper dark shale unit at Carcajou Ridge. The fauna strikingly resembles that from the earliest Frasnian of Timan and is indicative of the hmulicosta Zone. Duvernay Formation The unit was proposed (Geological staff, Imperial Oil Limited 1950, pp. 1817-19) for 175 feet of partly bituminous and calcareous brown and black shales in the Anglo- Canadian Beaverhill Lake 11-11 -50-1 7W.4M well. This locality is east of the Rimbey- Meadowbrook Leduc reef chain of the Edmonton area (Belyea 1956, fig. 1), where Duvernay lithology is developed below argillaceous limestones of the Ireton Formation and above clastic limestones of the Cooking Lake Formation. Although Manticoceras cf. simulator is said (Alberta Soc. Petrol. Geol. 1960, p. 107) to be one of the distinctive fossils of this formation, we have seen no ammonoids from it east of the reef chain. However, Calvinaria insculpta (McLaren), Leiorhynchus carya Crickmay, and WarreneUa nevadensis (Walcott) have been obtained from Duvernay cores in this region, indicating that the formation, in the vicinity of its type locality, is a correlative of the Perdrix Formation of the mountains. The statement that Manticoceras occurs in the formation is probably based on Crickmay’s (1950, p. 223) report that M. cf. simulator Hall was found in black shale between 7,720 and 7,980 feet in the Imperial Paddle River 5-17-56-8W.5M well. He has more recently stated (personal communication, 17 April 1962) that ‘some queer objects from 7,730 feet in that hole were at first thought by me to be Manticoeeras opercula, but later proved to be an arthropod — ’ and that the only true ammonoids were obtained from 7,913-7,932 feet. These have been re-examined and identified as Probe- loceras. They are poorly preserved juveniles and it is not possible to assign them specifically, but they are thought to indicate the hmulicosta Zone. Further specimens from between 9,990 and 9,995 feet in the Imperial Amerada Willesden Green 3-16-43- 5W.5M well are possibly the same. Both of these occurrences are west of the Rimbey- Meadowbrook reef chain, where the Duvernay Shale is thicker and extends down to the top of the Beaverhill Lake Formation and is therefore, in part, older than on the M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 499 eastern side of the reef chain. Both occurrences are less than 62 feet above the top of the Beaverhill Lake Formation, the upper part of whieh, on the western side of the reef chain, has yielded AUanaria minutilla Crickmay, Cahinaria athabascensis (Kindle), and other fossils indicative of the Maligne Formation of the mountains. The Duvernay- Beaverhill Lake formational boundary, therefore, west of the reef chain, falls within the hmuUcosla Zone and correlates with a horizon within the Maligne Formation of the mountains. That the older dark shales to the west of the Rimbey-Meadowbrook reef chain are equivalent to the limestones of the Cooking Lake Formation to the east has long been known, and they are now widely referred to as the Majeau Lake Member of the Cooking Lake Formation (Alberta Soc. Petrol. Geol. 1960, p. 213, and Alberta Oil and Gas Conservation Board 1960). At the time of writing, however, this unit has not been formally proposed and the older practice of applying the name Duvernay to the entire dark shale unit immediately west of the reef chain is followed on the correlation chart. Escarpment Formation Cameron (1918, pp. 25, 26) established the name Hay River for the entire Upper Devonian section exposed on Hay River and recognized in it an upper limestone and a lower shale division. Warren and Stelck (1950, fig. 1) expanded the term to embrace Hume’s D4 to D6 units in the Root River region, but Crickmay (1953, p. 11) restricted it almost to Cameron’s lower division by taking its upper limit at the base of Alexandra Falls on Hay River. This is stratigraphically about 50 feet higher than the upper limit of Cameron’s Hay River Shale, which was drawn at the base of Louise Falls on the same river (Belyea and McLaren 1962, p. 2). Douglas (1959) remapped the type area without applying the name to any of his map units. Belyea and MeLaren (1962, pp. 2, 3) rede- fined the formation yet again, by assigning to it all the beds below the base of Alexandra Falls and above the Slave Point Formation in the type area. They also introduced the term Esearpment Member for the upper 364 feet of the formation, which are more calcareous and biostromal and contain a rich brachiopod and coral fauna. In this paper the name Hay River is abandoned, the Escarpment Member is elevated to formational status, and the underlying beds down to the Slave Point Formation are assigned to the Fort Simpson Formation. The main components of the fauna of the Escarpment Formation are: Endothyra gallowayi Thomas, Phillipsastrea nevadensis Stumm, Macgeea proteus Smith, Tabido- phyllwn mcconnelli (Whiteaves), Nervostrophia maclareni Pedder, N. vestita Crickmay, Cahinaria variabilis (Whiteaves), Cyrtospirifer occidentalis (Whiteaves), C. thaJattodoxa Crickmay, Warreneila labrecqnei Crickmay, and Manticoceras sp. It is of interest to note that one of the two known specimens of Manticoceras sp. is the first indubitable Devonian ammonoid to have been found in western Canada. It was collected by MeCon- nell in 1887 and has been figured by both Whiteaves (1891) and Miller (1938). Although the specimens of Manticoceras are not identified specifically, they are consistent with the assignment of the Escarpment Formation to the lower cor datum Zone. Exshaw Formation At its type locality on Jura Creek, near Banff the formation consists of 31 feet of black shale, with a basal parting of sandstone, overlain by 37 feet of silty argillaceous 500 PALAEONTOLOGY, VOLUME 6 limestone. In spite of its thinness, the formation is widely distributed and can be traced throughout the Alberta Rocky Mountains and in wells eastwards to the pre-Cretaceous erosion surface. Warren (1937, p. 456) created the name but the shale was discovered as long ago as 1886 by McConnell (1887, p. 18). A species of Imitoceras occurs in the upper silty limestone on Jura Creek. Crickmay (1952, p. 590) identified it asAganidessp. close to A. discoidalis Smith ; previous identifica- tions include Clymenia (McConnell 1887, p. 18), Phylloceras (Raymond in Warren 1956, p. 141), and Tornoceras sp. cf. T. imiangidare (Conrad) (Miller 1938, p. 166; Warren 1937, p. 455; Warren and Stelck 1950, p. 63; 1956, pi. 29, figs. 1-3). At Crowsnest Pass, about 100 miles south-east of the type locality, Copeland ( 1960) described a Mississippian ostracod and conodont fauna from the same part of the formation. The lower black shale is normally without macrofossils, but a dwarfed and pyritized fauna, including Imitoceras aff. discoidale (Smith), I. spp., Protocanites cf. gurleyi (Smith), and Prodro- mitesl sp., is known from near Mount Arete, Alberta (Pamenter 1956; Schindewolf 1959). Although originally considered Devonian, the faunal evidence now indicates that this formation is very early Mississippian and that the Clymenia and Wocklumeria major Zones are absent in the Canadian Rocky Mountains, since the Exshaw Formation rests disconformably on the Platyclymenia fauna of the Palliser Formation. Flume Formation The excessive thicknesses quoted in the proposal of the formation (Raymond 1930, pp. 294, 295, 297) suggest that Cambrian as well as Devonian beds were included in it. Subsequently de Wit and McLaren (1950, p. 4) restricted the formation to Devonian carbonates below the Perdrix Shale and divided it into two members. The upper is a dark to black slightly argillaceous limestone, whereas the lower is a coarser and lighter cherty dolomite and limestone unit. At the type locality on Roche Miette, near Jasper, these members are 50 and 101 feet thick respectively (McLaren 1956, p. 26). The name was further restricted by Taylor (1957, p. 187), who confined it to the lower of de Wit and McLaren’s members and established a new formation, named the Maligne, for the upper. The revision is generally accepted and is incorporated in the correlation chart. On comparing the succession of faunas in the McMurray region with that in the Rocky Mountains, it is evident that the Flume Formation of the northern Alberta Rockies is older than that of the central and southern Alberta Rockies. This is probably due partly to onlap and partly to change from Ghost River to Flume facies. The older northern Flume Formation contains the Ladogioides kakwaensis fauna, but so far has not yielded ammonoids. Fort Creek Formation This is a dark bituminous shale with minor fine sandstones and limestones. In a recon- naissance study. Kindle and Bosworth assigned all Devonian beds of this lithologic type in the central and lower Mackenzie region to a single formation, which they named the Fort Creek Shale (Bosworth 1921n, p. 287; Kindle and Bosworth 1921, pp. 47, 48). Because shale of this type overlies Givetian limestone at Carcajou Ridge, it was regarded as being of Upper Devonian age everywhere. Other geologists (Warren 1944, p. 107; M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 501 Hume and Link 1945, p. 6; Stewart 1945, p. 3; Warren and Stelck 1949, p. 139; 1956, p. 11; Hume 1954, p. 34; and Storey 1961, p. 499) accepted the opinion. The type locality of the formation is at the mouth of Thunder (previously Fort Creek) River, where only about 175 feet and neither contact is exposed. Fortunately an Agonia- tites has been found there, indicating that the type Fort Creek Shale is at least partly Middle Devonian. Richer Middle Devonian faunas occur in the formation at several places in the Anderson drainage region. The lowest beds contain Leiorhynchus castanea (Meek), Cassidirostnim pedderi McLaren, WarreneUa cf. kirki (Merriam), and other forms, while what are believed to be higher beds have yielded Sellagoniatites jacksoni sp. nov., Wedekindella brilojtense (Kayser), W. aff. brdouense (Kayser), W. sp. (uncon- stricted form), Maeuicoceras cf. terebratum (G. and F. Sandberger), and M. sp. The goniatites are late Givetian and indicative of the terebratum Zone, although two of them had been previously misidentified as Upper Devonian forms (Warren and Stelck 1949, p. 142; 1950, p. 73; see House 1962, p. 255). Farther west the Fort Creek Forma- tion has yielded Agoniatites cf. fidgiiralis ( Whidborne) from a locality on Ogilvie River and Agoniatites sp. from Hungry Lake. These are also Givetian forms. In addition to this overwhelming evidence that much of the Fort Creek Shale in the Ogilvie, lower Mackenzie, and Anderson River regions is Middle Devonian, there are also indications that it is partly early Frasnian. On Ogilvie River the formation has yielded a single fragment of a goniatite body chamber with retracted growth-lines, which appears to be a species of Ponticeras. Furthermore, from the type section of the Fort Creek Shale, there are two anaptychi with outlines also suggesting a species of Ponticeras. Their preservation is similar to the anaptychi occurring in the upper dark shale at Carcajou Ridge, which is definitely of early Frasnian age. Fort Simpson Formation In the upper Mackenzie region, much of the Frasnian (and particularly the early part) is represented by meagrely exposed soft, grey, and partly calcareous shales. Cameron (1918, pp. 25, 26) introduced the name Simpson for limited exposures of these in the vicinity of Fort Simpson. The results of many wells drilled since Cameron’s work show that shales in the Fort Simpson region are part of the same lithological unit as the shales outcropping on Hay River below the Escarpment Formation. Cameron believed that the shales at Fort Simpson were older than those on Hay River. However, the opposite is true, since the fossils of the former (Kindle 1919) are now recognized as elements of the Calvinaria albertensis fauna. Douglas and Norris (1961, pp. 20, 21) proposed a slight modification of Cameron’s name to Fort Simpson, because Simpson is preoccupied by an Ordovician formation in Oklahoma (TalT 1902, p. 3). Belyea and McLaren (1962, p. 9) designated the section in Briggs Turkey Lake No. 1 well (61° 07' 30" N. ; 120° 22' 30" W.), where the formation is 1,940 feet thick, as the type. Funeral Formation West of the Nahanni and Camsell Ranges dolomites of the Arnica Formation are mostly replaced by grey argillaceous limestones and calcareous shales, for which Douglas and Norris (1961, pp. 17, 18) proposed the name Funeral Formation. The beds characteristically weather light rusty grey and are fissile, but also include, particularly in the northern part of their development, some purer and more massive weathering 502 PALAEONTOLOGY, VOLUME 6 limestones. In the southern Funeral Range, the Headless and Nahanni Formations simi- larly grade laterally into beds of this formation. In the regions of its greatest develop- ment the formation is nearly 2,000 feet thick. Kingston (1951, p. 2411) described a ‘ Devonian black fissile shale’ in the area in which it occurs, but his unit is a misidentified Silurian shale. Apart from abundant coniconchines the Funeral Formation is sparsely fossiliferous, although Gyroceratites (Lamellocems) sp. and Agoniatites sp. have been found low in the formation and Anarcestes (Latanarcestes) cf. praecursor Freeh is recorded 469 feet below the top. Both probably indicate Eifelian correlations. Grumbler Group Crickmay (1953, p. 1 1) divided the Frasnian limestones of the upper Mackenzie region into the Alexandra and overlying Grumbler Formations. These were new names for what was virtually the Hay River Limestone of previous workers (Cameron 1918, pp. 25, 26; Warren and Stelck 1950, fig. 1). Following more detailed surface and subsurface stratigraphical studies, Belyea and McLaren (1962, pp. 4-8) raised the Grumbler Formation to group status, recognized four formations in it, and since the Alexandra Limestone appears to be only a very local facies, demoted it to member status at the base of the group. In ascending order their formations are the Twin Falls (mostly lime- stone with minor shales), the Tathlina (a mixture of calcareous siltstones and silty limestones), the Redknife (argillaceous and silty limestones in the lower part becoming more argillaceous in the upper part), and the Kakisa (characterized by biostromal and biohermal reefs). At their type localities these formations are 521, 440, 231, and 187 feet thick respectively. About 50 miles west of Great Slave Lake, the Twin Falls and Tathlina Formations pass laterally into the Fort Simpson Formation and west of the lower Liard River the Redknife and Kakisa Formations pass laterally into the Carcajou Mountain Formation. Many beds in the group have yielded prolific coral and brachiopod faunas, but ammonoids are not common. There are several specimens from the Redknife Formation on Liard River and Warren and Stelck (1956, pi. 24, figs. 2, 4) have figured another from the same formation on Trout River. All are specimens of Manticoceras, but at the present time it is not possible to assign them to species. Hare Indian Formation Kindle and Bosworth (1921, p. 45) erected the name Hare Indian River Shale for soft, greenish-grey, argillaceous limestones and calcareous shales, which are weathered fawn, and lie below the purer and more massive limestones of the Ramparts Formation in the Fort Good Hope area. The type section, on Mackenzie River a few miles above the mouth of Hare Indian River, exposes only about the top 100 feet of the formation. Probably a much thicker part lies concealed below, since at Imperial anticline, 60 miles south-south- west, the formation is 750 feet thick. Geologists of the Canol Project (Hume and Link 1945, pp. 19, 20; Hume 1954, pp. 23-25) discarded the name on the grounds that the formation is not sufficiently distinct to be mapped in the Norman Wells area. This was not justified; their correlations were at least partly inaccurate and at Norman Wells, 80 miles south-east of the type section, the Hare Indian is replaced by a dark shale facies. Warren and Stelck (1950, fig. 1 ; 1956, pi. 3) applied the name to M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 503 older beds on Anderson River, which are now included in the Hume Formation. Bassett (1961, pp. 490-2) broadened the formation in another way by including in it the distinct, but equivalent, lower dark shales of the Norman Wells area. North of the mouth of Ontaratue River there is a similar change in facies to the Fort Creek Formation. Headless and Hume Formations The beds discussed under this heading are very fossiliferous brown and grey, partly argillaceous limestones. On Anderson River their measured thickness is 197 feet, but southwards they thicken to between 400 and 600 feet in the central Mackenzie area, and still farther south, on Ram Plateau, about 80 miles west-south-west of Fort Simpson, to as much as 829 feet. The Hume Formation has long been known and referred to under several names, such as Ramparts (Warren 1944, pp. 120-2) or Hare Indian (Warren and Stelck 1950, fig. 1 ; 1956, pi. 3) in the Anderson River area, and Lower Ramparts Lime- stone Member (Hume and Link 1945, p. 19; Hume 1954, pp. 23, 24), or ‘Ramparts’ (Warren and Stelck 1950, fig. 1), or Lower Hare Indian Shale (Warren and Stelck 1956, pi. 1), or basal Devonian limestone (Crickmay 1960a, p. 1) in the central Mackenzie region. Such names could only be provisional and Crickmay (19606, p. 877) introduced the Norman Wells Formation for the unit, with the 2,010 to 2,570 feet section in the Discovery No. 3 well of the Norman Wells oilfield as type section. Bassett (1961, pp. 486-90), in a publication which crossed Crickmay’s in press, introduced the Hume Formation for the same unit, based on a surface section exposed on Hume River. Although Crickmay’s name has about nine months’ priority, article 116 of the Code of Stratigraphic Nomenclature is invoked to retain the name Hume Formation. The Hume Formation can be traced southwards into the Nahanni Formation, from which it is distinguished only by its greater argillaceous content and softer and browner weathering characteristics. The Headless Formation was proposed by Douglas and Norris (1961, pp. 19, 20) for a more argillaceous facies equivalent of the Nahanni Forma- tion, most obviously developed north and west of Nahanni Butte. Although the name is provisionally retained, it is most likely that the Hume and Headless Formations will be considered synonymous, in which case the name Hume should be retained since it has about four months’ priority over Headless. The Hume and Headless Formations have yielded extraordinarily prolific early Givetian faunas with particularly abundant corals and brachiopods, but surprisingly ammonoids have not been found. Horn River Formation The higher Givetian carbonates of the Great Slave Lake area change facies north- westwards and are replaced by dark, partly bituminous shales and minor limestones in the Horn Plateau region. Whittaker (1922, p. 52) studied limited exposures of the latter on Horn River and named them after that river. The West Territories Westerol No. 7a well (61° 39' 13" N. ; 120° 43' 34" W.) encountered dark bituminous shales between 1,441 and 1,597 feet, for which Hunt (1954, pp. 2300, 2301) proposed the name Spence River Formation, but since they are part of the same unit as the Horn River Shales, Hunt’s name is suppressed as a synonym. Limestones in the formation on Horn and Willowlake Rivers contain Leiorhynchus eastanea (Meek). Thus a Givetian age is established for at least part of the formation 504 PALAEONTOLOGY, VOLUME 6 and at the present time the Middle-Upper Devonian boundary, over a wide area around Fort Simpson, is placed at the Florn River-Fort Simpson formational contact. Kee Scarp Formation For reasons given previously (Pedder 1963) this name is at present used only in a most restricted sense for the type Kee Scarp lentil. Maligne Formation The history of the study of these beds is outlined in the section on the Flume Forma- tion. The Maligne Formation includes the type horizon of Timanites occidentalism which is now regarded as being synonymous with T. keyserJingi Miller. The importance of this occurrence lies in that it firmly correlates the Calvinaria athabascensis brachiopod fauna with the upper part at least of the lunulicosta Zone. Manticocerasha.'t, been reported from the Maligne Formation on several occasions, but it is unlikely that the genus sensu stricto occurs in beds as old as these. The reports by Allan et al. (1932, p. 236) and Fox (1951, pp. 827, 828) of M. cf. oxy are based on the holotype of Timanites occiden- talis. Taylor’s (1957, p. 190) report of Manticoceras is based on material which neither he nor Shell Oil Co. are now able to locate, and McLaren’s material (1954, p. 169; Belyea and McLaren 1956, p. 89) was not among that loaned to the writers by the Geo- logical Survey of Canada. Mount Hawk Formation This formation was proposed by de Wit and McLaren (1950, p. 5) for fossiliferous, argillaceous limestones and dolomites occurring in the clastic facies of the Alberta Rocky Mountains. Its thickness is variable, but is typically about 550 feet. Since its inception the limits of the type section have been modified (McLaren 1956, pp. 15, 16), at the base by the addition of the upper 156 feet of the Perdrix Formation of the earlier classification, and at the top by the removal of 82 feet to the Alexo Formation. It has also been refined by division into members, but for the sake of clarity these are omitted in the correlation chart. There were several earlier classifications of the beds now assigned to this formation. The first of these was Raymond’s (1930, pp. 294-6, 300), which included the beds in the lower part of the Fiddle and Boule Formations. Apart from interests of refinement. Fox (1951, pp. 823, 825) has given good reasons why much of Raymond’s nomenclature is unsatisfactory. Other classifications included this formation in the upper part of the Blackface Formation (Kelly 1939, p. 2000) and in the lower part of the Cheviot For- mation (Fox 1951, pp. 830-4). As a result of considerable subsequent work, these terms have lost most of their earlier usefulness and are now obsolete. The first goniatites to have been recorded from this formation are said to be in the Cornell University collection. Merriam(1940, p. 73) referred to them as ‘goniatites which appear to be of the Manticoceras type’ and stated they came ‘from beds immediately above the Miette black shale’ (i.e. in the Mount Hawk Formation). Unfortunately, in spite of Professor Wells’s efforts to find these specimens, they remain lost, but copies of the plates in Bissell’s unpublished thesis (1930, pi. 8) clearly show that Manticoceras is present. Warren’s (1949, p. 569) report of Manticoceras in his jasperensis Zowq (i.e. in the Maligne to Mount Hawk Formations of present usage) is worded to suggest that M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 505 it is based on Merriam’s report and therefore on the same specimens. However, the specimens of Manticocems reported from the Mount Hawk Formation by McLaren (1956, pp. 25, 30; Belyea and McLaren 1956, p. 89) have, through his kindness, been restudied and are identified as M. cf. simioswn (Hall). Several other specimens, also referred to M. cf. sinuosum, have been collected from the Calvinaria albert ensis beds of the lower part of the formation. These occurrences clearly establish a correlation between part of the cordatum ammonoid Zone and the albertensis brachiopod Zone. Nahanui Formation Hage (1945, pp. 5, 6) established the Nahanni Formation for dark and fine-grained clastic limestones, that commonly form massive light grey cliffs. At the type locality it caps a prominent feature known as Nahanni Butte and is about 450 feet thick. Sections to the north and west, however, are thicker. Douglas and Norris (I960, pp. 14, 15) redefined the formation by removing the lower 130 feet of the original type section to what they later (1961, pp. 19, 20) called the Headless Formation. The limestones of the type section were first mentioned by McConnell (1891, pp. 56, 57), who in spite of noting that the only coral in them had the external appearance of a Carboniferous genus, seems to have considered them Devonian. Cameron and Warren (1938, p. 17), who also described these limestones, assigned them more definitely to the Devonian. Farther north, the upper 130 feet of Kindle and BoswortlTs (1921, p. 44) Lone Mountain section is the lower part of the Nahanni Formation and was correctly assigned to the Devonian. No ammonoids have yet been found in the Nahanni Formation, but Givetian corals and brachiopods are common in some beds. North Nahanni River Group The first classification of the carbonates forming this group was made when Mc- Connell (1891, pp. 14, 56) referred to the banded dolomites at Nahanni Butte as ‘un- doubtedly’ belonging to the Castle Mountain Group. Subsequently dolomites of the same unit, exposed on Lone Mountain near the mouth of North Nahanni River, were named the Lone Mountain Dolomite (Kindle and Bosworth 1921, pp. 44, 45), but since this name was preoccupied by a Silurian formation in Nevada (Hague 1883, pp. 253, 262, 267), Kindle (1936, pp. 14, 15) replaced it by the term North Nahanni River Dolomite. The type area was remapped by Douglas and Norris (1961, pp. 14-19), who deliberately ignored previous nomenclature and proposed three formations for the old Lone Mountain Formation. In descending order these are the Landry (grey crypto- crystalline limestone), Manetoe (coarse, vuggy and brecciated dolomites), and Arnica (dark, banded, fine-grained, and dense dolomites). Following these proposals the old North Nahanni River Formation is raised to group status to embrace the new forma- tions. Along its most eastern outcrop the group is between 2,000 and 3,000 feet thick, but about 45 miles to the west (that is, in those places where the group is not replaced by the Funeral Formation), thicknesses of more than 5,000 feet have been measured. The group yields so few determinable fossils that Bosworth (1921a, p. 287) once named it the Barren Series. However, Givetian corals have now been found in the group and also it can be shown in the field to pass laterally into the Funeral Formation, which 506 PALAEONTOLOGY, VOLUME 6 contains Eifelian ammonoids. Thus its age is Middle Devonian and earlier suggestions that it is Cambrian or Silurian are erroneous. PaUiser Formation Beach (1943, pp. 15-17) established this name for massive very fine-grained grey limestones and partly mottled dolomites, which constitute one of the great mountain building units of the Rockies and commonly exceed 1,000 feet in thickness, especially in the western ranges. De Wit and McLaren (1950, p. 6) refined it by proposing a lower Morro and much thinner upper Costigan Member. Most of Walcott’s (1924, p. 51) Pipestone Formation is identical with the Palliser, but the name never gained popular usage and in any case is preoccupied by a Pre-Cambrian quartzite in Minnesota (Win- chell 1888, p. xxii). A clymenid, tentatively identified as Platyclymenia, has been found in the Costigan Member. It occurs with Leioproductus sp. cf. L. coloradensis (Kindle), ‘ Niidirostra ’ ventri- cosa (Haynes), " Camaroloechia' nordeggi Kindle, and Cyrtiopsis sp. Thus the local ventricosa brachiopod Zone can now be tentatively directly correlated with the Platy- clymenia major Zone. Previously both Warren (1927, p. 20) and Taylor (1958, p.l5) hinted that clymenids are present in the Palliser Formation by listing Wlatyclymenia arnericana Raymond and cf. Platyclymenia respectively, from it. Unfortunately the specimens in both cases are now lost. Perdrix Formation The Perdrix Formation consists of dark to black, partly calcareous shales, or, especially towards the carbonate facies, of dark nodular argillaceous limestones. The fullest sections are between 500 and 600 feet thick, but the formation thins and becomes less distinct towards the carbonate facies. The unit was first named Miette by Kindle (1929, pp. 180, 184). This name, however, had already been pre-empted by Walcott (1913, pp. 335, 340) for Pre-Cambrian sandstones in British Columbia and Alberta. At about the same time the unit was also recognized by Raymond (1930, pp. 295, 296) and Kelly {in Allan et al. 1932, pp. 234, 235; Kelly 1939, p. 2000). A misunderstanding of the structure on Roche Miette led Raymond to confer two names on the formation, the Perdrix and the now discarded Kiln. Kelly termed the unit the Blackface, or Blackface Mountain Formation, but seems to have included in it beds now referred to the Mount Hawk Formation. In any case the name was never formally proposed and has passed out of use. The fauna described by Burgess (1931) is typical of the Perdrix Formation in the Jasper area. The more calcareous facies, however, contains Leiorhynchus carya Crickmay, Calvinaria inscidpta (McLaren), and Warrenella nevadensis (Walcott). The Perdrix goniatites mentioned by Raymond (1930, p. 296) and identified by Burgess (1931, p. 2000) as Tornoceras bicost at wn Hall, which is an Aulatornoceras, have been restudied and are now referred to Tornoceras {T.) sp. indet. The other specimens mentioned by Raymond (1930, p. 298) as Manticoceras sp. indet. and Gephuroceras sp. indet. have not been seen, but two specimens of Manticoceras cf. sinuosum (Hall) kindly loaned to the writers by the Geological Survey of Canada are from this formation. Although far from spectacular, these occurrences indicate that the inscidpta brachiopod Zone may be equated with part of the cordatum goniatite Zone. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 507 Slave Point, Amco, and Sulphur Point Formations Cameron (1918, p. 26) proposed the name Slave Point for limestones lying between the Simpson and Presqu’ile Formations in the Great Slave Lake area. It is now appreciated that the facies relationships in that area are more complex than envisaged by Cameron and that modifications of his classification are required. Law’s (1955, pp. 1943-9) suggestion that the Slave Point be restricted to beds above a thin shale interval named the Amco by Campbell (1950, p. 90) is followed here. The limestones between the Amco Shale and the Presqu’ile Dolomite are being named the Sulphur Point Formation (Norris, in press); this is also incorporated in the correlation chart. The Slave Point, Amco and Sulphur Point Formations on the southern side of Great Slave Lake are re- ported by Campbell (1950, pp. 90, 91) to be 170, 11, and 130-190 feet thick respectively. Precise correlations of these formations is not possible, mainly because of the lack of palaeontological control. However, Norris has found Stringocephahis in the Sulphur Point, Presqu’ile, and underlying Pine Point Formations, and Ladogioides has been found above the Slave Point Formation in the vicinity of Sulphur Bay (D. J. McLaren, personal communication). These occurrences confirm that the Slave Point and Sulphur Point Formations are Givetian. As pointed out by Law (1955, p. 1951), the Amco and at least some part of the Sulphur Point Formation are equivalent to the Watt Mountain Formation of north-western Alberta. SYSTEMATIC DESCRIPTIONS by M. R. House Family mimoceratidae Steinmann 1890 Genus gyroceratites von Meyer 1831 Type species by the subsequent designation o/ Schindewolf 1933: Gyroceratites gracilis Bronn 1835. Gyroceratites (Lamelloceras) sp. Plate 74, fig. 1 Material. One specimen preserved crushed and distorted in silty fissile shale. Remarks. The critical features of this specimen may be seen on the illustration. Crushed and displaced along the venter are shown paired ridges which appear to have extended as spiral sheets. This unusual feature is only known, among early tightly coiled goniatites, in certain species of Gyroceratites fox which Erben (1960, p. 78) has proposed the sub- generic name Lamelloceras, with G. dorsolamellatus Erben (1953, p. 185, pi. 17, figs. 7, 9, text-fig. 3; 1960 pi. 4, fig. 7, text-fig. 18fi) as type species. The ornament in the Canadian specimen is close to that known in the Eifelian G. (G.) gracilis, but may well be shown on species of Lamelloceras larger than those so far described. Lamelloceras ranges in Europe from the upper Emsian to the middle Eifelian. Horizons and locality. GSC 16921 (TOC G5058), collected by A. E. H. Pedder from 1,466-1,476 feet below the top of the Funeral Formation and 2,484-2,494 feet below the top of the Nahanni Formation (both measurements slightly excessive due to minor faulting in the section) in the northern Arnica Range, Northwest Territories, about 61° 55' N., 125° 13' W. 508 PALAEONTOLOGY, VOLUME 6 Genus teicherticeras Erben 1960 Type species by original designation: Gyroceratites desideratns Teichert 1948. Teicherticeras lenzi sp. nov. Plate 75, figs. 1-3, 10, 11; text-fig. 3 Material. One exquisitely preserved specimen in a black, styliolinid mudstone. The specimen is a testate phragmocone partly preserved in crystalline calcite. Dimensions {in mm.) D WW WH UW GSC 16929 510 130 19-2 21-8 32-0 8-5 11-8 13-5 111 3-2 4-1 4-4 TEXT-FIG. 3. Teicherticeras lenzi sp. nov. Cross-section at 52 mm. diameter based on GSC 16929 from 296 feet below the apparent Hume Formation on the Ogilvie River, Northwest Territories. Natural size. Description. Shell form laterally compressed, evolute, whorls advolute at least above 4-5 mm. diameter. Whorl section with flat dorsum and rounded umbilical shoulders : maximum width on the lower flanks, the flanks then converging towards a flat tabular venter. At the sharp ventro-lateral shoulders there appear to have been continuous raised flares. Along the course of the tabular venter are periodic indentations (PI. 75, fig. 2) which are related to grooves between slight ribs on the flanks. Growth-lines form a salient on the dorsal mid-line and another slight salient on the umbilical shoulder. Over the flanks growth lines are gently rursiradiate with a shallow sinus on the outer-mid-flanks and a salient on the outer flanks. The growth-lines pass back to a deep linguiform sinus on the venter. Before about 15 mm. diameter, the shell is ornamented by periodic strong rursiradiate ribs with a sharp crest commonly developed (PL 75, fig. 11). A septum seen at 50 mm. diameter shows a broad lateral sinus passing to a saddle centred on the dorsal mid-line which may have a slight lobe at its crest. Ventrally there is probably a small lobe, but this was not seen. Remarks. There is no evidence at any stage that there is an impressed area. This factor, and the paired ventral flares, would suggest that generic reference should be to Gyroceratites (Lamelloceras). But the growth- line pattern is utterly different from any known in Gyroceratites and is distinctly primitive. Similar patterns are shown in the early whorls of Teicherticeras lardeuxi Erben (1960, p. 68, pi. 5, figs. 5-7). The poor evidence of ornament in T. desideratum (Teichert 1948, p. 65, pi. 16, fig. 1) at larger diameters is similar to that of the Canadian specimen. The whorls of an undescribed species from Nevada are significantly more serpenticonic and are distinctly ribbed and have no marked lateral sinus. The new species is referred to Teicherticeras, but, as has been remarked elsewhere (House 1962, p. 252), the Teicher- ticeras group is a variable group which is still only imperfectly known. The horizon of the specimen is probably Emsian rather than Eifelian. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 509 Horizon and locality. GSC 16929 (CSC W6-58) from 296 feet below the base of the Hume equivalent on the Ogilvie River, Yukon, 65° 23' N., 138° 31' W. Family agoniatitidae Haug 1898 Genus agoniatites Meek 1877 Type species by the subsequent designation o/Foord and Crick 1 897 ; Goniatites expansus Vanuxem 1 842. Agoniatites cf. vainixemi (Flail) Plate 72, figs. 3, 4 Material. Two specimens preserved partly in a tentaculitid-rich limestone and partly as internal moulds of crystalline calcite. Dimensions (in mm.) D WW WH Wh UW GSC 16922 47-5 200 c. 20 0 c. 17-5 18-3 300 15-8 - - 100 GSC 16923 - - 23-8 190 - Description. Shell form compressed, sub-evolute, with whorl section of early volutions depressed, reaching WW = WH at about 47 mm. diameter and becoming slightly compressed thereafter. Whorl section with maximum width close to the steep umbilical wall and rounded shoulder. Flanks converge convexly towards a ventro-lateral furrow which is weakly double both on the shell and internal mould. The venter is almost flat but slightly convex. Suture shows a small V-shaped ventral lobe, and an immediately adjacent ventro- lateral saddle. The broadly rounded lateral lobe extends from the ventro-lateral saddle to an umbilical saddle which centres on the umbilical seam. Dorsal suture not seen. Remarks. Agoniatitids are notoriously variable, but the best of these specimens (GSC 19622) is almost identical with comparable inner whorls of type material of A. vanuxemi from the Cherry Valley Limestone of New York (compare with Miller 1938, pi. 10, fig. 3), but since no outer whorls are available the Francis Creek specimens cannot with certainty be assigned to that species. Horizon and locality. Both specimens, GSC 16922-3 (TOC H6053), were collected by E. W. Best and D. Barss from an unnamed dark shale unit (the Lower Fort Creek and Hare Indian of authors) 10-20 feet above the top of the Hume Formation in Francis Creek, Northwest Territories, about 65° 14' 20" N., 126° 23' 40" W. Agoniatites alf. fulguralis (Whidborne) Plate 70, fig. 1 ; Plate 73, fig. 5 1962 Agoniatites sp. House, p. 255. Dimensions (in mm.) D WH UW GSC 16925 75-0 c. 39 0 c. 12-0 GSC 16924 45-0 c. 23-5 c. 5-0 Description. Shell form probably compressed, involute with small open umbilicus. Ornament weak, consisting of growth-lines which form a prominent projecting ventro- lateral salient, a shallow sinus on the outer flanks and a weak salient on the lower C 1456 L 1 510 PALAEONTOLOGY, VOLUME 6 flanks. In the smaller specimen growth lines are marked by raised striae about 1-2 mm. apart on the mid flanks at about 33 mm. diameter. These appear to become weaker and closer set at larger diameters. Suture not seen. Remarks. Agoniatitids with subdued ornament and small umbilicus at diameters com- parable with these specimens are not numerous. Described species of the group include: A. fulgiiralis phillipsi Wedekind (1917, p. 112, pi. 16, fig. 1), A. roemeri (Holzapfel 1882, p. 234, pi. (45) 2, figs. 1, \a), A. verjwrhenanus (Maurer 1876, p. 821, pi. 1, figs, a-c), A. kayseri Wedekind (1917, p. 110, pi. 15, fig. 10), A. clariondi Fetter (1959, p. 77, pl.l, figs. 9, 9a), and A. occultus (Barrande 1865, p. 36, pi. 9, figs. 14-17). The presence of a prominent ventro-lateral furrow in the Canadian specimens is shown in A. vernorhenanus, A. occultus, and A. fulguralis fulguralis (Whidborne). Of these, comparison with the last is preferable, but the specimens show less sinuous growth-lines and a narrower umbilicus than the types of Whidborne (1890, p. 59, pi. 5, figs. 4, Aa). Horizon and localities. One specimen collected by A. E. H. Pedder, GSC 16924 (TOC D139) from an unnamed dark shale (Lower Fort Creek and Hare Indian of authors) about 20 feet above the top of the Hume Formation on Bosworth Creek, Northwest Territories, about 65° 20' N., 126° 51' W. The other specimen, GSC 16925 is from the Fort Creek Shale at an unknown horizon on the Ogilvie River, Yukon, at about 65° 20' N., 138° 44' W. Agoniatites cf. holzapfeli Wedekind Plate 70, fig. 8 Material. A single, incomplete, partly crushed specimen preserved in tough fissile dark-grey shale. Remarks. The specimen is a distinctively evolute agoniatitid. Although poorly preserved, the large protoconch shows clearly. This specimen is comparable to A. holzapfeli Wedekind (1917, p. 113, pi. 15, figs. 13, 14, text-fig. 21/) rather than A. costulatus (d’Archiac and de Verneuil 1842, p. 341, pi. 26, figs. 3, 3a, 3b), for the lateral growth- line sinus is low on the flanks and the ribbing becomes quite subdued by 22 mm. diameter. Horizon and locality. The specimen, GSC 16926 (CSC MC21-59), is from the Fort Creek Formation on the Ogilvie River, Yukon, about 65° 20' N., 138° 44' W. EXPLANATION OF PLATE 70 Fig. 1. Agoniatites di^. fulguralis (Whidborne) from the Fort Creek Formation on the Ogilvie River, Yukon. GSC 16925, X 1. Figs. 2, 3, 7. Maenioceras sp. from the Fort Creek Formation on the Carnwarth River. Respectively GSC 16941-2, x2; GSC 16948, X 1. Fig. 4. Agoniatites sp. from the Fort Creek Formation at Hungry Lake, Yukon. GSC 16927, X 2. Figs. 6, 10. Ponticeras sp. from an unnamed shale unit 10-40 feet above the top of the Beavertail Formation on Carcajou Ridge, Northwest Territories. 6, GSC 16947, X 2. 10, GSC 16945, showing sutures weathered out. Figs. 5, 9. Ponticeras cf. tschernyschewi (Holzapfel) from the same horizon and locality. GSC 16943 and GSC 16944 respectively, both X 1. Fig. 8. Agoniatites cf. holzapfeli Wedekind from the Fort Creek Formation on the Ogilvie River, Yukon. GSC 16926, X 2. Palaeontology , Vol. 6 PLATE 70 HOUSE and PEDDER, Devonian goniatites M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 511 Agonialites sp. Plate 70, fig. 4 Material. A single, incomplete specimen preserved in dark-grey shale. Remarks. The marked ribbing shown on the illustration of this specimen is due to crushing of the camerae of the phragmocone. The interest of the specimen lies in the unusually weak projection of the ventro-lateral salient and the apparent absence of a ventro-lateral furrow. This may represent a new species. Horizon and locality. The specimen, GSC 16927 (CSC WH23H-58), is from an unknown horizon in the Fort Creek Formation at Hungry Lake, Yukon, 65° 36' 30" N., 136° 24' W. Genus sellagoniatites nov. Type species. Goniatites discoides Waldschmidt 1885 (the holotype is refigured here: PI. 73, figs. 6, 7). Diagnosis. Involute, umbilicate, laterally compressed agoniatitids with a suture similar to Agonialites but with a mid-dorsal lobe and a shallow lobe on the ventral face of the ventral lobe. Growth-lines biconvex. Distribution. The type species occurs in the upper Givetian discoides Kalk in Germany (Schmidt 1958, p. 325) and is known in the upper Givetian of North Africa (Fetter 1959, p. 87). The genus may well be represented in New York State by S. unilobatiis (Hall) and is represented in the Northwest Territories by S.jacksoni sp. nov. Sellagoniatites jacksoni sp. nov. Plate 71, figs. 1-4; text-fig. 4a, b Material. One specimen only, the holotype, preserved as an incomplete internal mould in a grey calcareous mudstone. Dimensions (in nun.) D WW WH UW GSC 16928 156 c. 50 78 30-5 Description. Shell form involute, laterally compressed. Whorl section compressed, with maximum width two-fifths across the flanks from the umbilicus at the maximum diameter. The outer flanks slope flatly towards a well-rounded venter; the inner flanks slope convexly towards the umbilicus. Along the line of maximum whorl width are periodic nodes, and there are five of these, with weak traces of a sixth, in the last half whorl. The nodes become more prominent on the body chamber. Sutures (text-fig. 4a) show a V-shaped ventral lobe with sigmoidal sides, the lobe being extended apically along the course of the siphuncle. The ventro-lateral saddle is acutely rounded and the lateral lobe is very broad and semicircular. The dorsal suture has not been observed. Growth-lines (text-fig. 4b) show a deep linguiform sinus on the venter passing to a broad salient on the outer flanks. Laterally the growth-lines are concavely prosiradiate with a weak salient on the inner flanks. At the extreme apertural end of the body chamber is a very shallow constriction parallel with the course of the growth-lines. Remarks. This species is larger than the two other members of Sellagoniatites so far 512 PALAEONTOLOGY, VOLUME 6 recognized, 5'. imilobatiis (Hall 1879, pi. 71, figs. 15, 16; pi. 74, fig. 5) and the type species S. discoides. Illustrations of the latter are given here (PL 73, figs. 4, 6-8): one is the holotype and the other is a topotype in the Gottingen Museum, and a dorsal view of the impressed area (PI. 73, fig. 8) shows the mid-dorsal saddle which is the critical TEXT-FIG. 4. A, B, Sellagoiiiatites jacksoni gen. et sp. nov. a. Suture at c. 120 mm. diameter, b, Growth- line at 145 mm. diameter: the cross marks the position of a node. Both based on the holotype, GSC 16928, from the Fort Creek Formation on the Carnwarth River, Northwest Territories, c, D, Agoniatites vamixenii (Hall), c, Suture at 127 mm. diameter based on AMNH 4416/1 : i (figured Hall 1879, pi. 67, fig. 1). d. Growth-line at c. 91 mm. diameter based on AMNH 4416/1 : iii (figured Hall 1879, pi. 66, figs. 1, 2). Both from the Cherry Valley Limestone at Manlius, N.Y. E, F, Sellagoiiiatites unilobatus (Hall). E, Suture at 33 mm. diameter, f. Growth-line at 27 mm. diameter. Both based on the holotype, NYSM 3547, labelled ‘Hamilton Beds, Norton’s Landing, Cayuga Lake, N.Y.’ G, Sellagoiiiatites discoides (Waldschmidt). Suture at a whorl height of 31 mm. of a topotype in the GPIG from the discoides Schichten at Ense near Wildungen, Germany. All natural size. feature of the new genus. S. jacksoni differs from the type species in that the venter is flatly rounded rather than sub-acute. At comparable diameters, S. discoides appears to be more compressed. S. unilobatus, known only from the holotype from the top of the Hamilton Formation at Norton’s Landing, Cayuga Lake, New York, is too small for adequate comparison with either species since the maximum diameter of the holotype is only 36 mm. However, an examination of that specimen (NYSM 3547) suggests that the growth-lines have a deeper lateral sinus and more prominent salient on the inner M. R. HOUSE AND A. E. H. PEDDER; DEVONIAN GONIATITES 513 flanks than are shown by S. jacksoni at much greater diameters (text-fig. 4b, f). All species included within the new genus show a characteristic concavity at the sides of the ventral lobe (text-fig. 4a, e, g). Horizon and locality. The holotype, GSC 16928 (PAPC GS-4-59B), is from talus at the foot of an outcrop of the lower part of the Fort Creek Formation on the Carnwarth River, Northwest Territories, 7| miles downstream(direct)from the confluence with the Iroquois River, 68° IF 30" N., 129° 20' 24" W. Family anarcestidae Steinmann 1890 Genus anarcestes Mojsisovics 1882 Type species by the subsequent designation o/ Foord and Crick 1897: Goniatites plebeius Barrande 1865. Anarcestes (Latanarcestes) cf. praecursor Freeh Plate 74, fig. 3 1865 Goniatites plebeius (pars) Barrande, p. 37, pi. 7, figs. 3-9. 1897 Anarcestes praecursor Freeh, p. 169. 1913 Anarcestes praecursor Freeh, p. 14. 1959 Anarcestes praecursor Chlupac, p. 482, 492, 505. Material. One crushed specimen preserved in grey calcareous mudstone. Description. Shell form apparently sub-globular, involute with small open umbilicus. Growth-lines weakly shown, with a salient on the lower flanks and a slight sinus on the mid flanks which passes towards a ventro-lateral salient. Suture with a broad and very shallow lateral lobe and a saddle approximately centred on the seam. Suture not seen on the ventral and dorsal sides. Remarks. The weak growth-lines and subglobular form preclude this specimen from Agoniatites and indicate an anarcestid, and the shallow and wide lateral lobe shows that it should be referred to Anarcestes {Latanarcestes), among the described species of which only A. (L.) praecursor has so narrow an umbilicus. Indeed, the Canadian specimen has a smaller umbilicus than that shown on any of Barrande’s figures, but this is in part at least emphasized by crushing. Chlupac has shown that this species occurs in the Trebotov Limestone of Czecho- slovakia, and in the Suchomasty Marble associated with a lower Eifelian fauna. He has also demonstrated that Prantl’s statement (1954) that the species is Siegenian is false. Horizon and locality. The specimen, GSC 16930 (TOC G5170), was collected by A. E. H. Pedder from a horizon 469 feet below the top of the Funeral Formation and 1,525 feet below the top of the Nahanni Formation in the northern Funeral Range, Northwest Territories, about 61° 4F 30" N., 125° 05' W. Genus cabrieroceras Bogoslovski 1958 Type species by original designation : Goniatites rouvillei von Koenen 1886. Cabrieroceras karpinskyi (Holzapfel) Plate 72, figs. 1, 2, 7, 9-12; text-fig. 5 1895 Anarcestes karpinskyi Holzapfel, p. 77, pi. 3, figs. 15-20. 1917 Anarcestes Rouvillei (pars) Wedekind, p. 109. 514 PALAEONTOLOGY, VOLUME 6 1933 Werneroceras rouvillei (pars) Schindewolf, p. 98. 1958 Werneroceras karpinskyi Bogoslovski, p. 74. 1959 Werneroceras (?) karpinskyi Fetter, p. 105. Material Eight specimens preserved as internal moulds in crystalline calcite. Dimensions {in mm.) D WW WH UW GSC 16931 48-2 c. 22-3 15-9 20-3 GSC 16932 37-4 c. 190 c. 10-5 18-2 24-2 15-1 6-8 11-8 GSC 16933 24-3 15-7 8-4 11-5 GSC 16934 27-5 170 c. 9-5 c. 110 Description. Protoconch 1-4 mm. wide. Shell form rotund in the inner whorls with an open, expanding umbilicus; becoming compressed in the outer whorls. Whorl section TEXT-FIG 5. Cabrieroceras karpinskyi (Holzapfel). A, Cross-section based on GSC 16932. b, Tangential cross-section of GSC 16935. c, Cross-section based on GSC 16931. All from 0-10 feet above the top of the Hume Formation on Francis Creek, Northwest Territories. All natural size. extremely depressed up to about 20 mm. diameter with the whorl height about one-third of the whorl width; the venter is thus extremely broad and flat and the flanks limited and well rounded. Above 20 mm. diameter the section gradually becomes more reni- form, and the relative whorl height increases; thus the umbilical angle, which in early whorls is about 86°, becomes less in the outer whorls (text-fig. 5). The diameter at which this change takes place is rather variable as the figures given above show, but the quality of the material is not sufficient to analyse the variation critically. The suture has not been seen. Growth-fines are well shown and on the inner whorls occur as finely spaced striae backwardly directed from the umbilical seam. At 20 mm. diameter growth-lines on the broad venter form a shallow sinus distally and a rounded salient before passing back into a U-shaped sinus, about as broad as deep, on the mid-ventral line. EXPLANATION OF PLATE 71 Figs. 1-4. Sellagoniatites jacksoni gen. et sp. nov. Holotype, GSC 16928, from talus at the foot of an outcrop of the lower part of the Fort Creek Formation on the Carnwarth River, Northwest Territories. 1, 2, 4, xO-94; 3, x2. Palaeontology, Vol. 6 PLATE 71 HOUSE and PEDDER, Devonian goniatites e M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 515 Remarks. The only other species of Cabrieroceras known from North America is C. plebeiforme (Hall) from the Werneroceras Bed below the Cherry Valley Limestone in eastern New York State (Rickard 1952, House 1962), unless Werneroceras staujferi Sweet and Miller (1956) belongs here. Both C. plebeiforme and its probable European equivalent, C. crispiforme, show a regular increase in the width of the umbilicus and an almost constant whorl form throughout ontogeny without such prominent modifica- tions in the outer whorls as Holzapfel described in the type material of C. karpinskyi. These modifications were apparently overlooked by Wedekind and Schindewolf (1933, p. 94), who have regarded C. crispiforme {= C. ronvillei) and C. karpinskyi as synonyms. Stratigraphically C. crispiforme is characteristic of the lowest Givetian whilst C. karpin- skyi occurs in the middle and upper Givetian. Illustrations of one of Holzapfel’s original types are given here for comparison with the Canadian specimens (PI. 72, figs. 5, 6, 8). Horizon and locality. Eight specimens collected by E. O’Bertos, GSC 16931-8 (TOC H6054), from an unnamed dark shale unit (Lower Fort Creek and Hare Indian of authors) 0-10 feet above the top of the Hume Formation. Specimens of Cabrieroceras sp. also occur slightly higher in the succession. The locality is on Francis Creek, Northwest Territories, 65° 14' 20" N., 126° 23' 40" W. Family maenioceratidae Ruzhencev 1958 Genus maenioceras Schindewolf 1933 Type species by original designation : Goniatites terebratus G. and F. Sandberger 1850. Maenioceras sp. Plate 70, figs. 2, 3; Plate 72, fig. 16 Material. Two poorly preserved specimens in tough calcareous shale. Remarks. The specimens are too poorly preserved to warrant detailed description. The involute form, evidence of ventro-lateral furrows, angular lateral lobes (PI. 70a, fig. 3), and distinctive wrinkle layer (PI. 72, fig. 16; see House 1962, p. 269), all show that re- ference should be to the Maenioceras terebratiim group. The occurrence with these of specimens determinable as Wedekindella aff. brilonense (Kayser), and Agoniatites sp. confirms that the horizon is upper Givetian. Horizon and locality. Both specimens are from an unknown horizon in the Fort Creek Shale along the Carnwarth River, Northwest Territories, about 67° 40' N, 128° 22' W. The specimens are GSC 16941-2 (CSC RAS62-59). Family pinacitidae Schindewolf 1933 Genus wedekindella Schindewolf 1933 Type species by original designation: Goniatites retrorsus var. brilonense Kayser 1872 (the holotype is refigured here: Pl. 73, fig. 2). Several specimens of this genus are now known from several localities in the North- west Territories. A description has already been given of two specimens of W. brilonense from along the Carnwarth River (House 1962). New material includes a specimen of Wedekindella sp. (GSC 16959) from the Fort Creek Shale on the Anderson River (PI. 73, fig. 1) which appears to represent a non-constricted form. From the Fort Creek Shale is a specimen referred to W. aff. brilonense (PI. 73, fig. 3) which differs from W. brilonense 516 PALAEONTOLOGY, VOLUME 6 in showing evidence of a weak but continuous ventro-lateral furrow. None of the specimens referred to Wedekindella so far show the form of the suture. Horizons and localities. GSC 16939 (CSC RAS83-59) from an unknown horizon in the Fort Creek Formation, 45 miles NNE of the Thunder River Mouth, Northwest Territories, about 68° 00' N., 130° 10' W. GSC 16940 (CSC RAS62-59), from an unknown horizon in the Fort Creek Formation on the Carnwarth River, Northwest Territories, about 67° 41' N., 128° 22' W. AU 45447, from 130 feet above the base of the Fort Creek Formation on the Carnwarth River, Northwest Territories. Family gephuroceratidae Freeh 1902 Genus ponticeras Matern 1929 Type species by original designation: Ammonites aequabilis Beyrich 1837. Ponticeras cf. tschernyschewi (Holzapfel) Plate 70, figs. 5, 9, 10 Material. Seven specimens collected by A. C. Lenz and preserved crushed in grey fissile siltstone. Dimensions {in mm.) D WH UW GSC 16944 62-5 20-4 26-4 440 17-8 140 Description. Shell form evolute, apparently laterally compressed, with wide open umbiheus. Whorl section not discernible, but fractures suggest maximum whorl width was on the lower flanks and there was a ventro-lateral furrow with a slightly raised ridge on the umbilical side. Several specimens show a slight contraction of the whorl height in the outer part of the body chamber. Sutures show evidence of a trifid ventral lobe, a large lateral saddle with a crest on the mid-flanks and a rounded sub-umbilical lobe. Weathered sutures (PI. 70, fig. 10) falsely resemble those of Protornoceras. Growth-lines not seen. Remarks. Evolute ponticeratids of this type occur in the Lower Frasnian in many parts of the world. The specimens resemble P. perlatum (Hall) from the Geneseo Shale and Sherburne Sandstone of New York State, but the whorls of that species appear to expand more rapidly and show irregularities along the growth-lines simulating ribs (House 1962, pi. 45, fig. 10). Other closely related species are: P. tschernyschewi (Holzapfel 1899, p. 30, pi. 4, figs. 1-6; pi. 6, fig. 8), P. sandbergeri (Wedekind 1913), P. sahlgrundensis Matern 1931, P. pernai (Wedekind 1917). All these are Lower Frasnian forms apart from P. sandbergeri, which Wedekind records in the Middle Frasnian. Other known species of Ponticeras have too rotund a whorl section to be comparable. In the anticipa- tion that solid specimens will eventually appear from Canada some dimensions (in mm.) of the types of these European species will be given, based on new measurements, to facilitate future comparisons. P. tschernyschewi (probable syntypes in Gottingen Museum) P. sandbergeri, lectotype syntype P. sahlgrundense, syntype P. pernai, holotype D WW WH UW 450 17-0 7-8 15-5 - 19-5 9-9 - e. 17-0 4-7 5-5 c. 8 0 130 3-5 4-5 5-5 25-3 7-8 10-4 7-1 38-0 c. 120 14-6 11-5 c. 22-0 8-2 9-5 7-2 M. R. HOUSE AND A. E. H. REDDER: DEVONIAN GONIATITES 517 Of these species only P. tschernyschewi has been recorded at diameters comparable with the Canadian specimens, and agreement is close in other factors ; for example, some of Holzapfel’s figures (e.g. pi. 4, figs. 5, 6) show the same contraction of the adult whorl height, and others (e.g. pi. 4, fig. 3) show the same ventro-lateral groove and ridge on the internal mould. Horizon and locality. GSC 16943-6 (CSC L-8) collected by A. C. Lenz from an unnamed dark shale 10-40 feet above the top of the Beavertail Formation (Upper Fort Creek of authors) on Carcajou Ridge, Northwest Territories, about 65° 36' N., 128° 30' W. Genus probeloceras Clarke 1898 Type species by nionotypy: Goniatites hit her i Clarke 1885. The genus Probeloceras forms a very distinct group of lower Frasnian goniatites. In addition to the type species, the following species belong hQrt: P . forcipiferiim (G. and F. Sandberger), P. genundewa (Clarke), P. planorbis (G. and F. Sandberger 1851, pi. 9, fig. 3a here designated lectotype), P. (?) orientale Bogoslovski and P. applanata (Wede- kind). The group has tabular or grooved venters; commonly the grooved form is seen only in testate specimens whilst internal moulds often show no ventral groove. Very closely related to this group, and probably derived from it, are those species generally included in Manticoceras which have a similar shell form. These include M. neapolita- num (Clarke), M. holzapfeli (Clarke), and M. alveolatiim (Glenister). The only distinc- tion between the two groups is that the latter group have an umbilical lobe in addition to the lateral lobe. If a generic name is ever required for this group, then the somewhat inappropriate name Acanthoclymenia is available, the type species of which is M. nea- politamtm (see House 1961). Probeloceras sp. Plate 76, figs. 1-4, 7; text-fig. 6 a-c Material. Three small pyritized specimens, and three specimens crushed fiat in black shale which may belong here. Dimensions (in mm.) D WW WH GW IOC 258fl 17-2 - c. 5-0 c. 4-9 IOC 2586 8-8 2-2 3-3 3-3 5-9 F6 2-3 2-0 GSC 16972 4-5 1-3 1-75 1-4 Description. Shell form laterally compressed, evolute in innermost whorls, sub-evolute in outer whorls. Whorl section with slight impressed area from earliest whorls, with well- rounded flanks, weak ventro-lateral furrows, and sub-tabular venter by 6 mm. diameter becoming sharply tabular in outer whorls. Shell development shown in text-fig. 6a. At no stage seen is the internal mould grooved. Sutures are illustrated in text-fig. 6b. Growth-lines are not well seen, but are apparently biconvex. Remarks. The two species of Probeloceras from the cordatum Zone, or its equivalents, P. lutheri and P. genundewa, both have more pronounced mid-ventral saddles at com- parable diameters. Of those from the hmulicosta Zone, the holotype of P. forcipiferum, which reaches 18 mm. diameter, has a higher mid-ventral saddle and narrower lateral 518 PALAEONTOLOGY, VOLUME 6 saddle than the Canadian specimens at 10 mm. diameter, and the flanks of both P. forcipiferum and P. planorbis are considerably flatter. P. applanata and P. (?) orientale are apparently more involute and rotund, but there is a Pwbeloceras among specimens figured by Hall (1879, pi. 70, fig. 8) which has a suture at c. 24 mm. diameter (text-fig. 6d) with a relatively broader lateral saddle, but the flanks are moderately rounded. TEXT-FIG. 6. Probeloceras sp. from Canada and New York State. A, b, Whorl section at 8-8 mm. diameter and suture at 7 mm. diameter based on IOC 2586 from the Imperial Paddle River Borehole between 7,913 and 7,932 feet. Both x6. c. Suture at 9-5 mm. diameter based on IOC 258n from the same locality and horizon, X 6. d. Suture at 2-24 mm. diameter based on AMNH 5833/1:2 (figured Hall 1879, pi. 70, fig. 8) from the Geneseo Shale near Geneseo, N.Y. Reversed for comparison, x5. It is unfortunate that these various species are not better described. As a result the Cana- dian specimens cannot be critically compared with them at similar diameters. The specimens seem closest to those species from the hmulicosta Zone rather than the cordatum Zone. EXPLANATION OF PLATE 72 Figs. 1, 2, 5-12. Cabrieroceras karpinskyi (Holzapfel). 1, 2, GSC 16931, X 1. 7, 9-12, GSC 16933; 9-12, X L5; 7, x4. Both specimens from an unnamed shale unit 0-10 feet above the top of the Hume Formation on Francis Creek, Northwest Territories. 5, 6, 8, A cotype figured by Holzapfel (1895, pi. 3, fig. 18) in the Gottingen Museum from the Oderhauser Kalk at Ense, Germany. 5, 6, x2; 8, x4. Figs. 3, 4. Agoniatites cf. vanuxemi (Hall) from the same locality, 10-20 feet above the top of the Hume Formation. GSC 16922, xl. Fig. 13, 14. Agoniatites sp. from the Fort Creek Shale along the Carnwarth River, Northwest Territories. 13, GSC 16970, x4. 14, GSC 16969, x 1. Fig. 15. Tornoceras (T.) cf. westfalicum Holzapfel, from an unnamed shale unit 10-40 feet above the top of the Beavertail Formation on Carcajou Ridge, Northwest Territories. GSC 16949, X 1. Fig. 16. Maenioceras sp. Close-up photograph showing the wrinkle layer of a specimen figured on PI. 70, fig. 3. GSC 16942, X 10. Palaeontology, Vol. 6 PLATE 72 HOUSE and PEDDER, Devonian goniatites ft M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 519 Horizon and localities. IOC 25a, b from Imperial Paddle River 5-17-56-8W.5M well, core 85, at depths between 7,913 and 7,932 feet. GSC 16972 (loc. 16545) from the same borehole and core at 7,922 feet depth. Also placed here, but with little assurance, are three crushed specimens, GSC loc. 26784 from Imperial Amerada Willesden Green 3-16-43-5W.5M well, Alberta, between 9,990 and 9,995 feet. The horizon is the lower part of the Duvernay Formation {sensu lato). Genus manticoceras Hyatt 1884 Type species by original designation : Goniatites simulator Hall 1874. Determination of specimens of Manticoceras presents particular difficulty at present. Over sixty species and subspecies are referred to the genus, and almost all the types of these have never adequately been described, particularly with regard to the ontogeny of shell and suture form. Further, many represent large, usually limestone, specimens whilst others are based on small or minute pyritic specimens. Until a detailed study of the group has been accomplished and revision of the type material made, a large number of the available specific names are virtually useless, and the determination of material very difficult. Manticoceras occurs not uncommonly in the Frasnian rocks of western Canada, but most specimens are too poorly preserved either to warrant detailed description or to enable refined determinations to be made. M. oxy has been reported on the basis of speci- mens whose oxyconic form results purely from crushing. Only M. eordiforme and M. septentrionale (here shown to be synonyms) can be critically described. Other specimens shed little light on correlation with the successions known in Europe or New York. But most specimens appear to belong to the eordatiim Zone assemblages, rather than to other Frasnian levels. Manticoceras eordiforme Miller Plate 76, figs. 5, 6, 8-1 1 ; text-figs. 7a-d, 9f, h 1938 Manticoceras eordiforme Miller, p. 82, pi. 21, figs. 2-4. 1938 Manticoceras septentrionale Miller, p. 102, pi. 23, figs. 1, 2, text-fig. 19. 1945 Manticoceras intumescens Hume and Link, p. 39. 1954 Manticoceras intumescens Hume, p. 46. Examination and sectioning of the holotypes of M. eordiforme and M. septentrionale has led to the conclusion that they are conspecific. The name M. eordiforme is therefore used as it has page priority, but the name is somewhat innappropriate as the cordate section is emphasized by crushing of the holotype. Material. The holotypes of both species and six other specimens, some indifferently preserved, which are thought to be probably conspecific. The detailed description below is based on the holotypes of M. eordiforme and M. septentrionale alone. Dimensions {in mm.) D WH WIV Wh UW GSC 5139 c. 200 c. 93 c. 74 c. 70 37-5 c. 130 c. 72 c. 60 c. 50 21-4 c. 80 c. 40 c. 32 - 13-2 GSC 2393 720 31-5 26-2 - 12-6 26-5 11-5 110 8-6 5-5 10-5 4-9 5-1 3-6 2-9 520 PALAEONTOLOGY, VOLUME 6 Description. Shell form in the earliest whorls, before 5 mm. diameter, evolute and serpenticonic, and at subsequent diameters evolute, so that at 200 mm. diameter the form is sub-involute with UW/D = 19 per cent. Whorl section in the early stages is depressed with a very slight impressed area but by 15 mm. diameter WH = WW and the TEXT-FIG. 7. Manticoceras cordiforme Miller from the Carcajou Mountain Sandstone at Norman Wells, 4 5 miles north-west of Fort Norman, Northwest Territories, a. Actual cross-section of the holotype of M. ‘septeiitrioiiale' at 200 mm. diameter, b. Reconstructed cross-section. Both based on GSC5139. Both X 0 5. c. Actual cross-section of the holotype of M. cordiforme at 60 mm. diameter based on GSC 2393. xO-5. D, Graph showing the relation between diameter and whorl width (full line) and umbilical width (pecked line) of the holotypes of M. cordiforme and M. ‘septentrionale'. EXPLANATION OF PLATE 73 Fig. 1. Wedekindella sp. from the Fort Creek Formation, 45 miles NNE of the Thunder River Mouth, Northwest Territories. GSC 16939, X2. Fig. 2. Wedekindella brilonense (Kayser). Holotype of the species from the upper Givetian at Brilon, Germany, in the GPIG. X 2. Fig. 3. Wedekindella aff. brilonense (Kayser) from the Fort Creek Formation on the Carnwarth River, Northwest Territories. GSC 16940, x 2. Figs. 4, 6, 7, 8. Sellagoniatites discoides (Waldschmidt). 6, 7, Holotype from the discoides Schichten at Ense, Germany, X 1. 4, 8, A topotype in the GPIG. 4, X 1 ; 8, X 2-5. Fig. 5. Agoniatites aff. /«/g«ra/A (Whidborne) from an unnamed shale unit about 20 feet above the top of the Hume Formation on Bosworth Creek, Northwest Territories. GSC 16924, X 1. Palaeontology, Vol. 6 PLATE 73 HOUSE and PEDDER, Devonian goniatites M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 521 impressed depth has increased considerably. Above 15 mm. diameter or thereabouts the whorl form becomes progressively compressed. From about 5 mm. diameter to about 70 mm. diameter the whorl section shows a robust form with the flanks converging convexly towards the venter, but at the maximum diameter seen (text-fig. 7b) the flanks become more parallel sided. Suture first seen at 38 mm. diameter, where the subumbilical lobe is sub-acute; subsequently this lobe becomes sharply acute. At 57 mm. diameter all the sutural elements are seen and the mid-ventral saddle, 1 1 mm. high, shows only a slight lobe above the siphuncle. The broadly rounded lateral saddle is slightly asymmetric (text-fig. 9f, h) but the relative proportions of this and other elements appear to change little at subsequent diameters. Growth-lines at 58 mm. diameter (PI. 76, fig. 6) are almost convex but show a very slight sinus on the mid flanks. This form is also seen at 220 mm. diameter (PI. 76, fig. 10) where the growth-lines have a slightly rursiradiate course across the flanks as a whole. Remarks. The most distinctive feature of M. cordiforme is the growth-line pattern, which is intermediate between that of typical Manticoceras and Crickites. Several species with this type of growth-line have been recorded. M. sinuosum (Hall, lectotype selected by House 1962, p. 259 as Hall 1843, p. 243, fig. 106 : 8) shows this feature as does one of the syntypes of Goniatites patersoui Hall (1879, pi. 72, fig. 4 = AMNH 6242/1) from the ‘Chemung group. Pine Valley, town of Catherine, Schuyler county, N.Y. ’ These two species have been regarded as synonyms by Clarke (1898) and Miller (1938). But M. cordiforme differs from specimens referred to M. sinuosum by Miller in showing a rather more robust whorl section. Among European species, C. expectatum Wedekind, C. saJdgrundensis Matern, and C. scheldensis Matern show growth-lines which are nearly rectilinear across the flanks, but the holotype of C. scheldensis (SMF 328c) shows a weak lateral sinus which is lost by 25 mm. diameter. If a generic name for manticoceratids similar to M. cordiforme is ever required, then Gephuroceras (Hyatt 1884, p. 316) is available. With the present absence of detailed work on Manticoceras and its allies it is difficult to interpret the stratigraphic significance of the Canadian specimens. Wedekind (1917, p. 130), who did more work on this group than any of those who have since criticized him, noted that Crickites is rare in the upper cordatum Zone (of present usage) and common in the hohapfeli Zone. The New York species comparable, in growth-line form, to M. cordiforme occur in the Cashaqua Shale, which, on other grounds, is thought to represent the upper cordatum Zone (House 1962, p. 259). Thus the evidence suggests that the Carcajou Mountains Sandstone and its equivalents are upper Frasnian, but whether upper cordatum Zone or hohapfeli Zone age cannot at present be determined. Horizon and locality. GSC 5139 (the holotype), GSC 2393, and WM 51873, all from the Carcajou Mountain Sandstone at Norman Wells (Oil Creek), 45 miles north-west of Fort Norman, Northwest Territories. Manticoceras cf. sinuosum (Hall) Plate 74, figs. 6, 7; text-fig. 8a-d Material. Specimens from the Perdrix Formation and Mount Hawk Formation, mostly rather poorly preserved in dark-grey argillaceous limestone. PALAEONTOLOGY, VOLUME 6 522 Dimensions {in mm.) D WW WH UW Perdrix Formation GSC 16973 33 c. 11 c. 17 5-5 GSC 16974 c. 32-5 c. 12-5 c. 16-5 c. 4-6 Mount Hawk Formation GSC 16976 - 36-5 72 - SOC T6207 51-5 16-6 24 11-5 TEXT-FIG. 8. Manticoceras cf. sumosum (Hall) from the Mount Hawk Formation, a, Cross-section based on GSC 16975 from Deception Creek, b, Cross-section based on GSC 16976 from the North Saskatchewan River Gap. c, d, Cross-section and suture at 52 mm. diameter based on SOC T6207 from Crescent Creek, Nelson Range. All natural size. Remarks. The specimens from the Perdrix Formation are all poorly preserved, particu- larly specimens above 33 mm. diameter, but all show an acute lateral lobe and are similar in shell form to the Mount Hawk specimens except in showing a rather smaller umbilicus and an associated more rapid expansion in whorl height. The Mount Hawk specimens are illustrated in text-fig. 8. Noticeable is the development of a slender, compressed whorl section at larger diameters which is not wholly accountable by crushing. In this feature they are reminiscent of M. rhyuchostoma Clarke (see Miller 1938, p. 100, text- fig. 18h), but the inner whorls of the New York species are more depressed: no other species is comparable at such large diameters. There is one Perdrix specimen (GSC 16977) which reaches a whorl height of 65 mm. and appears to show an oxyconic form, but this is probably due to crushing. The missing specimens of Bissell (1930, pi. 8, figs. 1-3) from the Mount Hawk Formation may well be M. cf. sinuosum. Horizons and localities. From the Perdrix Formation is GSC 16973,4 (GSC loc. 36946) labelled 'Head Sulphur Creek, Miette area’, collected by E. J. Mountjoy. From the Mount Hawk Formation are GSC 16975 (GSC loc. 24560) collected by D. J. McLaren from unit 11 on Deception Creek, GSC M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 523 16976 (GSC loc. 18101) collected by D. J. McLaren from the north side of a gap on the North Saskat- chewan River, GSC 16978 (TOC RC4252) collected by A. E. H. Pedder from the Saskatchewan Gap in the Brazeau Range about 52° 25' N., 115° 50' W., 280 to 290 feet below the top of the formation, and SOC T6207 from Crescent Creek, Nelson Range, Alberta, about 53° 35' 30" N., 118° 30' W. Manticoceras spp. Plate 74, figs. 8, 9; text-fig. 9a, b, d, e 1938 Manticoceras aff. M. shuiosum Miller, p. 116, pi. 27, figs. 6, 7. 1956 Manticoceras sp. Warren and Stelck, pi. 22, fig. 17. 1956 Manticoceras oxy Warren and Stelck, pi. 24, figs. 1, 3. 1962 Manticoceras sp. House, p. 261, text-fig. 4a. Escarpment Formation. Two specimens are available from this horizon. The first, GSC 4294, was collected by McConnell in 1887 from opposite mile 23 on the original Mackenzie highway {sic): it was described with full synonymy by Miller (loc. cit.). A second specimen, GSC 16979 (GSC loc. 31261) collected by P. Harker, comes from 150 feet below a sandstone marker near the junction of Mills Lake Road and the Hay River: this specimen is figured here (text-fig. 9b, e; the section is very slightly widened by the section being cut just obliquely). Grumbler Group. Several specimens from the Liard Rapids, Northwest Territories, about 61° 26' N., 121° 38' W., belong here, probably from the Redknife Formation. All are somewhat distorted and preserved in siltstone. One, GSC 16964 (TOC NF9, PI. 74, figs. 8, 9), which externally appeared oxyconic, gave evidence on cross-sectioning (text-fig. 9a) of crushing with rounded venters and moderately compressed inner whorls. The specimen figured by Warren and Stelck (loc. cit.) as M. oxy is also from this locality and the specific assignment is probably similarly explained and incorrect. The suture of GSC 16964 is similar to others from this locality (House 1962, p. 261, text-fig. 4a) in showing a gentle rise from the acute lateral lobe in the umbilical seam, a distinctive feature shown also by M. cordiforme. Another specimen, GSC 16980 (GSC loc. 32966) collected by B. R. Pelletier from near the base of a 100-foot section of sandy argillaceous limestone on the left bank at Liard Rapids, also shows this feature. Genus timanites Mojsisovics 1882 Type species by original designation : Goniatites aciitus Keyserling 1846 (pi. 12. fig. 6) = Timanites keyserlingi Miller. Since Keyserling was incorrect in supposing his specimens belonged to Goniatites acutus Munster, no valid name had been applied to these forms until Miller (1936, p. 634) proposed the name T. keyserlingi. Hence it is not possible to agree with Letter’s resurrec- tion of Munster’s specific name. A new specimen from Canada shows that T. occidentalis should be regarded as a synonym of T. keyserlingi. The synonomy includes references to the types and Canadian specimens only. Timanites keyserlingi Miller Plate 75, figs. 6-9; text-fig. 10 1844 Goniatites acutus (non Munster) Keyserling, p. 232, pi. 1, fig. 6. 1846 Goniatites acutus (non Munster) Keyserling, p. 280, pi. 12, figs. 6o, b. 524 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 9. Manticoceras spp. from western Canada and New York State. A, D, Manticoceras sp., cross section and suture at 85 mm. diameter based on GSC 16964 from the Grumbler Group at the Liard Rapids. B, E, Manticoceras sp., cross-section and suture at 54 mm. diameter based on GSC 16979 from the Escarpment Formation near the junction of Mills Lake Road and the Hay River. c, Manticoceras siniiosum (Hall), growth-line of the lectotype at a whorl height of 48-7 mm. based on AMNH 5887/1 from the Cashaqua Shale, New York State. F, H, Manticoceras cordiforme Miller. F, Suture of the holotype at 63 mm. diameter based on GSC 2393; H. Suture at c. 105 mm. diameter based on the holotype of M. ‘‘ septentrionale' \ both from the Carcajou Mountain Sandstone at Norman Wells, 45 miles north-west of Fort Norman, Northwest Territories. All natural size. M. R. HOUSE AND A. E. H. PEDDER; DEVONIAN GONIATITES 525 1882 Timanites acutus Mojsisovics, p. 183. 1932 Mcmticoceras cf. M. oxy Allan et a!., p. 236. 1936 Timanites occidentalis Miller in Miller and Warren, p. 632, figs. 1-3. 1936 Timanites occidentalis Miller and Warren, p. 634, text-figs. 4, 5. 1938 Timanites occidentalis Miller, p. 131, pi. 29, figs. 4, 5. 1951 Manticoceras cf. M. oxy Fox, pp. 827, 828. 1954 Manticoceras cf. M. oxy Fox, p. 115. 1956 Timanites occidentalis Warren and Stelck, pi. 14, figs. 26, 27. 1959 Timanites acntns Fetter, p. 161. Material. Only two Canadian specimens are known, the holotype of T. occidentalis and a very fine specimen recently found by C. R. Stelck. Both are from the same horizon and approximately the same locality. Dimensions (in mm.) D W\V WH UW AU Dv684 (from Miller) 105 20 47 0 AU 38975 95 24-5 58 2-5 32 110 18-5 0 20-4 7-6 13-4 0 B D TEXT-FIG. 10. Timanites keyserlingi Miller from western Canada and Russia. A, B, D, Cross-section and sutures at 78 mm. and 20 mm. diameter based on AU 38975 from the Maligne Formation in Jasper Park, Alberta, a, b, X 1 ; d, x 5. c. Suture diagram from Miller (1938, p. 132, text-fig. 27a) of the holotype of T. ''occidentalis' from the same horizon and locality, based on AU Dv684, X 1. E, Suture diagram from Holzapfel (1899, p. 44, text-fig. 13) of a specimen from the lower Frasnian of the Timan Mountains, U.S.S.R. Description. Shell form compressed and discoidal, with umbilicus probably closed (by the shell) and venter acute in the adult. Early stages (text-fig. 10a) show a well-rounded whorl section but by about 15 mm. diameter the adult type acute venter is developed. Growth-lines, seen on the outer whorls only (PI. 75, fig. 8), show a slight salient on the C 1456 M m 526 PALAEONTOLOGY, VOLUME 6 lower flanks and weak sinus on the midflanks, and the growth-lines pass forward, and become emphasized at the ventro-lateral salient. Over the venter the growth-lines form an acute, V-shaped, sinus. The wrinkle layer is seen as a dorsal structure on top of the previous whorl at 30-35 mm. diameter (PI. 75, fig. 5; the same area is seen on fig. 6). The wrinkle layer striae pass slightly backwards from the umbilicus and continue in an irregular rectilinear course towards the venter, frequently bifurcating. By 20 mm. diameter (text-fig. IOd) the adult sutural elements have formed and the dorsal suture has a mid-dorsal lobe and a sub-umbilical lobe. The adult suture is illus- trated on an accompanying diagram (text-fig. 10b, c). Remarks. The new specimen, from the same formation as the holotype of T. occi- dentalis, shows even closer similarity to Russian specimens from the Timan Mountains. The slight differences in the umbilical portions of the suture of Miller’s specimen seem to be due, as he supposed, to distortion during fossilization. There seems no reason to keep a separate name for the Canadian species. Details of the wrinkle layer, and the inner whorls, now appear the same in the Russian and Canadian specimens. This genus appears to be restricted to the hmiilicosta Zone. In addition to the type species. Fetter (1959) has described two species from the Sahara from the hmiilicosta Zone as T. meridionalis and T. complanatwn. Certainly the former and probably the latter should be assigned to the genus Komioceras Bogoslovski (1958, p. 115). Horizon and localities. AU Dv684 from the Maligne Formation on the north-east shoulder of Roche Miette in Jasper Park, Alberta. AU 38975 is from the Maligne Formation at Sulphur Spring just east of a bridge over the Athabaska River in Jasper Park, Alberta. Family tornoceratidae Arthaber 1911 Genus tornoceras Hyatt 1884 Type species by original designation: Goniatites lanangidaris Conrad 1842. Tornoceras {Tornoceras) cf. westfaliciim Holzapfel Plate 72, fig. 15; text-fig. 11 Material. Five specimens preserved as solid limestone internal moulds of the body chambers and crystalline calcite internal moulds of the phragmocones. Some specimens are testate. EXPLANATION OF PLATE 74 Fig. 1. Gyroceratites (Lamelloceras) sp. from the Funeral Formation in the northern Arnica Range, Northwest Territories. GSC 16921, xF44. Fig. 2. IPlatyclymenia sp. from the Costigan Member of the Palliser Formation on the south-east slope of Mt. Lorette, Alb®rta. GSC 16963, X 1. Fig. 3. Anarcestes (Latanarcestes) cf. praeciirsor Freeh from the Funeral Formation in the Northern Funeral Range, Northwest Territories. GSC 16930, X T2. Figs. 4, 5. Tornoceras {Tornoceras) cf. crebriseptiim Raymond from Hume’s D5 unit, 6 miles south of the eastern tip of Carlson’s Lake. GSC 16954, X 1-67. Figs. 6, 7. Manticoceras cf. sinuosuni (Hall) from the Mount Hawk Formation on Crescent Creek, Nelson Range, Alberta. SOC T6207, X 1. Figs. 8, 9. Manticoceras sp. from the Redknife Formation at the Liard Rapids, Northwest Territories. GSC 16964, X 1-35. Figs. 10, 11. Imitoceras sp. from Hume’s D5 unit on the North Nahanni River, Northwest Territories. GSC 16960, x2. Palaeontology , Vol. 6 PLATE 74 HOUSE and REDDER, Devonian goniatites M. R. HOUSE AND A. E. H. REDDER: DEVONIAN GONIATITES 527 Dimensions {in mm.) D WW WH Wh UW GSC 16949 47-2 18-2 27-8 12-5 0 GSC 16950 37-2 15-5 c. 20-2 - 0 GSC 16951 c. 34 0 13-7 - - 0 GSC 16952 200 00 11-5 - c. 1-0 Description. Shell form laterally compressed, involute with closed umbilicus, subtrochoidal. Whorl section with maximum width close to the umbilicus and flanks sloping flatly to a broad well-rounded venter (text-fig. 11). Suture with V-shaped ventral lobe, a broad asymmetric lateral lobe and a short latero-umbilical saddle with a steep ventral face which flexures sharply over to the saddle crest. Suture similar at all stages seen. Growth-lines, typically for genus, form a weak salient on the lower flanks, a weak sinus on the middle flanks and then pass forward to a prominent ventro-lateral salient and back to a deep mid-ventral linguiform sinus. On one specimen (GSC 16950) there are several small rounded notches on the body chamber mould, presumably marking damage in life. The wrinkle layer is well seen on one specimen (GSC 16950) as fine posteriorly directed striae laid down on top of the preceding whorl in the dorsal part of the body chamber. TEXT-FIG. 1 1 . Tonioceras {Tonioceras) cf. westfaliciim (Holzapfel). Cross-section based on GSC 16949 from 10-20 feet above the top of the Hume Formation on Francis Creek, Northwest Territories. Natural size. Remarks. These specimens are interesting in showing a sub trochoidal shell form rather than the more typical subtegoidal form found commonly in Tornoceras. Also the umbilico-lateral saddle is unusually short. These are features shown especially by T. westfaliciim Holzapfel (1895, p. 104, pi. 4, figs. 11, 12) from the German Givetian, but the holographs of that species show a slightly open umbilicus with a shell ridge surrounding it. Unfortunately none of the larger species are testate in the umbilical region so that it cannot be demonstrated whether this feature also occurs. Unpublished work by the writer on the evolution of Tornoceras in the American Devonian suggests that species with such a sharp flexure between the ventral slope and crest of the umbilico-lateral saddle predominate in the United States at levels above the Centerfield Limestone, and T. westfaliciim occurs in Germany in levels which probably represent the upper Givetian. Horizon and locality. GSC 16949-53 (TOC H6053) collected by E. W. Best and D. Barss from an unnamed dark shale unit (the Lower Fort Creek and Hare Indian of authors) 10-20 feet above the top of the Hume Formation on Francis Creek, Northwest Territories, about 65° 14' 20" N., 126° 23' 40" W. Tornoceras {Tornoceras) cf. crebriseptiim Raymond Plate 74, figs. 4, 5 Material. Two specimens preserved as poor internal moulds of crystalline calcite and limestone. Dimensions {in mm.) D WW WH UW GSC 16954 17-4 8-0 9-8 1 GSC 16971 180 c. 8-0 - c. 0 528 PALAEONTOLOGY VOLUME 6 Description. Shell form very compressed, involute with closed umbilicus or nearly so. Maximum whorl width on the lower flanks, and the sides then converge convexly towards a well-rounded venter. Body chamber seen for half a whorl. Growth-lines and orna- ment not seen. Suture-lines not well preserved but showing a very small V-shaped ventral lobe, a ventro-lateral saddle close to the mid-ventral line passing almost rectilinearly back to a shallow lateral lobe. The ventral face of the umbilico-lateral saddle is moderately steep and the slope from the saddle crest to the umbilicus is gentle and even. Remarks. Unfortunately the suture is in neither case well preserved, but clearly it is not deeply undulating as in the Lower Famennian species T. acutnm Freeh, T. guestfalicum Freeh and others. The affinities clearly lie rather with the T. crebriseptiim group (see Miller 1938, p. 149), in which the sutural elements are relatively subdued. Locality and horizon. GSC 16954, 16971 (TOC G8353) collected by A. E. H. Pedder from Hume’s D5 unit, 564 to 574 feet below the top, and 614 to 604 feet above the base of the unit, from 6 miles south of the eastern tip of Carlson’s Lake, Northwest Territories, about 62° 2L N., 123° 43' W. Tornoceras (Tornoceras) sp. 1931 Tornoceras bicostatum Burgess, p. 200. 1938 (?) Tornoceras bicostatum Miller, pp. 168, 170. Material. The two plesiotypes of Burgess preserved in black mudstone. Remarks. Neither of these specimens show a ventro-lateral furrow so that they cannot belong to Aulatornoceras bicostatum. The larger specimen reaches 28 mm. diameter and shows biconvex growth-lines, with faint, sub-radial ribs on the lower flanks. The smaller specimen reaches 12-5 mm. diameter and shows a typical Tornoceras suture, with asym- metrical lateral lobe, but no evidence of growth lines. In both the umbilicus is closed and the shell form laterally compressed but well rounded. Horizon and locality. Both specimens are numbered MCZ 3473 and come from the Perdrix Formation (= Kiln Formation) from 3 miles south of Pocahontas on the east side of the Athabaska River, Alberta, and were found ‘in concretions scattered through a black fissile shale’. Genus lobotornoceras Schindewolf 1936 Type species by original designation : Goniatites ausavense Steininger 1853. EXPLANATION OF PLATE 75 Fig. 1, 2, 10, 11. Teicherticeras lenzi sp. nov. Holotype from 296 feet below the base of the Hume Formation equivalent on the Ogilvie River, Yukon. GSC 16929, X 1. 10, 11, The inner whorls of the same specimen, X 2. Fig. 4. Manticoceras sp. from the Mount Hawk Formation at the Saskatchewan River Gap, Alberta. AU Dvl707, X 1. Figs. 5-9. Timanites keyserlingi Miller from the Maligne Formation at Sulphur Spring just east of a bridge over the Athabaska River, Jasper Park, Alberta. AU 38975. 5, The wrinkle layer, X4. 6, 7, The inner whorls showing the dorsal septal line of the subsequent whorl, X 1. 8, 9, The outer whorls, Xl. Palaeontology, Vol. 6 PLATE 75 HOUSE and PEDDER, Devonian goniatites M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 529 Lobotornoceras aff. bilobatum (Wedekind) Plate 77, figs. 1, 2; text-fig. 12a Material. One specimen collected by A. E. H. Pedder preserved as an internal mould of the phragmo- cone in limestone. Form slightly crushed. Dimensions (in nnn.) D WfV WH UW GSC 16955 31-5 c. 13-8 18 0 c. 0 Description. Shell form laterally compressed, involute with closed umbilicus. Maximum whorl width close to the umbilicus and the flanks coverge sharply towards an apparently TEXT-FIG. 12. Lobotornoceras sp. from Germany and western Canada. A, L. aff. bilobatum (Wedekind). Suture at 34 mm. diameter based on GSC 16955 from Hume’s D5 unit near Carlson’s Lake, Northwest Territories. B. L. bilobatum (Wedekind). The original illustration of the sutures on the holotype (Wedekind 1908, pi. 39, fig. 35) from the CheUoceras Stufe at Engeberg, Germany. c, L. ausavense (Steininger). Suture diagram from Schindewolf ( 1936, p. 690, text-fig. 1 ) of a specimen from the Manticoceras Stufe at Biidesheim, Germany. All twice natural or stated size. well-rounded venter giving the whorl section a subtriangular form. Growth-lines and ornament not preserved. The suture is illustrated in text-fig. 12a. A small saddle can be seen on the specimen centred slightly dorsal of the umbilical seam. Remarks. The specimen clearly belongs to the Fammenian L. bilobatum-L. escoli group rather than the Frasnian group represented by the type species (text-fig. 12c). It seems probable these groups are phylogenetically unrelated, although both descended indubit- ably from Tornoceras s.s. Since the lateral lobe of the specimen from Canada is rounded, rather than pointed as in L. escoli (Freeh 1902, p. 48, pi. 3(2), fig. 19), the affinities are with L. bilobalum (Wedekind 1908, p. 579, pi. 39, fig. 35; pi. 40, fig. 8) from the Cheilo- ceras Stufe at Enkeberg, but the specimen differs from L. bilobalum in showing a wider 530 PALAEONTOLOGY, VOLUME 6 umbilico-lateral saddle (text-fig. 12b). Also, if the descriptions of Schindewolf (1923, p. 307) apply to this species, the whorl section is more flat-sided than L. bilobatum. Schindewolf (1923, p. 492) records the species in the lower Cheiloceras Stufe at Hof, and it occurs elsewhere in the Rheinische Gebirge both in the lower and upper parts of the Cheiloceras Stufe. Horizon and locality. GSC 16955 (TOC G854F), from talus about 150 feet above the base of Hume’s D5 unit, 6 miles south of the eastern tip of Carlson’s Lake, Northwest Territories, about 62° 21' N., 123° 43' W. Family cheiloceratidae Freeh 1897 Genus cheiloceras Freeh 1897 Type species by the subsequent designation of Wedekind (1917): Goniatites subpartitum Munster 1839. Cheiloceras {Cheiloceras) sacculum (G. and F. Sandberger) Plate 77, figs. 3, 4; text-fig. 13a, b 1851 Goniatites retrorsus var. saccidus G. andF. Sandberger, p. 109; pi. 10, fig. 22; pi. 106, figs. 7, 20, 22. 1908 Cheiloceras sacculum Wedekind, p. 584, pi. 39, fig. 2. 1917 Cheiloceras sacculum Wedekind, p. 146, pi. 18, fig. 11; text-fig. 46/;. 1923 Cheiloceras sacculum Schindewolf, p. 320. 1959 Cheiloceras sacculum Fetter, p. 233, pi. 17, figs. 17, \la, 18, 18u. Material. Two specimens preserved as phragmocones in solid limestone, mostly as internal moulds. Dimensions {in mm.) D WW WH UW GSC 16956 29 18 16 0 Description. Shell form rotund, laterally compressed, involute with closed umbilicus. Whorl section with maximum width close to the umbilicus and flanks sloping convexly towards a well-rounded venter. Shell form at early diameters is similar and is shown in text-fig. 13 a. Growth-lines have not been seen, but the wrinkle layer is shown as backwardly directed fine striae, spiral near the umbilicus, of the Tornoceras type. The suture (text-fig. 13b) shows a small, narrowly V-shaped ventral lobe and a broad flat ventro-lateral saddle: in the larger specimen (GSC 16956) the septal line, but not the suture, shows a weak concavity along this ‘ saddle ’. The lateral lobe is small and rounded and the ventral face of the latero-umbilical saddle rises steeply from it to form a wide, well-rounded saddle. Dorsal suture not seen. The smaller specimen (GSC 16957) shows, at a diameter of 19 mm., considerable approximation of the septa, for there are eight in the last quarter whorl, whereas the larger specimen has only eight in the last half whorl. EXPLANATION OF PLATE 76 Figs. 1-4, 7. Probeloceras sp. from the Imperial Paddle River 5-17-56-8W.5M well between 7,913 and 7,932 feet depth. 1, 2, GSC 16972, x2. 3, 4, IOC 258u, X 2-5. 7, A latex mould from IOC 2586, x2-5. Figs. 5, 6, 8-1 1. Manticoceras cordiforme Miller from the Carcajou Mountain Sandstone at Norman Wells, 45 miles north-west of Fort Norman, Northwest Territories. 5, 6, Holotype, GSC 2393, X 1. 8, 9, WM 51873, x 1. 10, 11, Holotype of M. “'septentrionale', GSC 5139, xO-5. Palaeontology, Vol. 6 PLATE 76 HOUSE and PEDDER, Devonian goniatites ft ft M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 531 Remarks. Of all the described species of Cheiloceras only C. sacculum shows such a rotund form and distinctive suture. C. amblylobiun (G. and F. Sandberger), which is the only other species of Cheiloceras to be described from North America (House 1962), does not show the small rounded lateral lobe. C. sacculum is typical of the upper Cheiloceras Stufe, that is the pompeckji Zone, in Germany, but Schindewolf ( 1923, p. 321) has noted that the species also occurs in the underlying curvispina Zone. Horizon and locality. Both specimens collected by A. E. H. Pedder; GSC 16956-7 (TOC G8342) come from scree below Hume’s D5 beds, from about 150 feet above the base, 6 miles south of the eastern tip of Carlson Lake, Northwest Territories, about 62° 21' N, 123° 43' W. TEXT-FIG. 13. Cheiloceras (Cheiloceras) sacculum (G. and F. Sandberger). a, Cross-section, b. Suture. Both based on GSC 16956 at 29 mm. diameter, from Hume’s D5 unit near Carlson’s Lake, Northwest Territories. Both X 2. Cheiloceras (C.) cf. sacculum (G. and F. Sandberger) A single specimen, GSC 16958 (TOC G5487), poorly preserved, from Hume’s D5 beds collected by I. Zemmels may be compared with C. sacculum. It comes from an isolated outcrop 5 miles SSW of the confluence of Battlement Creek and the North Nahanni River, about 62° 7' 35" N., 123° 38' W. Genus sporadoceras Hyatt 1884 Type species by original designation : Goniatites bidens G. and F. Sandberger 1851. Sporadoceras cf. primaevum Schindewolf Plate 77 figs. 6-7; text-fig. 14a-c 1923 Sporadoceras primaevum Schindewolf, p. 340, text-fig. 5a. Material. One specimen collected by A. E. H. Pedder preserved as a phragmocone in solid lime- stone mostly as an internal mould. Dimensions (in mm.) D WW WH UW GSC 16959 810 300 48-0 0 600 25-5 c. 32 0 1-5 532 PALAEONTOLOGY, VOLUME 6 Description. Shell form compressed, involute with closed umbilicus and subtegoidal in the adult, where the maximum whorl width lies close to the umbilicus and the flanks slope with increasing convexity to the well-rounded venter. The shell form at earlier diameters is illustrated by a cross-section which shows that the early whorls are globular with an open umbilicus (text-fig. 14a). Suture when first seen at about 45 mm. diameter shows a narrow ventral lobe, a broad subsemicircular ventro-lateral saddle with a slight flattening on the dorsal side, and a TEXT-FIG. 14. Sporadoceras cL primaeviim Schindewolf. A, Cross-section. B, Suture at 73 mm. diameter, c, Suture at c. 45 mm. diameter. All based on GSC 16959 from Hume’s D5 unit near Carlson's Lake, Northwest Territories. Alt natural size. subacute lateral lobe tilted ventrally at the apex (text-fig. 14c). At greater diameters the flattening on the dorsal face of the ventro-lateral saddle becomes more concave (text- fig. 14b). Dorsal suture not seen. Growth-lines are poorly seen but appear to be convex. Remarks. Some might argue that the A2 lobe is insufficiently developed to warrant calling this species Sporadoceras rather than Cheiloceras. Indeed, there is some justifica- tion for claiming that such a lobe is incipient until a radial line from the umbilicus would EXPLANATION OF PLATE 77 Figs. 1, 2. Lobotornoceras aff. bilobatum (Wedekind) from talus below Hume’s D5 unit 6 miles south of the eastern tip of Carlson’s Lake, Northwest Territories. GSC 16955, x2. Figs. 3, 4. Cheiloceras {Cheiloceras) saccuhan (G. and F. Sandberger) from the same locality and horizon. GSC 16956, X 1-25. Fig. 5. llmitoceras sp. from the same locality and horizon. GSC 16961, X 1. Figs. 6-8. Sporadoceras cf. primaevian Schindewolf from the same locality and horizon. GSC 16959, X M2. Palaeontology, Vol. 6 PLATE 77 HOUSE and PEDDER, Devonian goniatites X... ■m.t. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 533 cut both sides of the flexure. But it will avoid confusion to follow the usage of Schinde- wolf. The relations of this species have been discussed by Schindewolf (loc. cit.) and he has noted that the holotype was found in, and the range restricted to, the lowest beds of Zone IIj8 at Gattendorf, that is, the lower part of the pompeckji Zone. Correlation with this level is supported by the evidence of C. saccuhim, which is associated with S. cf. primaevum in Hume’s D5 beds. If Schindewolf’s illustration of the suture is natural size, and this is not stated, then the Canadian specimen shows a much later development of the Ag lobe than the holo- type. This suggests that it may be slightly younger than the true S. primaevum. Horizon arid locality. GSC 16959 (TOC 8342), collected from talus about 150 feet above the base of Hume’s D5 unit, 6 miles south of the eastern tip of Carlson's Lake, Northwest Territories, about 62° 21' N., 123° 43' W. Family imitoceratidae Ruzhencev 1950 Genus imitoceras Schindewolf 1923 Type species by original designation: Ammonites rotatorius de Konink. Imitoceras sp. Plate 74, figs. 10, 1 1 Material. One specimen collected by Dr. Paul Sartenaer preserved as an internal mould of the phragmo- cone in limestone. Dimensions (in mm.) D fTIT UW GSC 16960 22-5 15-3 0 Description. Shell form subglobular, slightly compressed, involute with closed umbilicus. Maximum whorl width close to the umbilicus ; well-rounded venter. Growth-lines not seen. No trace of constrictions. Suture shows typical elements of the genus (PI. 74, figs. 10, 11). Remarks. The speeimen is too poorly preserved to be adequately compared with the many Devonian described species. In shell proportions it approaches I. lineare (Munster) amongst others. Imitoceras first appears in Europe in the Cheiloceras Stufe, and this is the level of the Canadian specimen, showing that the genus was widely distributed even at its first appearance. Horizon and locality. GSC 16960 (loc. 38701) collected from the lower 355 feet of the outcrop of Hume’s D5 beds on the North Nahanni River, 62° 2T N., 123° 42' W. llmitoceras sp. Plate 77, fig. 5 A large specimen, GSC 16961 (TOC G8342), collected by R. de Wit from scree below the type section of Hume’s D5 beds, shows a typical Imitoceras-type lateral lobe but evidence at the umbilicus of a shallow saddle centred on the seam. Unfortunately the specimen is badly crushed. The same curious feature is shown on another poorly pre- served specimen (GSC 16962) colleeted by A. E. H. Pedder. This may be caused by crushing or by a dorsal displacement of the umbilical seam. Both specimens belonged 534 PALAEONTOLOGY, VOLUME 6 to the Triad Oil Co. collections and came from talus about 150 feet above the base of Hume’s D5 beds, 6 miles south of the eastern tip of Carlson’s Lake, about 62° 21' N., 123° 43' W. Family clymeniidae Edwards 1849 Genus platyclymenia Hyatt 1884 Type species by subsequent designation : Goniatites annulatus Munster 1832. 1 Platyclymenia sp. Plate 74, fig. 2 Remarks. This poorly preserved and weathered specimen shows no trace of ornament. It is crushed but gives the impression of having been laterally compressed in shell form, and evolute. These characters and the simple form of the suture (PL 74, fig. 2: the siph- uncle has not been seen) suggest reference to Platyclymenia rather than to any other Upper Devonian ammonoid genus. Horizon and locality. GSC 16963, collected by A. E. H. Pedder from 0-16 feet below the top of the Costigan Member of the Palliser Formation on the south-east slope of Mt. Lorette, Fisher Range, Alberta, about 50° 59' 30" N., 115° 08' W. A B C D TEXT-FIG. 15. Anaptychi from Carcajou Ridge and Thunder River, a, GSC 16965 (CSC L-17). [ c, GSC 16967 (CSC L-17). Both from the Fort Creek Shale at Thunder River, b, GSC 16966 (CSC ! L-8). D, GSC 16968 (CSC L-8). Both from an unnamed shale 10-40 feet above the Beavertail Forma- tion on Carcarjou Ridge. All natural size. Anaptychi Associated with Pontieeras cf. tschernyschewi in an unnamed shale 10-40 feet above the top of the Beavertail Formation on Carcajou Ridge, Northwest Territories, are several anaptychi collected by A. Lenz. These are poorly preserved as shiny films on bedding surfaces of the black shale. Outlines of two are illustrated here (text-fig. 15b, d). The rounded form, with shallow impressed region, is reminiscent of the whorl section of species of Pontieeras not uncommon in the German himilieosta Zone but for which there is no other evidence at Carcajou Ridge. From the type section of the Fort Creek Shale at Thunder River come two types of anaptychi. The first is long and narrow (text-fig. 1 5a) with a slight axial concavity which becomes more prominent towards the crest of the impressed depth. This form is not M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 535 dissimilar to the whorl section of P. tschernyscliewi itself (Holzapfel 1899, pi. 4, figs. 3a, 6a) and may represent the operculum of the comparable Ponticeras common at Carcajou Ridge. However, it is curious that this form has not been found at Carcajou Ridge and that Holzapfel recorded no anaptychi of this type from the Russian beds with P. tschernyscliewi. The second type of anaptychus from the Fort Creek Shale is more nearly rounded (text-fig. 15c), but with a deeper impressed depth than that of the commonest forms at Carcajou Ridge. Again the whorl section is reminiscent of some lower Frasnian species of Ponticeras. A poorly preserved specimen from the same horizon as the anaptychi may be a Ponticeras, but in any case the evidence seems to suggest that in the type area of the Fort Creek Shale both Givetian and lowest Frasnian faunas occur in the formation. Abbreviations. For simplicity the following abbreviations are used throughout the text. Measurements: dimensions are given in millimetres and apply to the stated diameter unless other- wise indicated. D =- diameter. UW = umbilical width. WH = whorl height. Wh ^ ^ distance from the ventral crest of the whorl to the ventral crest of the preceding whorl. WW ^ whorl width. Museum Collections: AMNH = The American Museum of Natural History, New York. AU = Geology Department, University of Alberta, Edmonton. CSC = The California Standard Company. GPIG = Geologisch-Palaontologisches Institut, Gottingen. GSC = Geological Survey of Canada, Ottawa. GSC loc. = Geological Survey of Canada locality number. IOC = Imperial Oil Ltd. MCZ = Museum of Comparative Zoology, Cambridge, Massachusetts. NYSM = New York State Museum, Albany. PAPC = Pan American Petroleum Corporation. SMF = Geologisches Abteilung Natur-Museum Senckenberg, Frankfurt am Main. SOC = Shell Oil Co. of Canada Ltd. TOC = Triad Oil Co. Ltd. WM = Walker Museum, University of Chicago, Illinois. Location of specimens. Through the generosity of the managements of the oil companies concerned it has been possible to present most of the mentioned material to the Geological Survey of Canada. In the text both the Survey collection numbers and the relevant oil company specimen numbers are given, the latter in parenthesis. Plaster casts of most figured specimens are deposited in the collections of the Geology Department, University of Durham. Acknowledgements. For providing specimens or information, or both, we are indebted to: E. W. Best of Triad Oil Co. Ltd., G. A. Cooper of the U.S. National Museum, C. H. Crickmay of Imperial Oil Ltd., D. E. Jackson of Pan American Petroleum Corporation, A. C. Lenz of the California Standard Com- pany, D. J. McLaren of the Geological Survey of Canada, M. H. Nitecki of the Walker Museum, Chicago, G. O. Raasch of Shell Oil Co. of Canada Ltd., C. R. Stelck of the University of Alberta, J. W. Wells of Cornell University and H. B. Whittington of Harvard University. One of us (M. R. H.) is indebted to the Commonwealth Lund of New York for the award of a fellowship which enabled the study of comparative material in the United States, and to Durham University for similar aid in Europe. We are also indebted to G. Dresser of Durham University and a grant from the D.S.I.R. for prepara- tion of all the photographs used on the plates, and to T. Glarum, of Triad Oil Co. Ltd., and B. Whan of New England University for preparing text-figs. 1 and 2. Einally we owe our thanks to the managements of the oil companies concerned who have so generously allowed us to study and publish on their material. 536 PALAEONTOLOGY, VOLUME 6 REFERENCES Alberta Oil and Gas Conservation Board. 1960. Table of Formations — Alberta. Calgary. Alberta Society of Petroleum Geologists. 1960. Lexicon of geologic Names in the Western Canada Sedimentary Basin and Arctic Archipelago. Calgary. ALLAN, J. A. et al. 1932. A preliminary study of the eastern ranges of the Rocky Mountains in Jasper Park, Alberta. Trans, roy. Soc. Can. (3), 26, sect. 4, 225-49. American Commission on Stratigraphic Nomenclature. 1961. Code of stratigraphic nomenclature. Bull. Amer. Ass. Petrol. Geol. 45, 645-55. d'archiac, a. and de verneuil, m. e. 1842. On the fossils of the older deposits in the Rhenish pro- vinces. Trans, geol. Soc. Lond. ser. 2, 6, 303-410, pi. 25-38. BARRANDE, J. 1865-7. Systeme Silurien dii centre de la Boheme. lere Parte: Recherches Paleonto- logiques. 2, Cephalopodes. Atlas (i), 1865. Text (i), 1867. Prague, Paris. BASSETT, H. G. 1961. Devonian stratigraphy, central Mackenzie River region. Northwest Territories, Canada. In Geology of the Arctic, 1, 481-98. Toronto. BEACH, H. H. 1943. Moose Mountain and Morley map areas, Alberta. Mem. geol. Surv. Can. 236. BELYEA, H. R. 1956 (1955 on cover is incorrect). Cross sections through the Devonian system of the Alberta plains. Pap. geol. Surv. Can. 55-3. and MCLAREN, D. J. 1956. Devonian sediments of Bow Valley and adjacent areas. Alberta Soc. Petrol Geol., Guidebook Ann. field conf. 6, 66-91. 1962. Upper Devonian formations, southern part of Northwest Territories, northeastern British Columbia, and northwestern Alberta. Pap. geol. Surv. Can. 61-29. BISSELL, B. 1930. The Devonian section at Roche Miette, Alberta. Unpublished M. Sc. Thesis, Cornell University, N.Y. BOGOSLOVSKi, B. I. 1958. The Devonian ammonoids of the Rudnyi Altai (in Russian). Trudy paleont. Inst. Akad. Nauk SSSR. 64, 1-155, pi. 1-9. BOSWORTH, T. o. 1921fl, June. The Mackenzie oilfield of northern Canada. Geol. Mag. 58, 287, 288. 19216, Oct. The Mackenzie oilfield of northern Canada. J. Petrol. Tech. 7, 276-9. BURGESS, c. H. 1931. The Kiln Shale fauna. Bull. Mus. comp. Zool. Harv. 72, 195-202. CAMERON, A. E. 1918. Explorations in the vicinity of Great Slave Lake. Summ. Rep. geol. Surv. Can. 1917, C, 21-28. and WARREN, p. s. 1938. Geology of South Nahanni River, N.W.T. Canad. Fid. Nat. 52, 16-18. CAMPBELL, N. 1950. The Middle Devonian in the Pine Point area, N.W.T. Proc. geol. Ass. Can. 3, 87-96. CHLUPAC, I. 1959. Facialni vyvoj a biostratigrafie bfidlic dalejskych a vapencu hlubocepskych (eifel) ve stfedoceskem devonu. Shorn. Ustr. Vst. geol. 25, 445-511, pi. 62-66. CLARKE, J. M. 1885. On the higher Devonian faunas of Ontario County, New York. Bull. U.S. geol. Surv. 16, 1-80, pi. 1-3. 1898. The Naples fauna (fauna with Manticoceras intumescens) in Western New York, Part 1. 16th Ann. Rep. N.Y. St. Geol. 29-161, pi. 1-9. CONRAD, T.A. 1842. Observations on the Silurian and Devonian systems of the United States, with descriptions of new organic remains. J. Acad. nat. Sci. Philad. 8, 228-80, pi. 12-18. COPELAND, M. J. 1960. A Kinderhook microfauna from Crowsnest Pass, Alberta. Trans, roy. Soc. Can. (3), 54, sect. 4, 37-43. CRiCKMAY, c. H. 1950. Some Devonian Spiriferidae from Alberta. J. Paleont. 24, 219-25. 1952. Discrimination of late Upper Devonian. Ibid. 26, 585-609. 1953. New Spiriferidae from the Devonian of Western Canada. Calgary. 1957. Elucidation of some Western Canada Devonian Formations. Calgary. 1960o. The Older Devonian faunas of the Northwest Territories. Calgary. 19606. Studies of the western Canada Stringocephalinae. J. Paleont. 34, 874-90. DOUGLAS, R. J. w. 1959. Great Slave and Trout River map areas. Northwest Territories. Pap. geol. Surv. Can. 58-11. and NORRIS, d. k. 1960. Virginia Falls and Sibbeston Lake map areas. Northwest Territories. Ibid. 60-19. 1961. Camsell Bend and Root River map areas. Northwest Territories. Ibid. 61-13. DRAKE, N. F. 1893. Report on the Colorado coal field of Texas. Rept. Texas geol. Surv. 4, 357-446. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 537 ERBEN, H. K. 1953. Goniatitacca (Ceph.) aus dem Unterdevon und dem unteren Mitteldevon. Neiies Jb. Geol. Palaont. 98, 175-225, pi. 17-19. 1960. Primitive Ammonoidea aus dem Unterdevon Frankreiches und Deutschlands. Ibid. 110, 1-128, pi. 1-6. EOORD, A. H., and crick, g. c. 1897. Catalogue of the Fossil Cephalopoda in the British Museum (Natural History), Part 3. London. FOX, F. G. 1951. Devonian stratigraphy of Rocky Mountains and Foothills between Crowsnest Pass and Athabaska River, Alberta, Canada. Bull. Amer. Ass. Petrol. Geol. 35, 822-43 (this also appeared in 1954 in West Can. seel, basin symp. 109-30, Tulsa). FRECH, F. 1897-1902. Lethaea geognostica: Theil 1, Lethaea palaeozoica, 2. 1902. Uber devoirische Ammoneen. Beitr. Palaont. Geol. Ost.-Ung. 14, 27-111, pi. 2-5. ■ 1913. Ammoniac Devonicae (Clymeniidae, Aphyllitidae, Gephyroceratidae, Cheiloceratidae). Fossilium Catalogus, 1. Animalia, 1-42. Berlin. HAGE, c. o. 1945. Geological reconnaissance along Lower Liard River &c. Pap. geol. Surv. Can. 45-22. HAGUE, A. 1883. Geology of the Eureka district, Nevada. Rep. U.S. geol. Surv. 3, 237-90. HALL, J. 1843. Geology of New York. Part IV, Comprising the fourth geological district. Albany. 1874. Descriptions of new species of Goniatidae, with a list of previously described species, p. 1-4. ?Albany (this also appeared in 1875 in the Rep. N.Y. St. Mas. 27, 132-36). 1879. Descriptions of the Gasteropoda, Pteropoda, Cephalopoda, Crustacea and Corals of the Upper Helderberg, Hamilton and Chemung Groups. N.Y. geol. Surv. Palaeontology of New York, 5, pt. 2. HANNA, G. D. 1926. Palcontology of Coyote Mountain, Imperial County, California. Proc. Calif. Acad. Sci. (4), 14, 427-503. HOLZAPFEL, E. 1882. Die goniatiten Kalke von Adorf in Waldeck. Palaeontographica, 28, 227-61, pi. 44-49. 1895. Das obere Mitteldevon (Schichten mit Stringocephalus Bartini und Maeneceras terebratum) im rheinischen Gebirge. Abh. preuss. geol. Landesanst., n.f., 16, 1-459, pi. 1-19. ■ 1899. Die Cephalopoden des Domanrk im siidlichen Timan. Man. Com. geol. 12, no. 3, 1-56, pi. 1-10. HOUSE, M. R. 1961. Acanthoclymenia, the supposed earliest Devonian clymenid, is a Manticoceras. Palaeontology, 3, 472-6, pi. 75. 1962. Observations on the ammonoid succession of the North American Devonian. /. Paleont. 36, 247-84, pi. 43-48. ■ 1963. Devonian ammonoid successions and facies in Devon and Cornwall. Quart. J. geol. Soc. Lond. 119, 1-27, pi. 1-4. HUME, G. s. 1922. North Nahanni and Root Rivers area and Caribou Island, Mackenzie River district. Summ. Rep. geol. Surv. Can. 1921, B, 67-78. 1954. The lower Mackenzie River area. Northwest Territories and Yukon. Mem. geol. Surv. Can. 273. and LINK, T. A. 1945. Canol geological investigations in the Mackenzie River valley. Pap. geol. Surv. Can. 45-16. HUNT, c. w. 1954. Normal Devonian sequence of southern Mackenzie basin, Western Canada. Bull. Amer. ^4^5. Petrol. Geol. 38, 2290-301. HYATT, A. 1883-4. Genera of fossil Cephalopoda. Proc. Boston Soc. nat. Hist. 22, 253-338. Imperial Oil Limited. 1950. Devonian nomenclature in Edmonton area, Alberta, Canada. Bull. Amer. Ass. Petrol. Geol. 34, 1807-25. KAYSER, E. 1872. Studien aus dem Gebiete des rheinischen Devon. 3,Die FaunadesRotheisensteinvon Brilon in Westfalen. Z. dtsch. geol. Ges. 24, 653-89, pi. 25-27. KELLY, w. A. 1939. Devonian and Mississippian stratigraphy of Jasper Park, Alberta. (Abstract.) Bull. geol. Soc. Amer. 59, 2000. KEYSERLiNG, A. G. 1844. Bcschrcibung einiger Goniatiten aus dem Domanik-Schiefer. Verb, russ.- Kayserl. min. Ges. St. Petersb. 218-55. KEYSERLING, E. 1 846. WissenschaftUclie Beobachtungen auf einer Reise in das Petschora-Land im Jahre 1843. St. Petersburg. 538 PALAEONTOLOGY, VOLUME 6 KINDLE, E. M. 1919. The discovery of a Portage fauna in the Mackenzie River valley. Miis. Bull. Can. geol. Siirv. 29. 1929. The succession of fossil faunas in the eastern part of Jasper Park, Alberta. Amer. J. Sci. (5), 18, 177-92. 1936. Formation names of the Mackenzie River valley. Science, 83, no. 2140, 14, 15. and BoswoRTH, t. o. 1921 . Oil bearing rocks of Lower Mackenzie River valley. Sitmm. Rep. geol. Surv. Can. 1920, B, 37-63. KINGSTON, D. R. 1951. Stratigraphic reconnaissance along upper South Nahanni River, Northwest Territories, Canada. Bull. Amer. Ass. Petrol. Geol. 35, 2409-26. LAW, j. 1955. Geology of northwestern Alberta and adjacent areas. Ibid. 39, 1927-75. MATERN, H. 1931. Das Oberdevon der Dill-Mulde. Abh. preuss. geol. Lanciesanst., n.f. 134, 1-139, pi. 1-4. MAURiER, F. 1876. Palaontologische studien im Gebiet des rheinischen Devon. 3, Die Thonschiefer des Ruppbachtales bei Diez. Neues Jb. Min. Geol. Paldont. (1876), 808-48. MCCONNELL, R. G. 1887. Report on the geological structure of a portion of the Rocky Mountains, &c. Ann. Kept. Geol. not. Hist. Surv. Can. 2, D, 1-41. 1891. Report on an exploration in the Yukon and Mackenzie basins. Northwest Territories. Ibid. 4, D, 1-163. MCLAREN, D. J. 1954. Upper Devonian rhynchonellid zones in the Canadian Rocky Mountains. In West. Canada sed. basin symp., 159-81. Tulsa. 1956 (1955 on cover is incorrect). Devonian formations in the Alberta Rocky Mountains between Bow and Athabasca Rivers. Bull. geol. Surv. Can. 35. MERRiAM, c. w. 1940. Devonian stratigraphy and paleontology of the Roberts Mountains region, Nevada. Spec. Pap. geol. Soc. Amer. 25. MILLER, A. K. 1938. Devonian ammonoids of America. Ibid. 14. and WARREN, p. s. 1936. A Timanites from Upper Devonian beds of America. J. Paleont. 10, 632-6. PAMENTER, c. B. 1956. Imitoceras from the Exshaw Formation of Alberta. Ibid. 30, 965, 966. REDDER, A. E. H. 1963. Alaiopliyllum mackenziense sp. nov., a Devonian tetracoral from Canada. Palaeontology, 6, 132-5, pi. 19. PETTER, G. 1959. Goniatites devoniennes du Sahara. Mem. Carte geol. Alger. Paleont. 2, 1-313, pi. 1-26. PRANTL, F. 1954. Biostratigraficky vyznam ceskych goniatitu. Cas. ndr. Mus. Odd. Pfirod. 123, 15-19. RAYMOND, p. E. 1930. The Paleozoic formations in Jasper Park, Alberta. Amer. J. Sci. (5), 20, 289-300. RICKARD, L. V. 1952. The Middle Devonian Cherry Valley Limestone of eastern New York. Amer. J. Sci., 250, 511-22. SANDBERGER, G. and SANDBERGER, F. 1850-56. Systcmatische Beschreibung und Abbildimg der Ver- steinerungen des rheinischen Schichtensystems in Nassau. 1-564, pi. 1-39 (Atlas). Wiesbaden. SCHINDEWOLF, o. H. 1923. Beitrage zur Kenntnis des Palaozoicums in Oberfranken, Osthiiringen und dem Sachsischen Vogtlande, 1, Stratigraphie und Ammoneenfauna des Oberdevons von Hof a. S. Neues Jb. Min. Geol. Paldont. 49, 250-357, 393-509, pi. 14-18. 1933. Vergleichende Morphologie und Phylogenie der Anfangskammern tetrabranchiater Cephalopoden. Abh. preuss. geol. Landesanst., n.f. 148, 1-115, pi. 1-4. 1959. Adolescent cephalopods from the Exshaw Eormation of Alberta. J. Paleont. 33, 971-6. SCHMIDT, H. 1958. Deutschland, Mittel- und Oberdevon. Lexique Strat. Int. 1, fasc. 5b, 305-54. SPURR, J. E. 1898. Geology of the Yukon gold district, Alaska. Rep. U.S. geol. Surv. 18, (3), 87-392. STEWART, J. s. 1945. Recent exploratory deep well drilling in Mackenzie River valley. Northwest Territories. Pap. geol. Surv. Can. 45-29. STOREY, T. p. 1961. Devonian stratigraphy — Norman Wells region (abstract). In Geology of the Arctic, 1, 499. Toronto. SWEET, w. c. and miller, a. k. 1956. Goniatites from the Middle Devonian Columbus Limestone of Ohio. J. Paleont. 30, 811-17, pi. 94. taff, j. a. 1902. Description of the Atoka quadrangle. Atlas U.S. geol. Surv. 79. TAYLOR, p. w. 1957. Revision of Devonian nomenclature in the Rocky Mountains. J. Alberta Soc. Petrol. Geol. 5, 183-95. M. R. HOUSE AND A. E. H. PEDDER: DEVONIAN GONIATITES 539 TAYLOR, P. w. 1958. Further data on Devonian correlations. Ibid. 6, 13-19. TEiCHERT, c. 1948. Middle Devonian goniatites from the Buchan District, Victoria. J. Paleont. 22, 60-67, pi. 16. VANUXEM, L. 1842. Geology of New York, Part III, Comprising the survey of the third geological district. Albany. WALCOTT, c. D. 1908. Nomenclature of some Cambrian Cordilleran formations. Smithson, misc. Coll. 53 (1804), 1-12. 1913. Cambrian formations of the Robson Peak district, British Columbia and Alberta, Canada. Ibid. 57 (12), 327-43. 1921. Explorations and field work of the Smithsonian Institution in 1920. Ibid. 72 (6), 1-10. 1924. Geological formations of Beaverfoot-Brisco-Stanford Range, British Columbia, Canada. Ibid. 75 (1), 1-51. WALDSCHMiDT, E. 1885. Uber die devonischen Schichten der Gegend von Wildungen. Z. dtsch. geol. Ges. 37, 906-27. WARREN, p. s. 1927. Banff area, Alberta. Mem. geol. Siirv. Can. 153. 1937. Age of the Exshaw Shale in the Canadian Rockies. Amer. J. Sci. (5), 35, 454-7. 1944. Index brachiopods of the Mackenzie River Devonian. Trans, roy. Soc. Can. (3), 38, sect. 4, 103-35. 1949. Fossil zones of the Devonian of Alberta. Bull. Amer. Ass. Petrol. Geol. 33, 564-71. 1956. The Exshaw Shale. J. Alberta Soc. Petrol. Geol. 4, 141, 142. and STELCK, c. r. 1949. The late Middle Devonian unconformity in northwestern Canada. Trans. roy. Soc. Can. (3), 43, sect. 4. 139-48. 1950. Succession of Devonian faunas in western Canada. Ibid. (3), 44, sect. 4, 61-78. 1956. Devonian faunas of western Canada. Spec. Pap. geol. Ass. Can. 1. WEDEKIND, R. 1908. Die Cephalopodenfauna des hoheren Oberdevon am Enkeberge. Neites Jb. Min. Geol. Paldont. 26, 565-634. 1913. Die Goniatitenkalke des unteren Oberdevon von Martenberg bei Adorf. Ges. Natiirf Freunde Berl. 23-77, pi. 4-7. 1917. Die Genera der Palaeoammonoidea (Goniatiten), (mit Ausschluss der Mimoceratidae, Glyphioceratidae und Prolecanitidae). Palaeontographica, 62, 85-184, pi. 14—22. WHiDBORNE, G. F. 1889-92. A monograph of the Devonian faunas of the South of England. 1. Palaeontogr. Soc. (Monogr.). WHiTEAVES, J. F. 1891. The fossils of the Devonian rocks of the Mackenzie River basin. Contr. Canad. Palaeont. 1, 197-253. WHITTAKER, E. J. 1922. Mackenzie River between Great Slave Lake and Simpson. Siimm. Rep. geol. Siirv. Can. 1921, B, 45-55. wiNCHELL, N. H. 1888. The geology of Minnesota. Final Rep. Minn. geol. nat. Hist. Siirv. 2. de WIT, R. and mclaren, d. j. 1950. Devonian sections in the Rocky Mountains between Crowsnest Pass and Jasper, Alberta. Pap. geol. Siirv. Can. 50-23. M. R. HOUSE Department of Geology and Mineralogy, University Museum, Oxford A. E. H. REDDER Department of Geology, University of New England, Armidale, N.S.W., Manuscript received 22 November 1962 Australia. ON GROWTH STAGES IN BRANCHIOS AURS by D. M. S. WATSON Abstract. Lower Permian material from Niederhasslich, Friedrichroda, and Odernheim in the author’s collec- tion and in the British Museum (Natural History) is reviewed and described. A new species, Branchiosaurus brachyrhy)ichus sp. nov., is erected for specimens from Friedrichroda; B. flagrifer Whittard is redefined. New combinations are Branchiosaurus {Micromelerpetoii) credneri (Bulman and Whittard) and B. (Leptorophus) levis (Bulman). The young amphibia of the Lowest Permian referred to the genus Branchiosaurus were exceedingly well described many years ago by Credner (1882, 1885, 1886) and placed by him in the order Phyllospondyli; he dealt with the abundant material from the middle Rotliegende of Niederhasslich, Plauen’sche Grund, Dresden. Since that time similar series of different stages of growth have been found at Odernheim, Rheinpfalz (in the highest part of the lower Rotliegende), and at Gottlob, Friedrichroda, Thuringia (Godlauterer Schichten; the highest part of the middle Rotliegende), though the material is less abundant. These do not provide so good a series as that set out by Credner, but so far as they go they agree in essence with it. Since this time various authors have contributed to our general knowledge of bran- chiosaurs, including Bulman and Whittard (1926), Bulman (1928), Whittard (1930), and Steen (1938). Romer (1939) discussed them, and showed that the order Phyllospondyli was apparently founded on larval specimens of labyrinthodonts, the adults of which have presumably in some cases been described under other names. This interpretation seems to be essentially correct (I continued to use the term Phyllospondyli in 1940 only because I had not then seen Romer’s paper), and naturally leads to a hunt for specimens of intermediate size connecting the branchiosaurs with the large labyrinthodonts into which they grew. In this paper Romer (1939, fig. 2) selects from Credner’s figures a series of eight contemporary skulls to show that Onchiodon may have developed from ^Branchiosaurus' \ the smallest is 10 mm. long, the largest 120 mm., and the intermediates vary between 22 mm. and 56 mm., thus covering what is evidently a very long period of growth. Although there are some difficulties — for instance f seems to me quite clearly not a member of the series to which its neighbours e and G belong — the figure does suggest that the growth of labyrinthodonts involves changes in skull character which are reasonably represented in it. Parrington’s paper on the labyrinthodont middle ear (1959) is important, for it discusses in a most helpful way the change of position of the tympanic membrane, brought out by comparison of a small branchiosaur with a large labyrinthodont. He accepts Romer’s figure 2 as a real growth series of the form Onchiodon, and points out that the tympanic membrane in the larva extends laterally to end above the attachment of the lower jaw, whilst in the adult, twelve times as long, the membrane lies high up, immediately lateral to the narrow skull roof in the occipital region, and does not reach the jaw articulation, but is nevertheless related to the same bones as it was in the larva. [Palaeontology, Vol. 6, Part 3, 1963, pp. 540-53.] D. M. S. WATSON; GROWTH STAGES IN BR ANCHIOS AURS 541 PRESERVATION AND PREPARATION OF MATERIAL I have in my collection some beautifully preserved specimens of branchiosaurs from Niederhasslich, Friedrichroda, and Odernheim, covering a considerable range in size, and have been able to borrow others from the British Museum (Natural History). It seems, therefore, that it may be useful to figure a size range from each of these places in order to see what the differences are. Careful restored drawings of individuals, when arranged in order of size, could be expected to show not only whether they belonged to the same species or not, but also proportionate changes due to growth, and thus give an indication of the probable nature of the adult into which they should have grown. The material used is as follows: From Niederhasslich, D.M.S.W., B. 91, 92, 97; B.M.N.H., R. 201 1, 271 1. From Friedrichrona, D.M.S.W., B. 36a, 37, 48-52; B.M.N.H., R. 5281-7, 5466/7, 5469. From Odernheim, D.M.S.W., B. 25-35, 39, 40, 44-47, 141; B.M.N.H., R. 5026, 5028, 6700. Several of the specimens show more than one individual. Some of them have already been figured by Bulman and Whittard (1926, 1928, 1930), but I have made new restorations of dorsal and some lateral aspects which differ slightly from theirs (at that time my collection had not been catalogued, and they gave to specimens temporary numbers which have since been replaced by permanent ones). The material from the respective localities does not differ greatly in geological age. The matrix of both Niederhasslich and Odernheim specimens is an exceedingly hard, very calcareous rock, fine-bedded when seen in broken section, and at Niederhasslich grading into a very tough but less notieeably bedded limestone, whose colour on a clean fracture is light brownish-grey, weathering considerably lighter on a joint face. The Odernheim matrix is dark grey on a fresh surface, again weathering lighter. The Friedrichroda matrix is a fine-bedded, black shale which breaks flatly and a little un- certainly into thin slabs, looking very like some Coal Measure shales. In these very small animals little preparation can be done, and one is dependent on the faets shown by the original fracture revealing the specimen; sometimes they may be improved by the use of acid to dissolve the bone, for plasticene squeezes made from such moulds often show better surfaee detail than bones prepared out. The material has limitations: drawings can, as a rule, only be made of the dorsal aspect of the skull, whose roof pattern (so far as the number and general relationships of the individual bones are concerned) is uniform and modified only by changes in proportion in the bones involved. In very small specimens the lachrymal is usually badly preserved, so that its apparent shape may be determined by the borders of the surrounding bones, and has no independent validity. The palate is not often shown, and in many specimens the postcranial region is of no help in determining affinities for it is missing. The practical difficulties of making the drawings were met by traeing from enlarged photographs of each skull on which the outline of bones (or their moulds) had been earefully inked in. The restorations were made by trial : it is assumed from the conditions found in adult labyrinthodonts, that the table between the otic notches is essentially flat, and that the parts lateral to it slope downwards at an angle, which cannot, of course, be determined directly by measurement, and is to that extent arbitrary. Into the area so C 1450 N n 542 PALAEONTOLOGY, VOLUME 6 marked out the pattern of the bones can be inserted, making allowance for the fore- shortening of the orbit and of those bones which, like the squamosal, lie at an angle to the skull roof; each drawing was checked when possible by a projection on to the other plane. Lateral views can only be drawn when the squamosal, jugal, and maxilla are well preserved and their articulation evident. The fundamental assumption of the names used in zoological nomenclature is that each such name should imply one particular kind of animal, at all stages of growth, even if it suft'ers a great metamorphosis in its life history, like a butterfly. The difficulty arising in the case of labyrinthodonts, whose growth stages may already have been referred to a variety of genera, raises the purely practical problem of what such in- dividuals should now be called, when it may well be impossible to discover the name of the adult into which they grew. Large forms have been found at Niederhasslich, though they are very rare, but none are yet known in association with the branchiosaurs of Odernheim, or of Friedrichroda, though it is assumed by analogy that they also were large when adult; in the other case in which we have a long gradated series of individual amphibia, Archegosaiirus (H. von Meyer, 1857, pis. 8u-23), which ranges from a skull less than 2 cm. in length to one (incomplete) at least 19 cm. long, it has never been doubted that the growth of a single species is represented. It should be evident that any name given to a branchiosaur is to be regarded, not as a normal specific name, but as a handle for convenience of reference, and on the whole it seems to me that the generic term Branchiosaurus may well stand for any labyrinthodont larva from Niederhasslich, Odernheim, or Friedrichroda whose adult has not yet been identified ; and further, that those larval individuals which can be distinguished from the rest on features not related to growth may be given ‘specific’ rank in that genus. MATERIAL FROM NIEDERHASSLICH Credner completed his work on branchiosaurs by discussing the growth of the amphibian, the whole material being summarized in two magnificent plates (1886, pis. 16, 17), one reproducing a series of eleven skeletons, the other a series of sixteen skulls ranging from about 5 mm. to some 20 mm. in length, all drawn unrestored as they lie with the jaws spread out laterally. They are figured from the dorsal side only, but the drawings, like all Credner’s work, are excellent, and show a gradual relative lengthening of the postorbital part of the skull, which becomes a little narrower proportionately; this is associated with the rapid growth of the brain. Text-fig. 1 shows restorations of three individual skulls from Niederhasslich now before me, and of Credner’s Onchiodon (Sderocephalus) lobyriiithicus {\S93, pis. 30, 31). In this series a and b show the relative lengthening of the postorbital part of the skull, but in c the preorbital part has begun to grow disproportionately, in order to provide a mouth and jaws of sufficient size to meet the needs of an animal whose weight is in- creasing as a cube of a linear dimension. A significant feature is the meeting of pre- and post-frontals, excluding the frontal from the border of the orbit, which is evidently related to a reduction in the proportionate size of the eye; at the same time the jugal lengthens proportionately, and the quadratojugal, while retaining its old length, deepens. B and c show the nature of the lachrymal particularly well; in a it can scarcely be seen, as D. M. S. WATSON: GROWTH STAGES IN BR ANCHIOS AU RS 543 is the case in most of Credner’s stages. In b this bone extends along the orbital margin, with a double opening for the duct, which is essentially surrounded by bone, and may not reach the nostril; in c the bone is beginning to be excluded from the orbit, and the duct is an open groove throughout its length, entering the nostril. It may be noted that the pineal foramen in A is on the level of the hinder border of the orbit, in b it lies en- tirely behind the orbit, in c it is relatively farther back. This depends on the fact thai brain development takes place during the early stages of growth, and ceases quite soon, a point very well brought out in Credner’s series. There seems no doubt that a, b, and c are part of a growth series, and in this connexion it is of interest that the snout of an individual almost the same size as c, showing an identical pattern of ornamented bones, and similar lachrymal ducts, lies close to R. 201 1 (a) on the same slab. If, as has been suggested (Romer 1939) the final term of this series be Onchiodon it can then be seen how far further growth has altered the proportions of the cranial roof, d shows that the snout continues to elongate, the lachrymal in con- sequence losing its contact with the orbit, being separated from it by a short but quite definite suture between the prefrontal and jugal. The orbit lies relatively far back com- pared with c, and is proportionately smaller still; its lateral border is separated from the border of the skull by rather more than its own width, in other words the jugal is now extremely deep. The pineal foramen is relatively even farther back. The table, and therefore the braincase which lies beneath it, is now very narrow compared with the whole width of the skull at the same point. The squamosal and quadratojugal have greatly increased to form the characteristically deep cheek seen most typically in large skulls of Eryops. Thus it seems evident that the large amphibian from Niederhasslich and the bran- chiosaurs found there are all members of the same growth series, and may be called Onchiodon (Sclerocephalus) labyrinthiciis Geinitz, a procedure justified by the fact that Credner’s work is confirmed by specimens not known to him, and a good intermediate stage has now been found between the small larva and the adult. It is interesting (text-fig. 1) to compare the ornament of the dermal skull roof of c with that of d (remembering that the former was drawn from a squeeze of a mould, and the latter from the mould itself). In c the ornament consists almost entirely of a series of pits, with very rare ridges and grooves only recognizable in some bones, the jugal and quadratojugal, for instance, In d the areas covered with pits, the growing points, are very small in comparison with the surrounding radially arranged grooves and ridges, which represent the extension resulting from growth; in other words the ornament sug- gests the direction of growth. But there are nearly twice as many elements in the pattern of D (in the postorbital, for instance) as in the smaller form c, suggesting that the orna- ment is not enlarged commensurately with the bone, and that new elements are in- troduced into the pattern. It may be interesting to recall that in reptiles sutures in the skull have been known to close, presumably implying that the individual is old; in am- phibia, so far as I know, the sutures never close, implying that growth remains possible even at extreme old age. 544 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 1 {continued on opposite page). Oncbiodon {Sclerocephahis) labyrinthicus Geinitz. Recon- structions of four skulls from Niederhasslich. A-c, Dorsal aspect; e-g, lateral aspect; xf, surface ornament taken from squeezes and thus re- producing the bone, d, dorsal and h, lateral aspect x J, the mould itself is drawn. A, B.M.N.H., R. 2011, about the same size as Credner’s no. 5 (1886, pi. 17), the smallest available skull from which a restoration could be made; mould of external surface of head. Fragmentary vertebral column and shoulder girdle present, and a pelvis and hind legs. B, D.M.S.W., B. 92; isolated skull shown as a perfect impression of the external dorsal surface and right cheek. ‘Branchiosaurus ainblystomus' (Watson 1940, fig. 22) was founded on it, but the new restoration modifies the nose (which is damaged) and orbit, and is confirmed by Credner’s no. 15 (1886, pi. 17) and B.M.N.H., R. 2711, which are almost the same size. c, D.M.S.W., B. 91 ; isolated skull shown as a sharp impression of the external dorsal surface and left cheek; new restoration of specimen figured as Oncbiodon (Watson 1951, figs. 36, 37); nearly half as long again as Credner’s largest (1886, pi. 17). D, Oncbiodon (Sclerocepbabis) labyrinthicus reconstructed from Credner’s best specimen (1893, pi. 30, fig. 1); anterior part of skull restored from two other specimens (pi. 30, fig. 2, and pi. 31, fig. 1). Note that this skull is five times as big as c. E, B.M.N.H., R. 2011, teeth after Credner. F, D.M.S.W., B. 92, left side restored from right. Teeth restored from short length of interlocking upper and lower ones, shown in three-dimensional detail; position of suture between jugal and lach- rymal not certian. G, D.M.S.W., B. 91, teeth restored from nearly complete series. H, Oncbiodon (Sclerocephahis) labyrinthicus reconstructed from Credner (1893, pi. 30, figs. 1-2, and pi. 31, figs. 1-2); lower jaw and shoulder girdle of same individual have been taken into consideration in determining height of skull. MATERIAL FROM FRIEDRICHROD A Branchiosaurus flagrifer Whittard, 1930 Text-fig. 2a-c Holotype. Specimen from D.M.S.W. Collection, now numbered B. 48. Other specimen. D.M.S.W. Collection, B. 36a. D. M. S. WATSON: GROWTH STAGES IN BR ANCHIOS AURS 545 Branchiosaurus brachyrhyncJms sp. nov. Text-fig. 2d, e Holotype. B.M.N.H. specimen R. 5466/7. Other specimen. B.M.N.H. specimen R. 5469. Discussion of both species. The material from Friedrichroda, which was first found in the nineteen-twenties, contains a branchiosaur of which fifteen specimens were examined by Whittard (1930) and assigned to a new species, Branehiosaurus flagrifer. His drawings (figs. 1-3) represent the skull of my specimen B. 48 (the holotype) which has an in- complete vertebral column, and the vertebral column of B.M.N.H., R. 5466/7 which is a very complete individual retaining not only the head and body but also a very long tail, represented for the greater part of its length by a sharply defined, narrow skin impres- sion. I have made a new restoration of the skull of B. 48 (text-fig. 2b, c) which differs somewhat from Whittard’s, the differences arising, I think, from the fact that it is ex- tremely difficult to draw such material consisting of a mould in which the relief is very 546 PALAEONTOLOGY, VOLUME 6 shallow; colour differences hinder rather than help. I had the advantage of good photographs at a considerable magnification on which the outlines of the bones could be inked. In text-fig. 2 this new restoration is compared with reconstructions of the smallest TEXT-FIG. 2. Reconstructions of four skulls from Friedrichroda, a-d, x|, e, x|; each represents the external surface of the cranial roofing bones. The lachrymal is uncertain in a, b, and c; in d it occurs but its shape as shown depends on the borders of the surrounding bones. A-c, Branchiosaiinis flagrifer. A, D.M.S.W., B. 36n, the smallest specimen figured in the paper; note the large, elongated pineal foramen. Skull preserved as a mould, with no trace of ornament. Also present vertebral column complete from skull to traces of the pelvis, and part of a shoulder girdle and fore limb, b. New restoration of D.M.S.W., B. 48, holotype of Whittard (1930, fig. 1), from a very good mould of the cranial roof and lateral parts of the palate and lower jaw. A series of very small scattered projections on the supratemporal and parietals are of the nature of ornamentation; there are, perhaps, traces of the sclerotic ring. Vertebral column containing about twenty-six vertebrae, extending from skull to region of the pelvis, present in the specimen; also a set of ribs, a shoulder girdle, humerus, and femur in poor preservation, c, D.M.S.W., B. 48, restored lateral view. D-E, Branchiosaiinis brachyrhynchiis, sp. nov. D, B.M.N.H., R. 5466/7, holotype. Skull of specimen on which restoration of postcranial region of the type of B. flagrifer was founded (1930, fig. 3). The part and counterpart show the specimen split through longitudinally, the mould of the skull roof suggests that a very shallow ornamentation exists. R. 5467 shows traces of a sclerotic ring in both eyes. E, B.M.N.FI., R. 5469. A particularly difficult skull to interpret as the surface is not well preserved and the sutures are obscure; they have been determined by the meeting of radial striation of the bones, which have traces of ornament. The specimen has a vertebral column in articulation from the skull to the pelvis, damaged in one place; a set of ribs is present, fragments of tail, a shoulder girdle, and nearly complete fore and hind limbs. suitable Friedrichroda skull I could find (B. 36a), R. 5466/7, which is larger, and the largest available one (B.M.N.H., R. 5469) which is not founded on such good evidence as the others, but the extreme width across the snout — with an enormous lachrymal and small external nostril — and the position of orbits and otic notches is clear, though the D. M. S. WATSON: GROWTH STAGES IN BRANCHIOS AURS 547 pineal foramen cannot be seen with certainty. These skulls agree in general structure, and it will be seen that in B. 36, B. 48, and R. 5469 the pre- and post-frontals exclude the frontal from the border of the orbit, whilst in R. 5466 this bone enters quite largely into the orbital margin; R. 5466 also differs from B. 48 in that the tabular appears not to meet the squamosal, and the postorbital is widely separated from the parietal, which suggests that this skull is different from it. Moreover, comparison of the postcranial region of B. 48 with that of R. 5466 confirms the difference in the specimens, for in B. 48 the vertebrae from skull to pelvis number about twenty-six and are short antero- posteriorly and closely packed, and the ribs immediately behind the shoulder girdle are long and slightly curved, and then shorten to half-length at about the ninth vertebra; while in R. 5466 the comparable number of vertebrae is about twenty, and they lengthen behind the pectoral region so that the interval between successive ribs is long, and the ribs are all short and straight. This leads to the unfortunate conclusion that the type skull is united in a drawing with a body belonging to a different species. Examination of the postcranial regions shows that B. 36 appears to have close-packed vertebrae like the type (though its ribs are straight), and R. 5469 has about twenty vertebrae (also with straight ribs) like R. 5466/7. None of the differences between the columns could readily be accounted for by changes due to growth, therefore as indicated formally above at least two different species occur. B.M.N.H., R. 5466/7 may be taken as the type of Branchiosaurus brachyrhynchiis sp. nov. ; I also place R. 5469 into this species on the grounds that its vertebral column matches, and that so much of its skull as can be seen conforms to what might be expected as a result of further growth. The logical step would then be to investigate the sixteen other individuals from this locality to see if they could be put into one or other of these species. This would involve an elaborate procedure of interpreting enlarged photographs, which I do not propose to enter on, but preliminary inspection of the materials seems to show that the two sorts of vertebral column and skull do occur. The ornament on the dermal skull roof of B. 48 is not well shown, but B.M.N.H., R. 5286, which is nearly the same size, and probably the same species, shows definite ornament in this region. It is not like that of a normal large labyrinthodont, but is shal- low. The central pitted region is surrounded by poorly developed radial striae, and the units of the pattern are large compared with the size of the bone, both features depend- ing on the youth of the individual. MATERIAL FROM ODERNHEIM Branchiosaurus credneri (Bulman and Whittard) comb. nov. Text-fig. 3a, b, d, f Lectotype (here chosen). D.M.S.W. Collection, B. 40 (text-fig. 3f). Branchiosaurus Jevis (Bulman) comb. nov. Text-fig. 3c, E Holotype. D.M.S.W. Collection, B. 44/45 (text-fig. 3c). 548 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 3 {continued on opposite page). Branchiosaums credneri (Bulman and Whittard) comb, nov., and B. levis (Bulman) comb. nov. Dorsal and some lateral reconstructions of the series of branchiosaur skulls from Odernheim, all X 2. a and b drawn from the under surface of the cranial roofing bones; c, d, e, and f from the upper surface. A, D.M.S.W., B. 27, B. credneri. New reconstruction of specimen figured as B. amblystomus Credner by Bulman and Whittard ( 1 926, fig. 1 , i). Their ‘lachrymal’ is probably a prefrontal, the strip of bone at the upper border of the orbit being part of the frontal; there may be a lachrymal in the specimen, but if so it cannot be seen. The specimen has seven vertebrae in articulation with the skull, also a fore limb. B, D.M.S.W., B. 39, B. credneri. Reconstruction of skull on which (with B. 40) Micronielerpeton credneri Bulman and Whittard (1926, figs. 1 1-13 and pi. 4u) was founded. Much ofthe palate can be seen, also both rami of lower jaw, and anterior part of body including shoulder girdle, but not the fore limb. c, D.M.S.W., B. 44/45 (part and counterpart), B. levis. New reconstruction of the type specimen of Leptorophiis levis Bulman (1928, figs. 1-A). Traces of ornament of the skull table are shown in very low relief, which is probably genuine, and not due to poor preservation. Specimen shows an indistinct shoulder girdle, and a well-defined area of ventral scutes in chevron-shaped rows. An oval area with a definite margin in the centre of the right orbit of B. 45 is presumably a trace of the crystalline lens. D, D.M.S.W., B. 46, B. credneri. Reconstruction of specimen on which (together with B. Ala, and B.M.N.H., R. 5026) the drawing of Pelosaurus laticeps Credner (Bulman and Whittard 1926, fig. 14) was founded. Most of the bone has survived and shows well-preserved ornament. A lachrymal is present, though its shape is uncertain, but no sclerotic plates are to be seen. There are no intelligible postcranial parts. E, D.M.S.W., B. Ala, B. levis one of the specimens on which Pelosaurus laticeps Credner (Bulman and Whittard 1926, fig. 14) was founded. Skull represented by a mould in which the table is slightly disarticulated, but its relation to the rest of the dorsal surface is evident. Ornament well preserved on some bones, such as the postorbital and tabular, but incomplete in the rest of the skull. There are about a dozen vertebrae in articulation with it, but no girdles or limbs. E, D.M.S.W., B. 40, B. credneri, Lectotype. Reconstruction of skull on which (with B. 39) Micro- melerpeton credneri Bulman and Whittard (1926, figs. 11-13 and pi. Aa) was founded; the table now lies about in the middle of the vertebral column, having been separated from the frontals before burial. The skull, drawn from a squeeze, is represented by a beautifully preserved mould, which shows the ornament exceedingly well; the presence of lateral-line grooves on the supratemporals suggests that the individual was still aquatic. A column of thirty presacral vertebrae is in articulation with it; a complete series of ribs and both limb girdles are present, with a nearly complete fore limb, and partial hind limb, but little of the tail remains. D. M. S. WATSON; GROWTH STAGES IN BRANCH lOSAURS 549 G, D.M.S.W., B. 27, B. credneri, lateral view. Teeth are present, but are only represented by sections of crowns and are not reconstructed. H, D,M.S.W., B. 46, B. credneri, lateral view. The tooth row is conventionalized, but in number and size the teeth are essentially correct; the premaxillary teeth are taller and perhaps more massive than those of the rest of the series. I, D.M.S. W., B. Ala, B. levis, lateral view. The teeth are restored from an incomplete row, the premaxil- lary teeth being longer, and of larger diameter, than the bigger ones at the anterior end of the maxilla. Discussion of both species. The six skulls in text-fig. 3 were certainly found in the same quarry, in rock of exactly the same character. It cannot be shown that they lie in the same bedding plane, but there is no reason to suppose that there was much variation in time between them, in fact they must represent a fauna of a quite small lake, c, d, e, and F show the upper surface of the skull roof, A and b are from its under surface, the only part exposed ; few specimens show the palate, and only one (f) has adequate postcranial elements, in this specimen there are thirty presacral vertebrae, which is a very large number. In contrast to the original figures of Bulman and Whittard, which often in- cluded facts drawn from several individuals not necessarily the same size, each drawing in this figure represents one individual only. There are points of general similarity between the members of this series, for instance the tabular horns of c (Leptorophus levis of Bulman), d, and e have much in common, and Bulman (1928, p. 255) says ''Leptorophus levis may thus occupy an intermediate position between the Micromelerpeton-Pelosaurus type and the geologically younger species L. tener\ but only c and e can be said to differ from the rest in significant ways; in c and e the pre- and post-frontals meet in suture above the orbit, and the suture between the dermosupraoccipital and the parietal meets the admedian border of the 550 PALAEONTOLOGY, VOLUME 6 supratemporal nearly at its mid point, making the parietals relatively shorter, whereas in A, B, D, and f the pre- and post-frontals are widely separated, and the anterior sutures of the dermosupraoccipitals and tabulars are very nearly continuous, the parietals being relatively long compared with their width. It should be pointed out that in c all the bones of the skull roof are extremely thin, whilst in e they are of normal thickness, although this could well depend on some accident of preservation. Also b as drawn ditfers from A, D, and F in that the postorbital does not meet the anterior end of the supratemporal at all, whereas in the other individuals it does; but in this specimen (b) the supratemporal is shown only as an impression of its lower surface, and is somewhat confused by being pressed down on the pterygoid, itself shown merely as a mould. It may be noted that the pineal foramen (which is not accurately circular) in the younger stages of these two species is in an anterior position, but in the two later stages, E and F, it is relatively farther back, implying that in these animals, as in vertebrates in general, the brain is early developing, so that in later stages its growth is greatly ex- ceeded by that of its surroundings. And in f the quadrate is far back showing the general trend of change of shape with age found in nearly all labyrinthodont skulls. Thus it appears, as indicated formally above, that at least two species of labyrintho- donts were present in the pool in which the rocks were laid down. COMPARISON OF MATERIAL When a comparison is made between skulls from these three localities one difference seems clear: in the Friedrichroda series the eyes are relatively farther forward, and the noses wider and blunter, than in those from the two other places. In the Friedrichroda series also the meeting of pre- and post-frontals above the orbit has taken place in a skull only 5-6 mm. in length; in the Niederhasslich series this event occurs in a skull measuring between 20-0 mm. and 36-0 mm. in length; and in the Odernheim series one of the two species has achieved it in a skull measuring 15-6 mm., while the other has not yet done so in a skull about the same size as the largest Friedrichroda one. There is also, perhaps, a difference in the posterior part of the skull table in the older and more characteristic members of each series ; in those from Niederhasshch the tabular horns tend to turn inwards, in the Odernheim forms they tend to turn outwards, and the Friedrichroda skulls differ from both. Enough has been said about the differences between individual skulls of one locality, and those of one locality compared with another, to suggest that the group of labyrintho- donts of this age was more elaborate than has yet been recognized, and that larvae can be determined as well as adults. SURVIVAL OF BRANCHIAL ARCHES The Odernheim series used in this paper covers the point of growth at which gills are lost, thus presumably representing the transition from a larval to an adult life. The following table shows these specimens arranged in order of skull length, with the nature of the gill apparatus of each individual indicated. It appears that those with a length of 11-2 mm. or less have internal gills, implied by the presence of gill rakers, and also (in some cases) external gills shown as a carbonaceous film; at 11-75 mm. and above no D. M. S. WATSON; GROWTH STAGES IN BR ANCHIOSAURS 551 indication of gills can be found, though a ventral remnant may be present in one in- dividual measuring 12-7 mm. Only a single individual with a skull length of 15-6 mm. (and possibly one other) shows the presence of ventral armour in the form of rows of scales in a chevron pattern, perhaps a sign of the approach of maturity. Branchiosaurs from Odernheim Registered number Species Skidl length in mm. Total length to pelvis in mm. Presence of gills Ventral armour R. 6700 undetermined 61 7 none B. 34 8-1 28-0 internal B. 27 B. credneri 8-2 „ B. 26 undetermined 90 38-5 ? internal, and external B. 30 91 41-0 internal B. 32 9-1 41-0 99 B. 35 9-5 46-0 ? external B. 33 9-5 47-0 7 R. 5028/9 no 49-0 internal traces B. 25 1L2 internal and external none B. 29 11-2 45-0 internal B. 28 95 11-75 none B. 31 12-7 51-0 ? ventral remnant 99 B. 39 B. credneri 13-0 none B. 44/45 B. levis 15-6 present B. 46 B. credneri 18-0 B. 47a B. levis 23-8 ? none R. 5026 undetermined 24-0 84-0 none B. 40 B. credneri 28-5 102-5 B. 141 undetermined 28-5 103-0 99 My Niederhasslich material is too scanty to allow of a comparison, but Credner (1886, p. 586) gives a table showing that in his individuals gills are found in those with a skull length of 14-0 mm. and under, above which size they are never seen; but those larvae which have gills lack ventral armour which is found in all the larger individuals. The Friedrichroda specimens range from 5-2 mm. to 29-0 mm. in skull length, and of these only four (B. 50 and B 51, B.M.N.H., R. 5285a, b) show the branchial arches, and none shows any sign of ventral armour. DETAIL OF BRANCHIAL SKELETON The specimen B.M.N.H., R. 5285 (individual a), from Friedrichroda, shows the bran- chial arches exceptionally well. They lie in position undisturbed, extending back almost as far as the shoulder girdle, each with a paired row of gill rakers. The structure of the anterior part of the hyoid arch is, however, better shown in one of my Odernheim speci- mens (B. 30) than in any other known to me (text-fig. 4). This skull, 9T mm. in length, is 552 PALAEONTOLOGY, VOLUME 6 seen from below. Both rami of the lower jaw are present as impressions. The premaxillae have been crushed down so that their posterior points, which should have articulated with the nasals, now lie directed forward and the impression of their outer surfaces is seen. The parasphenoid is shown as bone on the lateral parts of its widened hinder end, and by the impression of its processus cultriformis, superimposed on the impression of P.Mx. TEXT-FIG. 4. D.M.S.W., B. 30, a branchiosaur from Odernheim, x 6. Skull and shoulder girdle seen from below, drawn unrestored. Specimen almost complete to the fourth caudal vertebra. Reference letters: At., atlas; Ba.Hy., basihyoid; Br.Ar. 1 & 2, ventral elements of branchial arches 1 and 2 attached to the basihyoid; Br.Ar. 1-4, series of gill rakers attached to the upper ends of branchial arches 1-4; C/., clavicle; Ex.Oc., exoccipital; Fr., frontal; Lat.Hy., laterohyoid; L.Jaw, lower jaw; Mx., maxilla; Na., nasal; P.Mx., premaxilla; Par., parietal; Par.Sp., parasphenoid; Pt., pterygoid; Pt.Fr., postfrontal; ScL, sclerotic plates; St., stapes. the visceral surface of the skull roof. The stapes — pierced by a foramen — is present on each side, and the exoccipitals, pressed down into the general plane of the palate, are small, rather featureless bones widely separated dorsally. The anterior part of the hyoid apparatus — a basihyoid and laterohyoids — is seen in the region of the parasphenoid, followed by two pairs of shreds of bone which are the ventral attachments of the first and second branchial arches. Behind the stapes lie the dorsal ends of the branchial arches, now perished and indicated only by the rows of attached gill rakers, which open out- wards lateral to the dorsal ends of the clavicles. The left orbit has impressions of D. M. S. WATSON: GROWTH STAGES IN BR ANCHIOSAURS 553 sclerotic plates, and also a grey area which appears to be some part of the eye itself. (B.M.N.H., R. 6700, skull length 6-1 mm., also shows in both orbits carbonaceous impressions of the eye.) The vertebral column, complete as far back as about the fourth caudal vertebra, is in articulation with the skull. Both fore limbs are present, but not the hands. The two femora are shown, and the right tibia and fibula in a somewhat frag- mentary condition. In this specimen the gill rakers suggest that open gill slits of very considerable length existed, and that something of the nature of internal gills must have occurred. Since the branchial skeleton was so well developed it presumably carried out functional move- ments. Acknowledgements. My thanks are due to the British Museum (Natural History) for the loan of much material, and to Professors M. Abercrombie and P. B. Medawar, of the Department of Zoology, University College London, for hospitality. I owe Miss J. Townend thanks for continuous help, and am grateful to the Royal Society for making this help possible. REFERENCES BULMAN, o. M. B. 1928. Additional notes on some branchiosaurs from Odernheim. Ann. Mag. Nat. Hist., ser. 10, 1, 250-5, pi. 12. and WHiTTARD, w. F. 1926. On Branchiosaimis and allied genera ( Amphibia). Proc. Zool. Soc. Land. 96, 2, 533-79, pi. 1-4. CREDNER, H. 1882. Die stegocephalen und saurier aus dem Rothliegenden des Plauen’schen Grundes bei Dresden. Zeitsclir. deutsch. Geol. Ges., Teil 3, 34, 213-37, 2 pi. 1885. Ibid., Teil 5, 37, 694-736, 2 pi. 1886. Ibid., Teil 6, 38, 576-632, 4 pi. 1893. Ibid., Teil 10, 45, 639-704, 3 pi. VON MEYER, H. 1857. ReptiUcn aus der Steinkohlen-Formation in Deutschland. Palaeontographica, 6, 59-179, pi. 8a-23. FARRINGTON, F. R. 1959. A note on the labyrinthodont middle ear. Ann. Mag. Nat. Hist., ser. 13, 2, 24-28. ROMER, A. s. 1939. Notes on branchiosaurs. Amer. J. Sci., 237, 748-61. STEEN, M. c. 1938. On the fossil amphibia from the Gas Coal of Nyfany and other deposits in Czechoslovakia. Proc. Zool. Soc. Lond. B, 108, 2, 205-83. WATSON, D. M. s. 1940. The origin of frogs. Trans. Roy. Soc. Edin. 60, 1, 195-231. 1951. Paleontology and Modern Biology. Yale University Press, 1-216. WHITTARD, w. F. 1930. The structure of Branchiosauriis flagrifer, sp. n., and further notes on Branchiosaurus amblystomus, Credner. Ann. Mag. Nat. Hist., ser. 10, 5, 500-13, 2 pi. D. M. S. WATSON Department of Zoology, Manuscript received 20 November 1962 University College, London OBSERVATIONS ON THE PALAEOECO LOG Y AND AMMONITE SEQUENCE OF THE FRODINGHAM IRONSTONE (LOWER JURASSIC) by A. HALLAM Abstract. The Frodingham Ironstone is shown from the ammonite evidence to range from the top of the semicostatiim Zone to the top of the obtusum Zone of the Lower Lias. The rich invertebrate fauna, which is con- sidered from a palaeoecological viewpoint, includes traces of sediment-burrowing and shell-boring organisms which are described for the first time. Certain lamellibranchs in a pyrite-bearing bed are shown to be probably dwarfed. Data for such features as shell size, orientation, disarticulation, and wear are given and relationships between the shells and their sedimentary matrix described. The faunal characteristics of the different types of ironstone are discussed and an attempt is made, using both the petrological and palaeontological evidence, to outline the salient features of the environments of deposition. It is concluded that the Frodingham Ironstone was deposited on a marine shoal area isolated from the land and that it records a series of alternating episodes during which the degree of water agitation and the Eh varied considerably. The Frodingham Ironstone, of Lower Liassic age, outcrops in north-west Lincolnshire and dips gently eastwards beneath a cover of younger Jurassic rocks. Economically of great importance, its position of outcrop has determined the site of the steel town of Scunthorpe. Though it has long been known that the ironstone bears a rich and varied marine invertebrate fauna, little palaeontological work has been done since the pioneer in- vestigations of Cross (1875). In this study, an attempt is made to establish the complete sequence of ammonites and so determine the precise zonal position of the ironstone, and to describe the whole macrofauna from a palaeoecological viewpoint. Finally, the palaeoecological data are used in conjunction with the petrological data to further knowledge of the depositional environments. The rock succession. The distribution and variations in thickness and lithological char- acter of the Frodingham Ironstone have been fully dealt with by Wilson (in Whitehead et al. 1952). It is lenticular in form, passing northwards and southwards into shales, and reaches a maximum thickness of 32 feet near Santon, 3 miles east of Scunthorpe. It has proved economically exploitable over an 8-mile belt running north to south through Scunthorpe, and borings have proved an extension at least as far as the River Ancholme, 5 miles to the east. The author has found it desirable, for the purposes of internal consistency, to measure his own sections along the belt of opencast workings from Coleby in the north to Yar- borough in the south. Coleby Mine (SE/906193) ft. in. 6. Thinly-bedded ironstone, rather shaly in parts, with many fragments of mudstone. Piaror/iync/iia and Camptonectes abundant. ....... 2 0 5. Hard shelly ironstone ........... 1 3 4. Variable, rather thinly-bedded shaly ironstone with many shale fragments; pisolitic band 1 foot from top ........... [Palaeontology, Vol. 6, Part 3, 1963, pp. 554-74.] 3 A. HALLAM; OBSERVATIONS ON PALAEOECOLOG Y 555 3. Hard, current-bedded ironstone with abundant Cardiiiia. Many lenses, fragments and discontinuous thin bands of mudstone or shale; pisoliths . . . . .90 2. Shaly oolite ............. 1 0 1. Shelly ironstone, close to base, as indicated by drainage pump . . . .12 18 1 seen The successions at Thealby and Roxby Mines, a little to the south (SE/905177) are generally similar to that at Coleby Mine. The top foot or so contains abundant rhyncho- nelloids and pectens while the bottom 4 feet, at Thealby Mine, to the base of the quarry and a drainage dyke, consist of varied shaly ironstone, sometimes reddish in colour, sometimes oolitic shale. A thin and fairly persistent, non-oolitic, blue-black shale band, IJ inches thick, occurs 64 feet above the base. Crosby Mine (SE/907133) Dark shales of siinpsoiii Subzone. 8. Oolitic shale, very rich in belemnites and pelecypods ...... 7. Current-bedded reddish to yellow-brown ironstone with fragments of mudstone; Piarorhynchia and Camptonectes abundant ....... 6. Shaly oolite with abundant Gryphaea. ........ 5. Shaly oolite with hard shelly bands ......... 4. Hard, shelly ironstone with two Cardiuia bands ...... 3. Current-bedded ironstone with thin shreds of shale ...... 2. Yellow-brown to reddish shelly ironstone with many shreds and fragments of shale or mudstone and pisoliths; Cardiuia common ....... 1 . Soft, reddish, decalcified ironstone with Diplocraterion. ..... Grey silty shale (4 inches) passing down into shale. ft. in. 0 6 4 6 1 3 2 0 3 0 4 6 8 0 4 0 27 9 The section at Conesby Mine (SE/892147) a mile to the north-west, is essentially similar. Yarborough Mine, north-end (SE/928116) ft. in. 1. Shelly oolitic shale ............ 0 3-| 6. Reddish, partly decalcified shelly ironstone with abundant Diplocraterion, Pholadomya, and Plenromya ............ 4 0 5. Blue-grey pyritic ironstone (Snap Band); (c) Blue pyritic shaly oolite with abundant Gryphaea . . . .12 ib) Brown ironstone ........... 0 9 {a) Blue shaly ironstone, sparsely oolitic ....... 0 5 4. Yellow-brown ironstone with abundant Cardiina ....... 8 6 3. Shaly ironstone with distorted ooliths ........ 0 9 2. Yellow-brown ironstone with Cardiuia ........ 7 6 1. Soft, reddish decalcified ironstone with shreds of shale and Diplocraterion . .30 26 4i Towards the southern end of Yarborough Mine beds 2 and 5c become softer and more shaly and shells become less abundant. The suggested correlation between the sections, based on both lithology and fauna, is given in text-fig. 1. 556 PALAEONTOLOGY, VOLUME 6 Ammonite sequence. Cross (1875) was the first to record ammonites from the ironstone. A number of arietitids were recorded including "Ammonites' semicostatum, scipionanum, compressaries, brooki, and conybeari. These species signified to him a horizon at about the borderline of Lias a and ^ of the German classification, a very reasonable assigna- tion for its time. Ussher (1890) took this assemblage to represent the semicostatum Zone. Specimens of a large "Coroniceras', supposedly allied to C. gmuendense, were recorded — 4 miles > 2'h miles COLE BY CROSBY YARBOROUGH TEXT-FIG 1. Suggested correlation of the main sections of the Frodingham Ironstone. The numbers refer to beds described in the text. by the Survey from the bottom 5 feet (Lamplugh et al. 1920). Arnioceras semicostatum was claimed to range from 4 to 5 feet from the base to the top. Arkell (1933) assigned the ironstone to the bucklandi (pars) and semicostatum Zones. Further collecting in recent years by P. E. Kent and V. Wilson has led to the recognition of obtusum Zone ammonites at the top of the succession. On this basis a sequence ranging from basal semicostatum Zone to top obtusum Zone is now claimed to be present (Swinnerton and Kent 1949; Whitehead et al. 1952). This at once raises a problem, for assuming the accuracy of the older identifications there is no published evidence for the presence of the upper semicostatum, turneri or lower obtusum Zones. The amplification and revision of the zonal sequence proposed below is the result of re-examination of museum material and further collecting. The Scunthorpe Museum possesses a number of large ammonites with smooth compressed outer whorls of trigonal cross-section, which were obtained from the Frodingham Ironstone late last century. It was presumably these that Cross would have called Amm. scipionanum and compressaries. Examination of a series of specimens of various sizes collected in situ has made it clear A. HALLAM: OBSERVATIONS ON PAL AEOECOLOGY 557 that they belong in fact to the genus Eparietites of the upper obtiiswn Zone. Other large specimens in the museum belong to the genus Asteroceras, with close affinities with Sowerby’s species A. stellare. In addition there are one or two large specimens of the highly distinctive turneri Zone form Microderoceras birchi, hitherto unrecorded from the ironstone. As only the top 10 to 12 feet of the ironstone were worked when all these ammonites were collected (H. E. Dudley, personal communication), a rough lower limit to their occurrence is fixed. A number of small specimens in the museum have been identified by L. F. Spath and include Asteroceras stellare, Aegasteroceras spp., Eparietites tenellus, and Epophioceras carinatum, all obtusum Zone forms. This last, highly evolute, species has a close resemblance to Coroniceras (Metophioceras) conybeari of the lower bucklandi Zone, and it was presumably this which led to Cross’s very pardonable, but highly misleading, misidentification of Amm. conybeari from the ironstone assemblage. The British Museum (Natural History) contains Frodingham Ironstone specimens (C 25133, C 50157) identified by Spath as Epammonites scunthorpensis Spath. These specimens have been examined by the author, to whom they seem not readily distin- guishable from Pararnioceras alcinoe, a species now attributed to the upper semicostatum Zone (Dean et al. 1961). C 25133 belongs to the Wright collection and was figured by Wright (1878-86, pi. 1, figs. 1-3) as Arietites bucklandi and interpreted by Donovan (1954) as A. scunthorpensis. Another specimen (C 50156), Caenisites cf. brooki, a turneri Zone form, was collected from near the base of the ironstone at Coleby Mine. Field collecting has yielded the following information. At Crosby, Conesby, and Yarborough Mines Arnioceras aff. semicostatum occurs fairly commonly in the bottom 7 or 8 feet and has been found at Coleby close to the top. A band with Aegasteroceras and large Eparietites occurs at the top of bed 7, Conesby and Crosby, in the top 2 to 3 feet at Thealby and Coleby and at the top of bed 6, Yarborough. Wilson (in Whitehead et al. 1952) records Asteroceras stellare at 3 feet and 8 to 10 feet below the top of the ironstone at Coleby. Farge specimens of Pararnioceras aff. alcinoe occur in a band 6 feet and 74 feet above the base at Thealby and Yarborough respectively. Bed 8, Crosby, contains abundant Gagaticeras gagateum together with Oxynoticeras simpsoni. Wilson records a specimen of the latter species 3 feet from, and presumably below, the top of the ironstone at Roxby Mine. No other specimen of Oxynoticeras has been found below the horizon of bed 8, Crosby. Text-fig. 2 represents an attempt to give in diagrammatic form a general idea of the ammonite sequence in the Frodingham Ironstone outcrops. The total assemblage indicates the presence of the upper semicostatum Zone {sauzeanum Subzone), the turneri Zone {brooki and birchi Subzones), and the obtusum Zone (stellare and denotatus Subzones). Bed 8, Crosby, contains an exclusively oxynotum Zone fauna and may be excluded from the ironstone proper. The top of the obtusum Zone is best taken above bed 7, that is directly above the band with Aegasteroceras and Eparietites. The presence of abundant Arnioceras in the lower part of the ironstone is misleading as a stratigraphical index since the genus is the longest ranging of the arietitids and different species hard to distinguish. The bucklandi and semicostatum Zones have been recognized in a limestone-shale facies below the ironstone (Dudley 1942). As noted above, the latter zone ranges up into the ironstone, and the top should be taken above the horizon of Pararnioceras but below that of Caenisites and Micro- deroceras. The absence of Promicroceras planicosta ( J. Sowerby) from the Frodingham o o C 145G 558 PALAEONTOLOGY, VOLUME 6 Ironstone is interesting, since this species is one of the commonest of the obtusum Zone ammonites. As it is particularly characteristic of the lower obtusum Zone, for which there is no evidence, it is possible that a minor non-sequence is present. Alternatively the environment was unfavourable for this form. Top Base (AEGASTEROCERAS 1 EPARIETITES EPOPHIOCERAS GACATICERAS OXYNOTICERAS ASTEROCERAS ? MICRODEROCERAS CAENISITES PARARNIOCERAS ARNIOCERAS common TEXT-FIG. 2. Stratigraphical distribution of ammonites in the Frodingham Ironstone. Caenisites and Microderoceras have not been collected in situ and so their exact location is uncertain. Arnioceras is only common in the lower beds, but ranges almost to the top of the succession. LITHOLOGICAL CHARACTERISTICS Petrology. The petrology of the Frodingham Ironstone was first described in detail by Hallimond (1925), but most of our present knowledge derives from the work of Davies and Dixie (1951). (See also petrological notes in Whitehead et al. 1952.) The dominant characteristic of the ironstone is the abundance of limonite ooliths with an unusual flattened ellipsoidal shape. A further interesting feature is the presence of several bands of limonitic pisoliths up to 10 mm. in diameter. It has been established that the ooliths were composed originally of chamosite, small quantities of which may still be found. Davies and Dixie distinguished four types of ironstone which are listed here with the standard Survey nomenclature in parentheses. Type A {sideritic limonite-chamosite oolite). The closely packed ooliths are variably replaced by siderite, small crystals of which occur in the matrix. Type B (sideritic chamosite mudstone). This non-oolitic rock is the least common type and occurs only as discontinuous thin bands and fragments. Type C (chamositic-sideritic limonite-chamosite oolite). The ooliths are variably re- placed by siderite and occur in a fine-grained matrix of chamosite clay and scattered crystals of siderite. Grains of silt-grade quartz are present locally. Type D (calcitic limonite oolite). The ooliths are set in a matrix of coarsely crystalline calcite. The rock contains detrital fragments of other ironstone types and is held to signify mechanical disintegration of these ironstones during turbulent conditions. A. HALLAM: OBSERVATIONS ON PALAEOECOLOG Y 559 The upper boundary of the ironstone is sharply defined but the rock passes downwards more gradually into thin limestones and shales, with ooliths diminishing gradually in number and chamosite and siderite being progressively replaced by ordinary clay and calcite. Davies and Dixie gave diagrams illustrating the lateral variation of the different ironstone types. In the south of the ironstone field the bed becomes unworkable owing to the incoming of clay and detrital quartz in several bands. Quartz is also commoner at the northern end of the field. The following supplementary observations are the result of field and thin section ex- amination by the author. The ironstone types B and D are very distinct, the latter being characteristically developed as a hard yellow-brown rock often exhibiting false bedding. The softer, red- dish-purple to blue-grey types A and C tend to grade into one another. Type C is the more argillaceous, and becomes harder and reddish with increase in the proportion of interstitial siderite. Type D often contains localized patches of mudstone with indeter- minate boundaries, in addition to obviously derived fragments. Quartz silt is rare in this type, whereas it is variably common to uncommon in the other ironstones. Broken ooliths (Davies and Dixie 1951) are somewhat commoner than in the other ironstones. The hardness of type B varies with the siderite content ; when this is low the rock has the characteristics of a stiff clay. In section this rock is seen to be remarkably free of shells, but, unlike the other ironstones, minute shreds of bituminous matter aligned roughly parallel to the bedding are often present. Typical relationships of B and D ironstone in situ are illustrated from Crosby Mine in text-fig. 3. Text-fig. 3h shows how a seam of B ironstone may be split into two and both 3a and 3b illustrate seams which either taper to nothing in a few feet or are sharply broken across and locally displaced. There are two distinctive bands in the Yarborough section which call for special comment. Bed 3 in the southern part of the section is a C ironstone containing pale-brown chamo- site distorted ooliths or spastoliths (Rastall and Hemingway 1939) elongated parallel to the bedding, in a matrix of green chamosite clay. These ooliths, similar to those figured by Davies and Dixie (fig. vi 4), have characteristics, such as hooked junctions, closely resembling those described by Carozzi (1961), which are attributed to plastic deformation by reciprocal impact due to water agitation. Some have been so shredded out parallel to the bedding that sediment compaetion has probably also played a part in the distortion. Bed 5, known as the ‘Snap Band’, is rejected as ore beeause of its high sulphur content, which ranges up to 4-5 per cent. As recognized by Hallimond (1925), this is due to the abundance of scattered quantities of pyrite, which has locally replaced shells in bed 5c, a C ironstone, where it is commonest. Hallimond also observed that at this horizon ( = 5c) the ooliths have been dissolved out leaving oolicasts (Pettijohn 1957). Some outer rims still possess a little chamosite, however, whilst minute carbonate crystals may partly fill the interiors. A further feature of bed 5c, hitherto unrecorded, is the pres- ence of abundant light-brown phosphatic nodules with a concentric structure, up to 10 mm. in diameter. Diagenesis. It is convenient at this stage to outline briefly the main diagenetic changes which have affected the ironstone, with a view to establishing their relevance to the conditions of deposition. 560 PALAEONTOLOGY, VOLUME 6 As was recognized by Hallimond, the presence of limonite ooliths in a chamositic matrix implies contemporaneous oxidation on the sea floor, as do those ooliths showing alternations of chamosite and limonite laminae. The siderite is mostly, if not entirely, early diagenetic in origin and was presumably precipitated in slightly reducing alkaUne solutions in which an abundance of iron precluded the formation of calcite. In con- o o o o o o o o o o o o (°) ooo O O C o OOOOOO o o o o o o o o o o o o o o ooooo o o I Foot TEXT-FIG. 3. Diagrams illustrating relationships of oolitic ironstone and non-oolitic chamosite mudstone (black) at Crosby Mine. sequence, its presence has little direct bearing on the conditions of deposition and the distinction between A and C ironstones is not likely to be directly relevant to these conditions. Pyrite is abundant in certain clayey bands such as the Snap Band. It exhibits replace- ment of shells and matrix and is therefore diagenetic in origin. Calcite is essentially confined to shells and to the matrix of D ironstones. In the A and C ironstones thin shells have often been dissolved, as indicated by the presence of moulds. The drusy calcite seems to belong to a comparatively late diagenetic stage, having crystallized in cavities where the iron content of the interstitial waters had A. HALLAM; OBSERVATIONS ON PALAEOECOLOG Y 561 diminished. In some poor ironstones, as at the base of the succession, fine-grained calcite may replace siderite as an early diagenetic mineral. Shells in A and C ironstones have locally been partly or wholly replaced by chamosite at, presumably, an early stage of diagenesis. Tab c = common AMMONITES Aegasteroceras sagittariiim ( Blake) c Arnioceras aff. (Young and Bird ) c Asterocems stellare o Caenisites cf. brooki (J. Sowerby) o Eparietites denotatus c E. tenelliis (Simpson) o Epophioceras aft', carinatiim Spath o Oxynoticeras simpsoiu (Simpson) o Microderoceras birchi (J. Sowerby) o Paraniioceras aflf. alcinoe (Reynes) c Sidciferites sp. o NAUTILOIDS Nautilus striatus J. Sowerby o BELEMNITES Naunobelus brevis (Blainville) c LAMEEEI BRANCHS Anisomyaria Camptonectes lohbergensis (Emerson) c Chlamys cf. textoria (Schlotheim) c Entolium lunare (Roemer) c Gryphaea aff. arcuata c Lima gigautea (J. Sowerby) c L. succincta Schlotheim o ? Liostrea sp. o Meleagriiiella sp. o Modiolus sp. o Myoconcha sp. o Pinna hartmanni Zieten o Plicatula sp. o Terquemia arietis (Quenstedt) o LE 1 o = occurs Eulamellibranchia Astarte obsoleta Dunker c Cardinia concinna c C. Usteri (J. Sowerby) o Nippopodium ponderosum J. Sowerby o Plioladoniya ambigiia (J. Sowerby) c Pleuromya cf. striatula (Oppel) c Tutcheria cingulata (Goldfuss) o GASTROPODS Amberleya sp. o Pleurotomaria anglica (J. Sowerby) o Small forms seen in thin section c BRACHIOPODS Piarorhynchia juvenis (Quenstedt) c Spiriferina walcotti (J. Sowerby) o Zeilleria cf. perforata (Piette) o CRINOIDS Ossicles seen in thin section c ECHINOIDS Spines seen in section o EORAMINIFERA Seen in thin section c TRACE EOSSIES AND BORINGS ? Sponges c RhizocoraUium jenense Zenker o Diplocraterion paraUelum (Torell) c Zapfella pattei Saint-Seine c Algae c FAUNA Composition. The rich invertebrate fauna of the Frodingham Ironstone is dominated by lamellibranchs, as indicated in the faunal list of Table 1, in which the fossils are grouped taxonomically. It is more appropriate in this study, however, to discuss the fauna in terms of its broad ecologieal subdivisions. The organisms may accordingly be grouped as endobionts, more or less permanently occupying, during life, positions within the 562 PALAEONTOLOGY, VOLUME 6 sediment, epibiouts, living on the surface, and nekton, swimming in the waters above (Schafer 1956, Hallam 1960). The endobionts include two groups of lamellibranchs and soft-bodied forms repre- sented by trace fossils. The myas {Pholadomya and Pleuromya) occupied permanent burrows, while such unspecialized eulamellibranchs as Cardinia, Hippopodium, Astarte, and Tutcheria were very probably, by modern analogies, forms which spent most of their time occupying temporary positions just below the surface, but capable of slow movement on or through the sediment if disturbed. TEXT-FIG. 4. Diagrammatic representation of Diplocraterion parallelian (a and b) viewed normal to the bedding and Rhizocoralliiim jenense parallel to the bedding, to illustrate the surface sculpture. X f . The trace fossils include species of the ichnogenera Diplocraterion and Rhizocoralliiim. The commoner species, D. parallelum (Torell), is represented by small, protrusive, plugged U-tubes with a straight axis and vertical to the bedding, linked by a median laminated zone or Spreite. It resembles those recently described from the Blue Lias as Rhizocoralliiim (Hallam 1960). They are most conspicuous where marked out by differences in the sediment. Thus, in the bed directly below the base of the ironstone at Crosby Mine, oolite passes down into shale (text-fig. 4a) while in the overlying iron- stone clay-filled U-tubes lie isolated in the midst of oolite (text-fig. 4b). The absence of an overlying band of clay or mudstone implies a certain amount of erosion subsequent to the burrowing. In contrast to these examples, specimens in bed 6, Yarborough Mine, are preserved as oolite within oolite. Rhizocoralliiim differs from Diplocraterion in its greater size, its orientation more or less parallel to the bedding and in possessing a curved axis. Plugged U-tubes assignable to R. jenense Zenker, with an indistinct Spreite, occur a few feet from the top of the Yarborough succession. They possess discontinuous longitudinal grooves and bifurcat- A. HALLAM: OBSERVATIONS ON PALAEOECOLOG Y 563 ing thread-like ridges on their surface (text-fig. 4c). This species, not previously recorded in Britain, is much commoner and better preserved in the Cleveland Ironstone near Staithes, Yorkshire. The surface sculpture has been convincingly compared by Weigelt (1929) with the scratch-markings produced by certain burrowing crabs. Seilacher (1952) has interpreted Diplocraterion and Rhizocorallium as the infilled burrows of sediment eaters. The epibionts may be classified as vagile and sessile forms. The vagile epibionts include echinoids, known only from scattered spines, and, quite probably, small gastropods and foraminifera, common enough in thin section but virtually impossible to extract because of the hardness of the matrix. An idea of the probable composition of the foraminiferal fauna was obtained by extraction from the shale directly overlying the ironstone at Crosby (bed 8) of species of Fwndiciilaria, Dentalina, Lingulina, Marginu- lina, Nodosaria, and Plamdaria. The much larger group of sessile epibionts includes forms which were attached during life to the surface by means of byssi or pedicles (anisomyarian lamellibranchs and brachiopods) or rooted (crinoids). A small number of lamellibranchs (oysters, Plicatida) cemented themselves to Gryphaea shells lying on the sediment surface. Gryphaea itself was sessile but free lying, while the unattached pectens and limas probably had limited powers of swimming. (The Frodingham Ironstone Gryphaea resemble G. arcuata of the older Sinemurian but differ in being on the whole less tightly coiled and having a greater ratio of breadth to length.) Finally there are traces of a number of shell-boring organisms. Many Gryphaea in bed 6, Crosby, contain small cavities with structures indistinguishable from borings of the cirripede Zapfella, as described by Saint-Seine (1954). The cavities as exposed on the shell surface are elongate and sometimes slightly arcuate, rounded at one end and tapering to a point (text-fig. 5a), with dimensions averaging T5xO-5 mm. Where the outer layers of the shells have been worn away the cavities are seen to expand downwards into larger, sac-like structures, ranging up to 5x1 mm., which may be filled with sediment. Some at least of the cirripedes seem to have bored into the shells after death of the host, since they occur on the median part of the left valve, which dur- ing life must have lain adjacent to if not actually embedded in sediment. A second type of boring is common in many Gryphaea shells (text-fig. 5fl). This consists of minute, superficial perforations ranging up to 0-5 mm. in diameter, variably normal or oblique to the shell surface. In size and shape these perforations bear some resemblance to those produced by clionid sponges (Topsent 1887), but this assignation is tentative. Yet a third type is represented by borings that closely resemble those of certain algae (Duncan 1876), very common in and seemingly confined to shells in type D ironstones. These constitute a ramifying network of fine tubes averaging 5 p in diameter and infilled by finely crystalline siderite (text-fig. 5b). The straight or slightly sinuous borings are aligned mostly normal to the shell margin and are occasionally seen to branch. There is no apparent relationship to the structure of the shells, even when they have been reerystallized. They are apparently confined to lamellibranch shells (cf. Cayeux 1914). Algal borings have been recorded by Cayeux from French Liassic ironstones, and it was presumably such structures as these that Hallimond (1925) referred to in passing as ‘algues perforantes’. The last group, the nekton, consists of varied cephalopods which probably had but transient contact with the sea bottom during life. 564 PALAEONTOLOGY, VOLUME 6 Distribution. Many of the beds in the ironstone contain abundant fossils, with Gryphaea and Cardinia the most conspicuous. Individual species, however, vary considerably in abundance as the succession is ascended, first one and then another coming into prominence. Certain thin bands especially rich in certain species may indeed be useful TEXT-FIG. 5. a, Cirripede and ?sponge borings in a Gryphaea shell from bed 6, Crosby, x 3. b. Algal borings in a lamellibranch shell, Crosby, x 100. for local correlation. For example, Gryphaea is very common in bed 6, Crosby, and bed 5c, Yarborough. Cardinia is especially common in bed 3, Coleby, and beds 2 and 4, Yarborough and Crosby. Piarorhynchia is abundant in the top 4 feet of all sections but uncommon below this level. Camptonectes is very abundant also in the top 2 feet of all sections. Certain ammonites such as Pararnioceras, Eparietites, and Aegasteroceras are largely confined to thin bands, whose position in the succession has already been noted. Laterally, over a small distance, shells tend to be uniformly distributed, the only (rather dubious) suggestion of clustering being with Piarorhynchia in bed 7, Crosby. A. HALLAM: OBSERVATIONS ON PALAEOECOLOGY 565 Size. A conspicuous feature of the ironstone is the large size of many of the ammonites and pelecypods, the largest being ammonites of maximum diameter 620 mm. {Eparietites) and 300 mm. (Asterocems). What appears to be a genuine case of dwarfing occurs in bed 5f, Yarborough, the upper part of the Snap Band, which is crowded with lamellibranchs in the northern part of the section. The most amenable fossils for analysis are specimens of Cardinia concinna Growth ri nqs TEXT-FIG. 6. Graphical analysis of the spacing of growth-rings in shells of Cardinia concinna. The broken lines represent specimens from the Snap Band. Length of each specimen given in mm. and Pholadomya ambigua. The Cardinia shells have distinct growth striae, which are numerous, fine and closely spaced, and growth-rings marked by acute depressions on the shell surface, which are much fewer in number, regularly spaced, and far more prominent. The maximum size of four specimens of Cardinia collected from the bed in question is 35 mm., as opposed to 125 mm. for normal representatives of the species elsewhere in the ironstone. Full measurement of growth-rings was possible on two of these specimens and a graph was plotted (text-fig. 6) to compare the spacing of these rings with those of a sample of normal Cardinia collected near the base of the ironstone. It is seen that the shells fall on different curves, with those from the Snap Band having much more closely spaced growth rings, indicating that they are not merely juveniles. A 566 PALAEONTOLOGY, VOLUME 6 similar result was obtained for internal moulds of Pholadomya, comprising small forms from the Snap Band and large forms from bed 6, Yarborough (text-fig. 7). Although the numbers of specimens from the Snap Band are small, this does not affect the validity of the interpretation, as a correlation of size and growth rate has been demonstrated. A further possible case of dwarfing is suggested by collection from bed 5c of three 65 60 55 50 45 E E 40 o ju E 35 E 2 30 ! Q 20 15 lO 5 O i 2 3 4 5 6 7 8 9 iO i1 12 13 14 15 \6 IT 18 19 2 0 21 22 23 2:4 25 2 6 21 28 G row t h r i nq s TEXT-FIG. 7. Graphical analysis of the spacing of growth-rings in specimens of Pholadomya ambigita. Conventions as in text-fig. 4. small specimens of Lima gigantea (maximum length 16 mm.), a species which elsewhere in the ironstone attains ten times this size. It is very unusual to find specimens as small as 16 mm. outside the Snap Band. Unfortunately, proof of dwarfing is not possible in the absence of pronounced growth-rings. There are a number of other small shells including Astarte obsoleta, Camptonectes lohbergeusis, Pseudolimea pectinoides, and Amberleya sp., but as these are naturally small species there is no reason to presume dwarfing. By far the commonest fossil is Gryphaea afif. arcuata. Though the shells are mostly small, the length of the right valve ranges up to 45 mm., which is not appreciably smaller than the A. HALLAM: OBSERVATIONS ON PAL AEO ECOLOGY 567 maximum of 65 mm. for the ironstone as a whole. Growth-rings are too irregular for reliable measurement but from the size distribution and comparisons of tightness of coiling it seems likely that at least the bulk of these oysters are juveniles. Shell orientation and disarticulation. Large ammonites always lie parallel to the bedding but small specimens have a more irregular orientation. This situation compares with that in the Blue Lias of Dorset and is explicable as a purely mechanical phenomenon w r CA Bed 3, Coleby v_y Cardinia 313 34 37 6 Bed 6, Crosby Gryphaea (left valve 58 82 50 10 where disarticulated) Bed 7, Crosby Camptonectes 173 61 Chlamys 12 8 Gryphaea (left valve) 12 15 Pseudolimea 17 13 Bed 8, Crosby Camptonectes 7 7 Cardinia 1 4 3 Chlamys 23 24 14 Gryphaea (left valve) 34 75 42 „ (right valve) „ (articulated) 44 41 25 3 Pholadomya Pseudolimea 2 5 3 TEXT-FIG. 8. Quantitative estimate of disarticulation and orientation of lamellibranch shells from several ironstone beds. The three columns on the left represent disarticulated valves in various orientations, the two on the right articulated shells. (Hallam 1960). Belemnites lie mostly parallel or slightly oblique to the bedding. Their orientation in the planes of bedding was not determinable. An attempt has been made in text-fig. 8 to give some quantitative expression to the orientation and degree of disarticulation of the commoner lamellibranch shells in typical samples of the various ironstones. Orientation of the shells and disarticulated valves is indicated diagrammatically. In ironstones of type D, where Cardinia is very common, it is readily apparent from general observation that there is a high degree of disarticulation and a high predomin- ance of valves convex upwards. Bed 3, Coleby, is a typical example, studied in vertical section. Bed 7, Crosby, is a variable A-D ironstone. There is again a preponderance of valves convex upwards, though less pronounced than in bed 3, Coleby. Disarticulation is complete. Bed 6, Crosby, is a rather clayey type C ironstone containing abundant large gryphaeas. There is a high degree of disarticulation, with a predominance of incurved left valves convex downwards. As Zeuner (1933Z?) observed, this orientation is an ex- ception to the general rule among lamellibranchs and is a consequence of the unusual shape of the left valve and its reaction to bottom currents. If such a valve lies on its side. 568 PALAEONTOLOGY, VOLUME 6 a regular water current moving across it will cause a shallow pit to be excavated in the sediment around its upstream side so that the valve will eventually come to be embedded in a horizontal position, predominantly convex downwards (Zeuner 1933n). Though this orientation is similar to the probable life position there is every indication that the shells have been disturbed after death. The degree of disarticulation is high and oc- casionally left valves are found locked in each other. In one case the valve of an oyster was found cemented on the underside of a left valve convex downwards, implying disturbance subsequent to the growth of the incrusting oyster. Bed 8, Crosby, a shale with limonite ooliths, contains Pholadomya in growth position together with numerous disarticulated epibiont lamellibranchs. Neither the small pectens nor the right valves of Gryphaea show any preferred orientation in the horizontal position, a natural consequence of the almost planar shape. In contrast the left valves of Gryphaea have a pronounced tendency to lie convex downwards. The results for these two argillaceous beds compare well with those obtained by Zeuner (19336) for left valves of Gryphaea in clayey sediment. Comparison is made here in terms of percentages. Convex down Convex up oblique Bed 6, Crosby 43 31 26 Bed 8, „ 50 22 28 Zeuner, 19336 43 23 34 Zeuner found that in fine sandstone there is a much larger proportion (72 per cent.) of valves convex downwards, a result attributed to the stronger action of currents embed- ding shells in the substratum. In contrast to the highly disarticulated epibiont and ‘conditionally vagile’ endobionts {bedingt vagile Endobionten of Schafer, 1956), such as Cardinia, the deep-burrowing myas are invariably found undisturbed in growth positions with the anterior ends pointing obliquely downwards at a high angle to the bedding. Disarticulation of crinoids (into small ossicles) seems complete, whereas in the brachiopods it is negligible. Shell wear and fragmentation. Gryphaea exhibiting considerable signs of wear are present in bed 5c, Yarborough, and bed 8, Crosby. Many of the smaller shells are worn smooth, with a total loss of growth ridges. Sharp edges are truncated and, in extreme cases, left valves resemble subrounded pebbles. A few belemnites in bed 8, Crosby, collected within shale, have been broken near the apex and the jagged edges worn smooth. Generally speaking, evidence of wear among other shells in the ironstone is not conspicuous. Only a rough qualitative estimate is possible of the degree of shell fragmentation. Broadly, fragmentation of large, thick shells is slight, in contrast to small, thin shells. Fragmentation, highest in the D ironstones, is never extreme and no shelly layers com- posed entirely of finely comminuted debris have been seen. Shell diagenesis. Original lamellar structure is preserved in brachiopods and in certain lamellibranchs such as Lima gigantea and many Gryphaea. The thick prismatic layer of Pinna hartmanni is unaltered. All this suggests original calcitic shells. Extensive re- crystallization of CaCOg has taken place in some Cardinia and Gryphaea, while the A. HALLAM: OBSERVATIONS ON PALAEOECOLOG Y 569 aragonite of ammonite shells has recrystallised to calcite. In many clayey ironstones, a lot of solution of thin ammonite and lamellibranch shells has been effected; myas are invariably represented by moulds. It is possible that the myas possessed aragonitic shells at least in part, but it is not necessary to assume this in accounting for their in- variable absence, since Hecht (1933) has demonstrated that thin, large shells, such as were possessed by the myas, are readily dissolved within a muddy sediment in the presence of decaying organic matter. Chamosite has replaced shells quite extensively in I I Calcite shell liiMI Mudstone Drusy calcite 1S53 Oolite TEXT-FIG. 9. Shell-matrix relationships in the Frodingham Ironstone, as drawn from actual specimens. some A and C ironstones and chamositized shells are to be found not uncommonly in D ironstones. Local replacement by siderite and pyrite is recognizable. Shell-matrix relationships. Ammonites have been examined by means of polished sagittal and transverse surfaces. The body chambers contain sediment similar to the external matrix. The sediment of the inner chambers of Aegasteroceras in type A ironstone at Yarborough Mine tends to be rather less oolitic than the external matrix. The inner chambers in some D ironstones are filled with variable amounts of oolite, mudstone, sometimes containing scattered ooliths, and drusy calcite, which may only partly fill the cavities. In some large specimens a few inner chambers may be markedly oolitic compared with their neighbours, suggesting that local fracture of the shell wall has allowed the influx of coarse material from the enveloping sediment. A particularly interesting matrix structure is exhibited by a specimen of Epophioceras from Thealby (text-fig. 9a). Considering the sectioned chambers on the right, the lowest parts are filled with oolite, which is overlain in two of them by mudstone. The upper part of the chambers is filled by drusy calcite. Small quantities of mudstone lie directly beneath the shell. 570 PALAEONTOLOGY, VOLUME 6 Some interesting relationships are shown by cardinias in the D ironstones. Articulated shells variably contain oolite, sideritic chamosite mudstone with scattered ooliths, and drusy calcite. Where both mudstone and oolite are present the latter may occupy the lower part of the shell interior (text-fig. 9b, cf. text-fig. 9a). In the same type of ironstone small pockets of mudstone frequently occupy positions beneath the ‘canopies’ repre- sented by Cardinia and Gryphaea valves convex upwards (text-fig. 9c, cf. text-fig. 9a). Sometimes the canopy is directly underlain by drusy calcite. Text-fig. 9d illustrates a sectioned specimen of an Entolium valve resting on a Cardinia valve, with mudstone beneath the former resting on oolite. No pockets of mudstone are found, however, in association with valves convex downwards. Many small gastropods and other microfossils are seen in thin sections of oolites to have interiors filled with mudstone or drusy calcite. RELATIONSHIP BETWEEN LITHOLOGY AND FAUNA The sideritic mudstones of type B are clearly distinct in that there is virtually no trace of organic life. In contrast the clayey oolites (types A and C), which are not readily distinguishable from a faunal point of view, abound in fossils. Endobionts including myas in growth positions and sediment eaters (Diplocraterion, Rhizocorallium) are an important element of the fauna which is frequently dominated, however, by varied epibionts and nekton. The high degree of lamellibranch and crinoid disarticulation and the evidence of orientation, fragmentation, &c., implies a fair amount of disturbance of shells after death, but there is no reason to suppose that the organisms in question did not live in the environment reflected in the sediment. The faunal peculiarities of bed 5c, Yarborough, can be correlated with two distinctive sedimentary features, the high sulphur content and the presence of phosphatic nodules. Analyses of two samples of shelly rock containing the dwarfed lamellibranchs gave sulphur contents of 0-84 and 0-93 per cent. These analyses are well below the bulk sulphur content of the Snap Band as a whole (Whitehead et al. 1952), due most probably to the absence of large clots of pyrite from the samples. They are nevertheless appreciably higher than most values for the ironstone. The D ironstones contain most of the fossils present in the clayey ironstones with the notable exception of myas, Diplocraterion and Rhizocorallium. On the other hand, Entolium is commoner, and traces of boring algae are abundant. Text-fig. 8 shows that the degree of disarticulation is no higher than in the A and C ironstones whereas the percentage of Cardinia valves oriented convex upwards is conspicuously higher than in pelecypods from these latter types. It is apparent from the experimental work of John- son (1957) that valves lying in the more stable position convex upwards (excepting Gryphaea) are more readily buried by the action of bottom currents on the sediment. Such a pronounced preferred orientation in the D ironstones could be a consequence, either of the action of strong currents, or of prolonged exposure to moderate currents. In view of the evidence of erosion in the D ironstones, noted by Davies and Dixie and in this paper, the important question arises to what extent the fauna has been derived from the clayey ironstones. Most of the organisms represented as fossils evidently flourished in muddy conditions, as they abound in many Lower Liassic shales, and it is A. HALLAM: OBSERVATIONS ON PALAEOECOLOG Y 571 very doubtful whether they could have tolerated the turbulent conditions implied by oolite in which mud is largely absent. Certainly the scattered chamositized shells would seem to be exotic to this environment. It is here that a consideration of the shell-matrix relationships becomes critical to an understanding of the depositional environment. The presence of pockets of mud ad- jacent to the inner side of valves convex upwards (text-fig. 9c) could be held to indicate removal by currents of a mud-filled shell from its original environment, and subsequent disarticulation and loss of most of the mud filling. This interpretation would fail, how- ever, to explain the absence of mud on the inner side of valves concave upwards, nor could it account for the presence in some specimens (text-fig. 9a, b) of mud overlying type D oolite, which must have entered the empty shell first, or of the pockets of mud underlying the outer side of the ammonite of text-fig 9a. The one feature these mud pockets have in common is that they lie in positions which would have been protected by shells on, or just below, the sediment surface from the agitation of bottom waters. The mud seems, in fact, to have been deposited in situ and protected from subsequent winnowing. No macro-fossils have been discovered with interiors of ordinary A or C ironstone. It follows from these facts that deposition of the D ironstones may have been ac- complished in several phases, periods of strong water agitation, during which mud was removed, alternating with quieter periods when fine particles were enabled to settle once more. Rather than the ooliths having been transported far, therefore, much of the D ironstone could be residual in origin, having been derived from unconsolidated sedi- ment. While many of the fossils doubtless are exotic to the turbulent environment in this sense, others could perhaps have lived during the quieter phases. The presence of scattered ironstone pebbles does signify, nevertheless, some erosion of consolidated rock. ENVIRONMENTS OE DEPOSITION The absence of fossils and the presence of shreds of bituminous matter aligned parallel to the bedding suggest that during deposition of the sideritic mudstone (type B) the bottom waters were anaerobic and therefore unfavourable for bottom life. If this were not so, burrowers would have destroyed all trace of lamination and the bituminous matter would have been lost through oxidation. It is possible, therefore, that the siderite was precipitated syngenetically. The clayey oolites, however, contain a rich bottom fauna, lack bituminous matter and any trace of microlamination, and bear evidence of moderate water agitation, all signifying well-oxygenated bottom waters. Bottom disturbance was sufficient to move the shells of epibionts and ‘conditionally vagile’ lamellibranch endobionts, but too weak to winnow away mud or erode deep-burrowing myas. Faunal evidence could well have a bearing on the controversy about whether cha- mosite ooliths were formed in situ within the sediment (Pulfrey 1933; Caillere and Kraut 1954) or by the rolling action of gentle currents on the sea floor (Whitehead et al. 1952; Dunham 1960); the latter implies that chamosite would remain stable in an oxidizing environment. Those ooliths with alternating laminae of chamosite and limonite signify either a regular shuttling to and fro on the sediment surface between regions of differing 572 PALAEONTOLOGY, VOLUME 6 oxidation-reduction potential or periodic exposure and reburial. The author finds the latter alternative more plausible, but conclusive evidence is lacking as yet. In regard to bed 5c, Yarborough, though the presence of diagenetic pyrite does not necessarily signify anaerobic conditions in the bottom waters during deposition, a high pyrite content very probably implies the existence of fair quantities of H2S in the original sediment. As deficiency of oxygen is known to be a potent factor inhibiting growth the dwarfing of the endobionts Cardinia and PboJadomya may reasonably be attributed to this factor. The abundance of shells and of highly worn specimens of Gryphaea and the presence of phosphatic nodules suggests that, in addition, this part of the Snap Band represents an episode of very slow sedimentation. The calcitic limonite oolite (type D) signifies a comparatively turbulent environment and was largely produced by erosion of the other ironstones. While erosion of con- solidated rock played a role, as evidenced by the presence of pebbles, it is here suggested that the most significant process might have been the winnowing away of unconsolidated mud from the clayey oolites by strong water agitation, not necessarily accompanied by much transport of ooliths. The only organisms which were certainly well adapted to these conditions were boring algae and the epibiont Entolium. Much of the remaining fauna was, in all likelihood, derived from the A and C ironstones though some species may have been able to tolerate periods of quieter water when mud was enabled to settle. The ironstone succession, as a whole, signifies a series of alternating phases of quiet and disturbed water and of reducing and oxidizing bottom conditions. In regard to the likely depth of formation of the D ironstone, Nadson (1927) observed that boring algae in the Black Sea are most abundant from sea-level down to 20 to 25 m. and he found no trace below 40 m. This seems plausible enough bearing in mind the light requirements of algae, but boring algae have been claimed also in deep sea corals (Duncan 1876). This apparently anomalous result seems to warrant further investigation. There is no evidence to indicate the immediate proximity of a shoreline, and the ironstone may have passed in all directions into marine shales, as it certainly does to the north and south. The comparative rarity of detrital quartz, except at the margins of the ironstone field, may signify the presence of an adjacent ‘clastic trap’ in slightly deeper water, as first suggested by Cayeux (Whitehead et al. 1952). We may accept with a fair degree of confidence that the ironstone formed on a shoal isolated from the land. Sedimentation rates seem generally to have been rather low, since the ironstone is markedly condensed compared with shales of the same age elsewhere. Periodically, shells exposed on the sediment surface were subjected to prolonged episodes of wave or current disturbance, boring, incrustation, and wear. Although, according to orthodox theory, the bottom waters must have been enriched in iron compared with the present day, the fauna gives little indication of having been affected. It is typical, with few exceptions, of that which flourished in Lower Sinemurian muds elsewhere. Compared with the nearest well-exposed shale succession of the same age, at Robin Hood’s Bay on the Yorkshire coast (which the author has examined in detail), Cardinia and EntoUum are relatively more abundant and the normally ubiquitous Oxytoma inaequivalve apparently rare or absent. Brachiopods have not been found in Yorkshire. These slight differences, and the large size to which many of the species grew, could be attributable, partly to higher temperatures and better oxygenation on the shoal, and partly to lower rates of sedimentation than on the surrounding sea floor. A. HALLAM; OBSERVATIONS ON PALAEOECOLOG Y 573 In the relative abundance of ammonites and endobionts and the rarity of brachiopods, except at one horizon, the Frodingham Ironstone differs considerably from the Marl- stone Rock-bed ironstone of the Middle Lias. It is similar in these respects, however, to the Cleveland Ironstone of Yorkshire, excepting that brachiopods are somewhat com- moner in the latter, though never dominant as in the Marlstone. These differences probably relate to the presence of muddy sediment in the original depositional environ- ments of the Frodingham and Cleveland Ironstones, suggesting that the controlling factor might have been the degree of water agitation. Acknowledgements. The author is grateful to representatives of the United Steel Company (Ore Mining Branch) and the Richard Thomas and Baldwin Steel Company for permission to study their workings. Messrs. Gathercole and Kirby, successive curators at the Scunthorpe Museum, and Dr. M. K. Howarth kindly made available ammonites in their charge. Thanks are due also to Miss J. Tarrant for the photographs. Miss F. Coxon for the diagrams and the Geochemical Department of the Grant Institute of Geology, Edinburgh, for the sulphur analyses. REFERENCES ARKELL, w. J. 1933. The Jurassic System in Great Britain. Oxford. CAiLLERE, s., and KRAUT, F., 1954. Les gisements de fer du bassin Lorrain. Mem. Mas. Hist. Nat. Paris, (C), 4, 1-175. CAROZZi, A. V. 1961. Distorted oolites and pseudo-oolites. J. Sed. Petrol. 31, 262-74. CAYEUX, L. 1914. Existence denombreuses traces d’Algues perforantes dans les mineraisde fer oolithique de France. C.R. Acad. Sci. 158, 1539-41. CROSS, J. E. 1875. The geology of north-west Lincolnshire; with an appendix by R. Etheridge. Quart. J. geol. Soc. Land. 31, 115-29. DAVIES, w., and dixie, r. j. m. 1951. Recent work on the Frodingham Ironstone. Proc. Yorks, geol. Soc. 28, 85-96. DEAN, w. T., DONOVAN, D. T., and HOWARTH, M. K., 1961. The Liassic ammonite zones and subzones of the north-west European province. Bull. British Mus. (Nat. Hist.): Geology, 4, 437-505. DONOVAN, D. T., 1954. Syiioptic supplement to T. Wright’s monograph on the Lias ammonites of the British Islands (1878-86). Pcd. Soc. Loud. DUDLEY, H. E. 1942. The Lower Lias beds in the Frodingham Railway Cutting, north-west Lincolnshire. Proc. Geol. Lond. 53, 152-5. DUNCAN, p. M. 1876. On some thallophytes parasitic within recent Madreporaria. Proc. Roy. Soc. Lond. 25, 238-57. DUNHAM, K. c. 1960. Syngenetic and diagenetic mineralisation in Yorkshire. Proc. Yorks, geol. Soc. 32, 229-84. HALLAM. A. 1960. A Sedimentary and faunal study of the Blue Lias of Dorset and Glamorgan. Phil. Trans. Roy. Soc. (B), 243, 1-44. HALLiMOND, A. F. 1925. Special reports on the mineral resources of Great Britain, XXIX. Iron Ores: Bedded ores of England and Wales. Petrography and Chemistry. Mem. Geol. Surv. HECHT, F. 1933. Der Verbleib der organischen Substanz der Tiere bei meerischer Einbettung. Sencken- bergiana, 15, 165-249. JOHNSON, R. G. 1957. Experiments on the burial of shells. J. Geol. 65, 527-35. LAMPLUGH, G. w. et al. 1920. Special reports on the mineral resources of Great Britain, XX. Bedded ores of the Lias, Oolites and later formations in England. Mem. Geol. Surv. MONAGHAN, p. H., and LYTLE, M. A. 1956. The origin of calcareous ooliths. J. Sed. Petrol. 26, 111-18. NADSON, G. 1927. Les Algues perforantes de la mer Noire. C.R. Acad. Sci. 184, 896-8. PETTiJOHN, F. J. 1957. Sedimentary Rocks. New York. 2nd ed. PULFREY, w. 1933. The iron-ore oolites and pisolites of North Wales. Quart J. geol. Soc. Lond. 89, 401-30. C 1456 PP 574 PALAEONTOLOGY, VOLUME 6 RASTALL, R. H., and HEMINGWAY, J. E., 1939. Black oolites in the Dogger of north-east Yorkshire. Geol. Mag. 76, 225-33. SAINT-SEINE, R. DE. 1954. Existence de Cirripedes acrothoraciques des le Lias: Zapfella pattei nov-gen., nov-sp. Bull. Soc. geol. Fr. 6, (4), 447-51. SCHAFER, w. 1956. Wirkungen der Benthos-Organismen auf den jungen Schichtverband. Sencken- bergiana Leth. 37, 183-263. SEiLACHER, A. 1952. Zur Eintcilung und Dcutung fossllcr Lcbensspuren. Uuiv. Tiibingen dissertation. swiNNERTON, H. H., and KENT, p. E. 1949. The gcology of Lincolnshire. Lines. N. H. Brochure, 1 (Lines. Nat. Union). TOPSENT, E. 1887. Contribution a I’etude des Clionides. Arch. Zool. exp. gen., ser. 2, suppl. no. 4, 1-165. ussHER, w. A. E. 1890. Gcology of parts of north Lincolnshire and south Yorkshire. Mem. Geol. Surv. WEiGELT, J. 1929. Fossile Grabschachte brachyurer Decapoden als Lokalgeschiebe in Pommern und das Rhizocorallium-BxohlQm. Zeitschr. fiir Geschiebeforsch. 5, 1-42. WHITEHEAD, T. H. et al. 1952. The Liassic Ironstones. Mem. Geol. Surv. WRIGHT, T. 1878-86. Monograph on the Lias ammonites of the British Islands. Pal. Soc. Lond. ZEUNER, F. E. 1933«. Die Lebensweise der Gryphaen. Palaeobiologica, 5, 307-20. 19336. Die Lage der Gryphaea arcuata im Sediment. Zbl. Miner. Geol. Paldont. B, 568-74. A. HALLAM Grant Institute of Geology, University of Edinburgh, Edinburgh, 9 Manuscript received 15 June 1962 A NEW HETEROMORPH AMMONITE FROM THE LOWER CRETACEOUS OF YORKSHIRE by J. C. DOYLE Abstract. Shasticrioceras anglicum sp. nov., which seems to be the earliest known species of the genus, is described and the significance of its occurrence in the Compound Nodular Band D.I., at Speeton in east York- shire, is discussed in the light of current knowledge. Recent collecting at Speeton has brought to light several genera of ammonites which are new to Britain. The present paper deals with one of these discoveries, apparently an ammonite of the genus Shasticrioceras. If this identification is correct, then so far as is known, this will be the first European record of this genus. Previously Shasticrioceras has been recorded from North America: California, Oregon, Washington (Anderson 1938), and from south-western parts of British Columbia bordering on the State of Washington. In these localities Shasticrioceras poniente Anderson has been selected as the index fossil of the lowest zone of the Barremian stage (Murphy 1956). The genus has also been described from Japan (Matsumoto 1953), where S. nipponicum Matsumoto together with Pulehellia ishidoensis Matsumoto form an ammonite zone which is taken to correspond with the Barremian stage. Several species of Shasticrioceras occur in the Barremian strata of the above loealities, but only in the U.S.A. has the genus previously been found as low as the Upper Hau- terivian (Popenoe et al. 1960). The occurrence of this genus at such a low Hauterivian horizon as the Compound Nodular Band in England would, if confirmed by later collecting of more satisfactory material, be of great interest and importance. Acknowledgements. I am greatly indebted to Dr. D. Jones of the U.S. Geological Survey for kindly supplying plaster casts of Anderson’s type material, and for permission to reproduce text-fig. \b. I would also like to thank Dr. J. W. Neale for his critical appreciation of the manuscript and helpful co-operation at all stages in the production of this paper. SYSTEMATIC DESCRIPTION Eamily ancyloceratidae Meek 1876 Subfamily crioceratitinae Wright 1952 Genus shasticrioceras Anderson 1938 Shastierioceras anglicum sp. nov. Plate 78, figs. 1-3; text-fig. \a Material. One small whorl fragment, D. 680, in the author’s collection. Horizon of holotype. The holotype was found on the upper surface of one of the large ‘compound nodules’ which characterize D.T. (for description of bed see Lamplugh [Palaeontology, Vol. 3, Part 1, 1963, pp. 575-8, pi. 78.] 576 PALAEONTOLOGY, VOLUME 6 1889). It is partly preserved in brown phosphate, and partly in the grey limestone of which the nodules are largely composed. Some of the shell is still retained, and the ammonite is almost certainly indigenous, for there is little indication of rolling. Spath (1924) has shown that the Compound Nodular Band represents a low Hauteri- vian horizon, and he places it in the zone of Lyticoceras noricus. However, Arend Thier- mann of Hamburg University is working on the Lyticoceratid fauna of an equivalent horizon in Germany, and his work should eventually make it possible to determine the horizon of D.I. more accurately. For the terminology of the detailed succession of ‘D’ beds, see Neale (1960). Dimensions of holotype. Maximum height of whorl ..... 20-5 mm. Width of whorl at maximum whorl height . . 10 0 mm. Width of venter measured at maximum whorl height 2-5 mm. Description. Whorl fragment exhibits slight coiling in one plane, and increases in height very slowly. Shell compressed in section, the height being a little more than twice the width. Ornament consists of dense, single, slightly sinuous ribs, which cross the narrow venter without interruption. At the ventro-lateral shoulders, the ribs are pinched into short, sharp spines, projecting backwards slightly, rather like the teeth of a saw. The characteristic form of these spines is only seen where the original shell is preserved. The latter is absent on one part of the holotype, where the low, blunt, internal cast of a spine base is exposed. The spines are displaced to a certain extent (PI. 78, fig. 2) thus disturbing slightly the bilateral symmetry of the specimen. This appears to be an original feature and not due to subsequent crushing; its significance is not apparent, but incipient helical coiling may be indicated. The suture-line is unfortunately not fully displayed, much of it being covered by the original shell. Sufficient is exposed, however, to show that the second lateral lobe is almost equal in length and breadth to the first lateral lobe (text-fig. la). Affinities. Although the material leaves much to be desired, it is none the less sufficient to allow comparison with, and distinction from, other known forms. Since little is known of Lower Cretaceous heteromorph ammonites (Casey 1960, p. 12), it is desirable to consider whether the specimen definitely belongs to the genus Shasticrioceras. Compression is acute and the peculiar spines are carried on the shoulders of an exceptionally narrow venter. These characters could be considered of specific rank if it could be shown that the Speeton form agreed in coiling with Shasticrioceras, and indeed it would appear to coil in a loose criocone like S. whitneyi Anderson. Dr. R. Casey, however (in litt. 1961), has suggested that it might possibly be part of the straight EXPLANATION OF PLATE 78 Figs. 1-3. Shasticrioceras anglicnmsp. nov. Holotype. Bed D.I, Lower Hauterivian; Speeton, Yorkshire. 1, Side view, X 2-5. 2, Ventral view, showing slightly oblique ventral ribbing, X 2-5. 3, Reverse side of holotype showing suture line, X 2-5. Figs. 4, 5. Nallaniphiura felli sp. nov. Holotype, CPC 4642, F22163, Locality T.T. 51. Bathurst Island, Northern Territory, Australia; Cenomanian. 4, General view, x4. 5, Enlarged view of portion, Xll. Palaeontology, Vol. 6 PLATE 78 DOYLE, Shasticrioceras angliciim SKWARKO, Nullamphhira felli J. C. DOYLE; A HETEROMORPH AMMONITE 577 shaft of some unnamed genus. Unfortunately the suture-line is not well enough shown to allow that to be used for a full comparison with the suture of a true Shasticrioceras. Dr. D. Jones of the U.S. Geological Survey {in lift. 1961) believes that the species of Shasticrioceras described by Anderson (1938) intergrade and are not so morphologically distinct as might be supposed. After seeing plaster casts of the holotypes I am inclined to agree with him, but whilst it is not the object of this paper to discuss the validity of these forms, in my opinion S. whitneyi Anderson is sufficiently distinct to justify specific differentiation. This species seems to be most closely related to S. anglicwn on morpho- logical grounds, both species being well compressed and apparently loosely coiled. The TEXT-FIG. 1. a, Shasticrioceras anglicinn sp. nov.; part of suture-line of holotype, x2-5. b. Suture-line of Shasticrioceras sp. USGS Loc. 1519, X 2-5. Other species of Shasticrioceras tend towards having an inflated whorl section and much tighter coiling. S. whitneyi also seems to be most closely related to S. anglicwn on stratigraphical grounds, for Imlay (1960) indicates that S. whitneyi and allied forms first occur in California in the zone of Hertleinites aqitila, which is generally supposed to represent a high Hauterivian horizon. However, S. whitneyi would appear to range into the Lower Barremian. All the other known species of Shasticrioceras occur in associa- tion with typically Barremian faunas. A specimen from the zone of Hertleinites aquila of Shasta County, California, de- scribed by Imlay (1960) as S. aff. whitneyi Anderson, possesses a rather narrow venter, but otherwise does not closely resemble S. anglicwn. In the same paper, Imlay mentions a specimen of Shasticrioceras, which, according to Murphy (quoted in Imlay, op. cit.), approximates to S. whitneyi in its openness of coiling, while differing in the ornamenta- tion, which is similar to that of S. poniente. This observation may be significant, since unlike S. whitneyi, where the ribs are straight, those of S. poniente are flexuous, as in S. anglicwn. However, it will be necessary to study many more specimens before any attempt at unravelling the lineage of the group can be made. b 578 PALAEONTOLOGY, VOLUME 6 REFERENCES ANDERSON, F. M. 1938. Lowcr Cretaceous deposits in California and Oregon. Spec. Pap. geol. Soc. Amer. 16. CASEY, R. 1960. The Ammonoidea of the Lower Greensand. Palaeontogr. Soc. (Monogr.) IMLAY, R. w. 1960. Ammonites of Early Cretaceous Age (Valanginian and Hauterivian) from the Pacific Coast States. Prof. Pap. U.S. geol. Siirv. 334-F, 167-225, pi. 24-43. LAMPLUGH, G. w. 1889. On the subdivisions of the Speeton Clay. Quart. J. geol. Soc. Loud. 45, 575-617. MATSUMOTO, T. 1953. Selected Cretaceous leading ammonites in Hokkaido and Saghalien. In Ap- pendix to The Cretaceous System in the Japanese Islands. Japanese Soc. for the Promotion of Scientific Research, Veno, Tokyo, Spec. Pub. 243-324, pi. 1-20. MURPHY, M. A. 1956. Lower Cretaceous stratigraphic units of northern California. Bull. Amer. Ass. Petrol. Geol. 40, 2098-2119. NEALE, J. w. 1960. The subdivisions of the Upper ‘D’ Beds of the Speeton Clay of Speeton, east Yorkshire. Geol. Mag. 97, 353-62. POPENOE, w. p., IMLAY, R. w., and MURPHY, M. A. 1960. Correlation of the Cretaceous Formations of the Pacific Coast (United States and North Western Mexico). Bull. geol. Soc. Amer. 71, 1491-1540. SPATH, L. F. 1924. The ammonites of the Speeton Clay and the subdivisions of the Neocomian. Geol. Mag. 61, 73-89. WRIGHT, c. w. 1952. A classification of the Cretaceous Ammonites. J. Paleont. 26. 213-22, figs. 1, 2. J. C. DOYLE 291 Ware Road, Manuscript received 9 February 1963 Hertford, Herts. A NEW UPPER CRETACEOUS OPHIUROID FROM AUSTRALIA by S. K. SKWARKO Abstract. A new ophiuroid, Nidlamphiura felli sp. nov., is described from Cenomanian (Upper Cretaceous) strata of Bathurst Island, Northern Territory, Australia. This is the first recorded fossil occurrence of the cosmo- pohtan living genus Nullamphiura Fell 1962. Only one other ophiuroid, Ophiacantha (Ophioglyphoida) fasten Chapman 1934, is known to have been described from Australian Cretaceous beds. During systematic investigation of the Mesozoic strata of the Northern Territory the writer visited Bathurst Island, north of Darwin, where rich collections of Cretaceous fossils have been collected in the past. The fossil collections made by the writer consist almost exclusively of Mollusca, but at Mirindow Point, on the southern coast of the island, a wave-cut cliff yielded a single incomplete specimen of an ophiuroid. As far as can be ascertained this is only the second ophiuroid to be found in Australian Cretaceous strata, though Brunnschweiler (1953, p.48) recorded an ‘abundance of small Ophiuroidea in certain beds’ of the Upper Jurassic Alexander Formation, Canning Desert, Western Australia. SYSTEMATIC DESCRIPTION Class ASTEROZOA (Subdivided after Fell 1960) Subclass OPHIUROIDEA Order ophiurida Mueller and Troschel 1840 Suborder ophiurae Mueller and Troschel 1840 Family amphiuridae Ljungman 1867 Genus nullamphiura Fell 1962 Type species. Amphiara psilopora H. L. Clark 1911. Nullamphiura felli sp. nov. Plate 78, figs. 4, 5; text-figs. 1, 2 Material. Single oral impression of an almost complete specimen; several poorly preserved impressions of arms of other specimens. Aboral aspect not available for examination. Specimen embedded in bluish clay of Cenomanian age; collected in a wave-cut cliff at Mirindow Point, southern coast of Bathurst Island, Northern Territory, Australia. Locality number, T.T. 51. Holotype, CPC 4642, fossil registration number, F22163; lodged at the Bureau of Mineral Resources, Geology and Geo- physics, Canberra, A.C.T. Diagnosis. Disc scaled; ratio of the size of ventral arm plates to lateral arm plates inereases distally along arms. Tentacle pores large, lacking tentacle scales; arm spines robust, tapering, two to eaeh lateral arm plate. Description. Specimen about 16 mm. across; width across the disc 3-3 mm. ; diameter of Palaeontology, Vol. 6, Part 3, 1963, pp. 579-81.] 580 PALAEONTOLOGY, VOLUME 6 the mouth, 0-7 mm. ; length of each arm, 6-5 mm. ; breadth of each arm within the limits of the disc, 0-7 mm. ; length of a spine, 0-45 mm. Individual extremities of mouth rounded rather than pointed. Jaw structures not preserved. Adoral shield very small, consisting of a thin curved plate concave to the periphery of an arm. The ventral surface of the disc scaled. b TEXT-FIG. 2a-c. NiiUamphium felli sp. nov. Holotype, CPC 4642, F22163. Locality T.T. 51, Bathurst Island, Northern Territory, Australia; Cenomanian; X 33. a. Proximal ventral arm plate with tentacle pores; b, distal ventral arm plate with spines and tentacle pores; c, proximal lateral arm plate with spine and tentacle pores. Visible ventral surface of the ventral arm plates approaches a heart-shape, with the base of the heart pointing orally. The most proximal plates outside the disc are narrower (parallel to the length of the arm) than the distal plates ; progressive distal elongation of ventral arm plates is accompanied by a slight inward migration of tentacle pores, which perforate both lateral margins of each ventral arm plate; plate surface pitted; highest relief of plate along aboral edge; plates touching or out of contact with each other. Shape of the inner and outer surface of the lateral arm plate not known ; lateral plate narrow in ventral view, with shape of a modified triangle with adoral and lateral margins S. K. SKWARKO: AN UPPER CRETACEOUS OPHIUROID 581 concave, aboral margin convex, and all three corners attenuated; each plate armed with two robust elongate tapering arm-spines, each about 0-45 mm. long. The lateral arm plates become relatively smaller in relation to the ventral arm plates nearer the arm-tip. The tentaele pores are large and have no tentaele scales. Remarks. The preservation of Niillamphiiira felli sp. nov. does not allow description of its internal structures or of its aboral aspeet ; even in the oral view not all structures are preserved. The nature of the jaws and teeth is not known so that some reservation is necessary on whether the new species is an amphiurid. On the other hand, features such as the scaled nature of the ventral portion of the disc, the distally increasing relative size of the ventral plates, and large tentacle pores without tentacle scales, all appear in the newly described genus Niillamphiura Fell 1962, and the similarities of the Bathurst Island specimen to this genus are so striking that little misgiving is felt about including it in the Amphiuridae. Niillamphiura Fell includes eight species. It is a cosmopolitan genus and ranges from the littoral zone to 1,800 metres (see Fell 1962). No fossil representatives of the genus have been described. The distinctive specific character of the new species is the remark- ably robust elongate tapering arm-spines. The only Cretaceous ophiuroid which has been described from Australia is Ophia- cantha {Ophioglyphoida) fosteri Chapman, from the Tambo( ?) beds (Lower Cretaceous) at Cleeve, near Longreach, Queensland (Chapman 1934). The new species can be dis- tinguished from it by its ventral arm plates, which distally become larger in relation to the lateral arm plates, and also an apparently smaller number of spines on each lateral arm plate. The shape of the ventral arm plate is quite distinctive in the Queensland form. N. felli sp. nov. does not contribute to our knowledge of the age of the strata from which it was derived, nor does it extend the range of ophiuroids in general. It is, however, the first recorded fossil form of the genus Nullamphiura, and its value lies in increasing our knowledge of these rare fossils and their distribution. The new species is named in honour of Professor H. Barraclough Fell, Zoology Department, Victoria University of Wellington, New Zealand. REFERENCES BRUNNSCHWEiLER, R. o. 1953. Mcsosoic Stratigraphy and history of the Canning Desert, Western Australia. J. geol. Soc. Aiist. 1, 35-54. CHAPMAN, F. 1934. A Lower Cretaceous brittle-star from Queensland. Proc. rov. Soc. Viet., n.s., 46, (2), 195-9, pi. 7. FELL, H. B. 1960. Synoptic keys to the genera of Ophiuroidea. Zool. Pttbl. Vic. Univ. Wellington, N.Z. 26. • 1962. A revision of the major genera of Amphiurid Ophiuridea. Trans, rov. Soc. N.Z., Zool. 2(1), 1-26. RASMUSSEN, H. w. 1950. Cretaceous Asteroidea and Ophiuroidea with special reference to the species found in Denmark. Damn. geol. Unders. ser. 2, 77, 1-134, pi. 1-18, figs. 1-8. S. K. SKWARKO Bureau of Mineral Resources, Geology and Geophysics, Childers Street, Manuscript received 22 October 1962 Turner, Canberra, A.C.T. THE RHAETIC-HETTANGIAN BIVALVE GENUS PTEROMYA MOORE by L. R. cox Abstract. Investigation of the genus Pteromya Moore, originally founded on external characters, has led to the conclusion that it should be referred to the family Ceratomyidae, which includes the genera Ceratomya and Gresslya. In addition to its type species, P. crowcombeia Moore, of the Rhaetic, it is considered that Pteromya should also include the basal Hettangian species ‘Plearomya' tatei Richardson & Tutcher, the Upper Rhaetic species "P.' langportensis R. & T. (described as a variety of P. tatei), and a new Lower Hettangian species Pteromya wilkesleyensis, found in cores from a borehole in Cheshire. As the Plearomya Beds of the basal Hettangian were so termed because of the abundance of 'P.' tatei in them, the discovery that this species is not a Plearomya renders the name inappropriate. The genus Pteromya was founded by Moore (1861, p. 505) for the reception of the two nominal species, P. crowcombeia and P. simplex, the type specimens of which came from blocks of fossiliferous Rhaetic limestone excavated during the construction of a canal tunnel at Bere Crowcombe, near Ilminster, Somerset. P. crowcombeia was designated as its type species by Stoliczka (1871, p. xv), but there can be no doubt that this and P. simplex should be regarded as synonyms. The systematic position of Pteromya, a genus based solely on the external characters of the shell, has hitherto been doubtful. Moore stated that it ‘evidently belonged’ to the ‘Myadae’, a family (now Myidae) then in- terpreted in a much broader sense than at the present day. Stoliczka (1871, p. xv, foot- note) remarked that ‘except by its thin structure, this shell does not appear to differ from Corbula', and much more recently Yokes (1945, p. 28) has expressed the same opinion. On the other hand, Tate (1876, p. 406) relegated Pteromya to the synonymy of Plearomya. In the hope of elucidating this question, I have examined the available material in the British Museum (Natural History), the Geological Survey Museum, the Geological Department of Bristol University, and the C. Moore Collection, at present being re- installed in its former home, the Museum of the Bath Royal Literary and Scientific Institution. I must express my thanks to Mr. P. Pagan, Director of the Victoria Art Gallery and Municipal Libraries, Bath, and his Committee, for permission to borrow the specimens from the C. Moore Collection; to the authorities of the Geological Survey for the loan of specimens and to Mr. H. Ivimey- Cook, of that institution, for calling my attention to the borehole specimens upon which the new species Pteromya wilkesleyensis is based; and to Dr. R. J. G. Savage, of the Geological Department of Bristol University, for looking out and sending to me specimens from the Departmental collection, as well as for help with various inquiries. Mr. C. P. Palmer, of the British Museum (Natural History), has rendered photographic assistance and also paid a special visit to Pinhay Bay, near Lyme Regis, where he was successful in obtaining specimens of F. tatei, a species not previously recorded from that locality. For reasons explained in the present paper, I consider that 'Plearomya' tatei Richard- son & Tutcher, from the basal Hettangian Prt-Planorbis Beds of England, and ‘P’. langportensis of the same authors, from the Upper Rhaetic Langport Beds (the White Lias proper), should be included in Pteromya in addition to the new species just men- tioned. The available material of the four species now recognized and of a form from the [Palaeontology, Vol. 6, Part 3, 1963, pp. 582-95, pis. 79-80.] L. R. COX: RHAETIC-HETTANGIAN BIVALVE 583 Rhaetic which I propose to record as Pteromya aff. crowcombeia is described below. A discussion of the affinities of Pteromya follows the specific descriptions. SYSTEMATIC DESCRIPTIONS Family ceratomyidae Arkell, 1934 Genus Pteromya Moore, 1861 Pteromya eroweombeia Moore Plate 79, figs. 1-5 1861 Pteromya eroweombeia Moore, p. 506, pi. 15, figs. 22, 23. 1861 Pteromya simplex Moore, p. 506, pi. 15, fig. 24. 1864 ? Pholadomya eorbiiloides Desh. ; Levallois, p. 395, pi. 6, figs. 2-4. 1893 Pleuromya eroweombeia (Moore); Woodward, p. 50, text-fig. 33. 1903 Pleuromya eroweombeia (Moore); Vaughan & Tutcher, p. 51. 1903 Pleuromya eroweombeia (Moore); Vaughan, p. 400. 1913 Pteromya simplex Moore; Jeannet, p. 293. 1916 Pteromya eroweombeia Moore; Richardson & Tutcher, p. 51, pi. 8, figs, lo, b, 2. 1933 Pteromya eroweombeia Moore; Arkell, pi. 29, fig. 9. 1945 Pteromya eroweombeia Moore; Vokes, p. 28, pi. 4, figs. 28, 29 (copied from Moore). Type speeimeus. Of nine syntypes of P. eroweombeia preserved in the C. Moore Collection at Bath, the original of Moore’s fig. 22, a left valve (PI. 79, fig. 1 of the present paper) registered as M. 136, is now designated as lectotype of the species. According to the collection label, the right valve represented in Moore’s fig. 23 is a specimen, registered as M. 137, which is of about the same size as the figure, but the posterior end of this specimen is defective and (unless the shell has been damaged since it was illustrated) it seems probable that the figure is composite, the posterior end being drawn from the smaller syntype (M. 138fl) represented in PI. 79, fig. 2 of the present paper. The holotype of P. simplex is a left valve registered as M. 139. Remarks. Moore’s syntypes came from the Lower Rhaetic of Bere Crowcombe, a locality where material can no longer be collected. Their matrix is a hard grey limestone. The specimens are elongate-ovate, inequilateral, Mya-Mkt shells up to about 26 mm. in length. There is a marked difference in shape between the left valve originally figured (the lectotype) and the right valve. In the former (PI. 79, fig. 1 of the present paper) there is, as stated in Moore’s description, an obtuse ridge passing from the umbo to the postero-ventral corner of the shell, separating a flattened postero-dorsal area, with a well-defined, obtuse outer angle, from the flank of the shell. In Moore’s figure of a right valve the postero-dorsal and posterior margins form a strongly convex, uninterrupted curve, which terminates in a sharp angle at the posterior end of the ventral margin, and there is no trace of a posterior ridge as in the other valve. Moore stated that ‘no right valve possessed the ridge and extended area’, and the marked difference between the two valves was the main basis for the erection of the new genus Pteromya. It is evident that if, as Moore apparently thought, the left and right valves of the same shell in this species differed greatly in outline as well as in the development of the diagonal ridge, the margins of its two valves could not have been exactly in contact when the shell was closed, but that one must have overlapped the other, as in certain Cor- bulidae. Unfortunately, however, all specimens of P. eroweombeia so far collected are single valves, so that the relationship of the left valve to the right cannot be determined 584 PALAEONTOLOGY, VOLUME 6 by direct observation. Examination of the somewhat limited material which is avail- able is sufficient to show that both valves are very variable in shape and that it can- not be assumed that in the same shell the left and right valves diifered so greatly in outline as those figured by Moore. It is true that in most of the available right valves from the type locality the dorsal and posterior margins form a strongly convex curve, as in the shell illustrated in PI. 79, fig. 2, but there are more elongate specimens (PI. 79, fig. 3) in which they meet in a distinct, obtuse angle. In at least one right valve from Bere Crowcombe, moreover, there is a trace of a posterior ridge. The left valve from this locality represented in PI. 79, fig. 4 is of interest on account of the strong upward-facing convexity of its ridge and of the narrowness of its postero-dorsal area. In some right valves from the same locality the posterior end of the shell is bent to the right, this feature being particularly noticeable in one of the syntypes registered as M. 137. It led Moore to conclude that the posterior end of the shell gaped. The ornament of the typical P. crowcombeia is of irregular concentric folds and ridges, which are present on the postero-dorsal area of the left valve as well as on the flank. Valves of F. crowcombeia with the shell preserved occur in a bed belonging to the Lower Rhaetic at Blue Anchor Point and at St. Audries Bay, both near Watchet. A posterior ridge is present in some left valves from these localities, but it is absent in others. Examination of right valves from this district clearly shows that in some specimens (PI. 79, fig. 5) there is a sharp, diagonally directed ridge near the posterior end of the shell from which a narrow area falls away steeply to the posterior and postero-dorsal margins, which themselves meet in an obtuse angle. Moore’s statement that in this species the ridge is always absent in the right valve is, therefore, incorrect. It is evident that the shell was inequivalve in that the right valve was more inflated than the left. The question whether the two valves of one and the same shell differed in outline, the margin of the right valve overlapping that of the left, must, however, remain open until specimens are discovered with the two valves still in juxtaposition. There is no evidence that the valves differed in this manner in the other species discussed in the present paper. The hinge structure of P. crowcombeia has been investigated by means of transverse sections perpendicular to the hinge-line. Unfortunately the material available for sectioning was very limited and in some specimens the shell structures proved not to be clearly distinguishable from the limestone matrix. In a right valve from Blue Anchor a section, which probably passed some distance posterior to the beak, shows the shell wall to be much thickened along the dorsal margin and to have a flange extending towards the left valve (text-fig. \d). A section through a left valve, probably in about the same position, shows the marginal region of the shell wall to be thickened, extended towards the right valve, and reflected upwards distally so as to have a hook-like cross-section. These features suggest that the hinge apparatus was similar to that of P. tatei, as revealed by the clearer transverse sections described under that species. In view of the very uniform character of the Rhaetic fauna throughout Europe (ex- cept where the Megalodon limestone facies occurs), it is remarkable that P. crowcombeia does not seem to have been recognized on the Continent, although Jeannet (1913, p, 293) recorded its synonym P. simplex from the Tours d’Ai, in the Prealpes Vaudoises. However, Mr. C. P. Palmer has discovered in the old collection of Gideon Mantell, in the British Museum (Natural History), some pieces of infra-Liassic sandstone from Vic-sur- Seille, east of Nancy, Lorraine, containing numerous ill-preserved specimens which L. R. COX; RHAETIC-HETTANGIAN BIVALVE 585 evidently belong to the species from that locality figured by Levallois as Pholadomya corbuloides Desh. (apparently previously a MS. name). These specimens are unmistak- ably referable to Pteromya and belong either to P. crowcombeia or to the form recorded below as P. aff. crowcombeia. Localities and horizons of material examined. Bere Crowcombe, near Ilminster, Somerset (type locality) ; Lower Rhaetic, Westbury Beds, in the so-called ‘flinty bed’. Blue Anchor Point, near Watchet, Somer- set ; Lower Rhaetic, Westbury Beds, Bed 2 1 of Richardson (1 9 1 1 , p. 1 7). St. Audries Bay, near Watchet, Somerset; Lower Rhaetic. Almondsbury, 8 miles north of Bristol, Gloucestershire; Lower Rhaetic. TEXT-FIG. 1. Transverse sections through four valves of Pteromya, crossing the hinge-margin pos- terior to the beak in each case, left and right as oriented: a, F. tatei (R. & T.), Lower Hettangian, Puriton, Somerset; left valve, no. LL. 30748, x2. b, P. tatei. Lower Hettangian, Somerset; right valve, no. L. 66273, x2. c,P. /‘a/‘ei,LowerHettangian, West Hatch, Somerset ; right valve, no. LL. 30749, X 2. d. P. crowcombeia Moore, Lower Rhaetic, Blue Anchor, Somerset; right valve, no. L. 25353, X 4. 1860 Myacites miiscidoides (?) Schlotheim; Wright, pp. 378, 385, 388 {non Schlotheim). 1865 Myacites muscidoides Schloth. ; Etheridge, pp. 226, 228, 229, 234 (non Schlotheim). 1871 Myacites muscidoides (dwarfed); Phillips, p. 107, pi. 7, fig. 36 {non Schlotheim). 1872 Myacites muscidoides Schl., dwarfed form; Blake, p. 142 (non Schlotheim). 1876 Anoplophora muscidoides (Schlotheim); Woodward et al., p. 88 (non Schlotheim sp.). 1876 Pleuromya crowcombeia (Moore); Tate, p. 406, pi. 13, fig. 10. Remarks. This form occurs in abundance in one particular bed of the Upper Rhaetic at Garden Cliff, Westbury-on-Severn. The specimens are all moulds of separate valves, strewn over a bedding plane, the largest being about 22 mm. long. There is no obvious difference between left and right valves as regards convexity, but this has probably been diminished by pressure in the course of fossilization. In most specimens of both valves the dorsal and posterior margins meet in a well-marked, obtuse angle. Traces of a posterior ridge are seen in some specimens of both valves, although it would not be expected that a ridge on the interior of the shell would always be represented on in- ternal moulds. In no specimens do the dorsal and posterior margins form a continuous, highly convex curve, as in many specimens of P. crowcombeia from its type horizon. No Pteromya aff. crowcombeia Moore Plate 79, figs. 6, 7 586 PALAEONTOLOGY, VOLUME 6 trace of the muscle scars or pallial line is seen on any of these specimens, nor do they throw any light on the hinge structure of the shell. Tate (1876), who figured a very similar specimen from Yorkshire, considered that the Garden Cliflf form, which earlier workers had recorded and Phillips had figured under the name Myacites musculoides (a Pleuromya of the German Trias), belonged to Pteromya crowcombeia, which he transferred to Pleuromya. It was, however, the opinion of Rich- ardson and Tutcher (1916, p. 52) that this form belonged to their Pleuromya tatei laugportensis, discussed below. In size and proportions the Garden Cliff specimens are much more similar to P. crowcombeia, as P. laugportensis is a larger and relatively less elongate form. They are, however, less variable than specimens of P. crowcombeia from its type horizon, and it is difficult to assess the significance of the absence among them of valves with the peculiar posterior outline just mentioned. In view of their later geological age, it is possible that they should be regarded as a distinct subspecies of P. crowcombeia, but much more material from its type horizon should be examined before a definite decision could be reached on this point. The Garden Cliff form is, therefore, here recorded as Pteromya aflf. crowcombeia Moore. Locality and horizon of material examined. Garden Cliff, Westbury-on-Severn, Gloucestershire; Upper Rhaetic, Cotham Beds, Bed 17 of Etheridge (1865, p. 238), immediately below the Meleagrinella fallax Bed, formerly called the Monotis Bed or Psendomonotis Bed. I have not been able to examine specimens from Yorkshire, such as were recorded by Blake (1872) and Tate (1876). Pteromya laugportensis (Richardson & Tutcher) Plate 79, figs. 8 a-c 1916 Pleuromya tatei var. laugportensis Richardson & Tutcher, p. 53, pi. 8, fig. 5. Type specimen. The specimen (L. 70446^) figured by Richardson and Tutcher and here reillustrated is designated as lectotype. ^ Specimens with registration numbers prefixed by ‘L.’ or ‘LL.’ are in the British Museum (Natural History). EXPLANATION OF PLATE 79 Figs. 1-5. Pteromya crowcombeia Moore. 1, Lower Rhaetic, Westbury Beds, Bere Crowcombe, Somerset. Lectotype, C. Moore Coll. (Bath), no. M. 136, x2; left valve. 2, Same horizon and locality. Paralectotype, same coll., no. M. 138u, x2; right valve. 3, Same horizon and locality, B.M., no. LL. 23113, x2; right valve. 4, Same horizon and locality, G.S.M., no. 90656, X L5; left valve. 5, Same horizon. Blue Anchor, Somerset, B.M., no. LL. 23114, X 3; right valve. Figs. 6, 7. Pteromya aff. crowcombeia Moore. Upper Rhaetic, Cotham Beds (Bed 17), Garden Cliff, Westbury-on-Severn, Gloucestershire. B.M., nos. 67461 and 38237, both Xl; groups of internal moulds, including both valves. Figs. %a-c. Pteromya laugportensis (Richardson & Tutcher). Upper Rhaetic, Langport Beds, Radstock Grove, Radstock, Somerset. Holotype, B.M., no. L. 70446, X 1 ; u, right valve; b, anterior view: and c, dorsal view with right valve uppermost. Figs. 9-13. Pteromya tatei (Richardson & Tutcher). 9, Lower Hettangian, West Hatch, Somerset, B.M., no. L. 69286, x 1-5; left valve. 10, Same horizon and locality. Lectotype, B.M., no. L. 70442, X 1 ; right valve. 11, Same horizon, Filton, near Bristol. Holotype of the ‘var. altior', B.M., no. L. 70445, X 1; right valve, internal mould. 12, Same horizon. Stoke Gifford, near Bristol, B.M., no. L. 77280, X 1 ; right valve. 13, Same horizon, ‘Somerset’, B.M., no. L. 66273, X 1 ; right vale. Palaeontology, Vol. 6 PLATE 79 COX, Ptewmya L. R. COX: RHAETIC-HETTANGIAN BIVALVE 587 Remarks. This species occurs in abundance in the White Lias proper (the Langport Beds) of Somerset, in the form of internal moulds of shells in which both valves had remained in juxtaposition. Its usual length is about 30 mm., the largest specimens seen being nearly 40 mm. long. It is characterized by its relatively high and unelongated form and by its strongly and asymmetrically convex ventral margin. The shell was evidently moderately inequivalve, as not only is the right valve the more strongly inflated, but its umbo is distinctly higher than that of the left valve (PL 79, fig. 8h). There was evidently no gape of the valve margins, as the margin of the internal mould, where intact, is every- where a sharp ridge. This margin does not lie exactly in a plane in every specimen. In some (L. 74404, L. 61420) the posterior end of the shell is distinctly bent to the right (a condition noted above in some specimens of P. crowcombeia), but in one (L. 30977) it is just as clearly bent to the left. There is no evidence that the margin of the right valve overlapped that of the left (a point raised when discussing P. crowcombeia), although it is doubtf^ul if internal moulds would give any decisive information on this question. There is no trace of a posterior ridge on the internal mould of either valve, nor is there a distinct postero-dorsal angle. In one specimen (L. 18059) there is some indication of what may have been a pallial line without a sinus, but it is obscure and its interpretation doubtful. There is no reason to dispute the view of Richardson and Tutcher that this form is congeneric with their "Pleiiromya' tatei, discussed below, but it seems to differ from it sufficiently to rank as a distinct species rather than as a ‘variety’. Localities and horizon of material examined. Radstock Grove, Radstock (type locality); Rockhill, Radstock; Burnett, near Keynsham; Queen’s Camel; Butleigh, near Glastonbury; Shepton Mallet; Langport; Weston, Bath; all in Somerset. Upper Rhaetic, Langport Beds (White Lias proper). Pteromya tatei (Richardson & Tutcher) Plate 79, figs. 9-13; Plate 80, figs. 1, 2 1893 Pleiiromya crowcombeia (Moore); Woodward, pp. 76, 82, 83, 119, 141, 152 (non Moore sp.). 1916 Pleuromva tatei Richardson & Tutcher, p. 52 (excluding part of synonymy), pi. 8, figs. 3a-c, 6. 1916 Pleiiromya tatei var. altior Richardson & Tutcher, p. 52, pi. 8, fig. 4. 1933 Pleiiromya tatei Richardson & Tutcher; Arkell, pi. 29, fig. 8. Type specimens. A nomenclatural complication arises from the fact that its authors, when describing this species, referred to it as a ‘nom. nov.’ rather than as a ‘sp. nov.’ and included in its synonymy ‘'Myacites mitscidoides, Geol. Surv., Phillips, Geology of Oxford, p. 107, plate vii, fig. 36’, representing the Garden Cliff form recorded in the present paper as Pteromya aff. crowcombeia Moore. Use of the term ‘nom. nov.’ properly implies the renaming of a homonym, although some authors have em- ployed it when founding new species on published figures they have considered to be misidentified. Thus, even if we may assume that Phillips had not intended to erect a new species Myacites mitscidoides, which was accidentally homonymous with Schlotheim’s species of that name, but, without citing the author’s name, was identifying his specimen with Schlotheim’s species on the authority of some unnamed member of the Geological Survey, it could be argued that the species Pleiiromya tatei was founded on the specimen figured by Phillips rather than on those described in the paper by Richardson and Tutcher. As, however, it was clearly the intention of these authors to found a new species primarily for the form from the basal Hettangian illustrated in their paper, it seems preferable to ignore their use of the term ‘nom. nov.’ and to accept their figured and other specimens as syntypes from which a lectotype can now be designated. I therefore now choose as lectotype of P. tatei the right valve (L. 70442) represented in pi. 8, fig. 3a, b of Richardson and Tutcher and reillustrated in PI. 79, fig. 10 of the 588 PALAEONTOLOGY, VOLUME 6 present paper. Of the figured residual syntypes (‘paralectotypes'), the original of pi. 8, fig. 3c is registered as L.70443 and the group represented in fig. 6 as L. 70444. The holotype of the variety altior (PI. 79, fig. 1 1 of the present paper), which does not seem worth distinguishing from the typical P. tatei, is registered as L. 70445. Remarks. This is the species that occurs in abundance in the lower beds of the Het- tangian of Somerset, Gloucestershire, and other counties. By some authors it was formerly recorded as Pleuromya [or Fteromya] crowcombeia, from which it is evident that its similarity to the Lower Rhaetic species was noticed. Its presence suggested the name "Pleuromya Limestones’ or "Pleuromya Beds’ for the strata in which it occurs, but as it is shown in the present paper that it does not belong to Pleuromya, the name is inappro- priate and its use might well be discontinued. Specimens in the R. F. Tomes Collection in the British Museum (Natural History) were found identified as Myacites jurassinus Quenstedt. Through the kindness of Professor H. Holder, of Tubingen, I have, however, examined the holotype of that species (Quenstedt, 1856, p. 49, pi. 5, fig. 1) and ascer- tained that it is a distinct form and a true Pleuromya. The largest specimens of P. tatei which have been examined are 40 mm. long. The species occurs mainly as isolated valves and there is no obvious dilference in convexity between left and right ones. The only specimen (Brit. Mus. 67458) with the two valves in association which has been seen is from Thurlbeer, Somerset. The right valve seems to be the more convex, although the dilference is not pronounced, and the right umbo is slightly higher than the left one ; the specimen has, however, been slightly sheared, and it is difficult to assess to what extent these were original features of the shell. The posterior end of the specimen is broken and it cannot, therefore, be observed if the valve margins gaped. Ventrally and anteriorly the margins are exactly in contact; here there was certainly no overlap of the margin of the right valve over that of the left. In some specimens of both valves a very obtuse posterior ridge is well seen (PI. 79, figs. 9, 10) and forms the boundary of a postero-dorsal area from which the concentric ribs found on the flank of the shell are absent. There are, however, many specimens of both valves in which a ridge cannot be observed at all, or in which it is seen only near the posterior end of the shell. There are right valves (PI. 79, figs. 11, 13) in which the posterior and dorsal margins form a strongly convex, uninterrupted curve exactly as in many right valves of P. crowcombeia. The external features of the shell certainly do not serve as a basis for the generic separation of tatei from crowcombeia. In some specimens of P. tatei concentric ribbing is present on the whole of the flank of the shell, whereas in others it is absent from later growth stages or fades away on the posterior part of the flank. The exterior of the shell, well seen in many specimens from the Bristol district, appears to be devoid of the minute pustules present in Pleuromya. EXPLANATION OF PLATE 80 Figs. 1, 2. Pteromya tatei (Richardson & Tutcher). 1, Lower Hettangian, West Hatch, Somerset, B.M., no. L. 70444, X 1 ; group of specimens, including left and right valves. 2, Same horizon, Hewitt’s quarry. Patchway, near Bristol, B.M., no. L. 77282, X 1; left valve. Figs. 3-5. Pteromya wilkesleyensis, sp. nov. Lower Hettangian, Geological Survey borehole at Wilkesley, Cheshire. 3, Pre-planorbis Beds, depth 503 ft. 3 in. Paratype, G.S.M., no. 108104, X 1; right valve. 4, Pre-planorbis Beds, depth 487 ft. Holotype, G.S.M., no. 108103, xl; left valve. 5, Planorbis Zone, depth 470 ft. 1 in. Paratype, G.S.M., no. 108100, X 1; right valve, associated with the ammonite Psiloceras planorbis. Palaeontology, Vol. 6 PLATE 80 COX, Pteromya 1^?; L. R. COX: RHAETIC-HETTANGIAN BIVALVE 589 Although a few specimens are preserved as internal moulds, and parts of the shell wall have peeled off in others, no muscle scars or clear impressions of the pallial line can be seen. As in the case of P. langportensis, there is, however, a distinct suggestion that the pallial line was without a sinus. Several transverse sections have been prepared passing through one of the valves a short distance behind the beak. In the left valve (text-fig. la) the dorsal marginal region is thickened and reflected upwards, its cross-section recalling a tobacco pipe. This is precisely the structure found in the genera Ceratomya (text-fig. 2b) and Gresslya (text-figs. 2a, 4b, c). Transverse sections through the right valve (text-fig. lb, c) just behind the beak show that the margin is not reflected but is much TEXT-FIG. 2. Transverse sections through dorsal half of shells of Gresslya and Ceratomya, passing well posterior to beaks, left and right as oriented: a, Gresslya abdiicta (Phillips), Inferior Oolite, England, no. LL. 8049, x2. b, Ceratomya bajociana (d'Orbigny), Inferior Oolite, England, no. 67257, X 1. thickened, the thickened portion projecting into the cavity of the valve and being ex- cavated in one or more places on its lower side. Here, again, there is a general similarity to the marginal region of the right valve in Ceratomya (text-fig. 2b) and Gresslya (text- figs. 2a, 4b, c). In Lang’s Bed H 1 at Pinhay Bay, west of Lyme Regis, a bedding plane of shaly limestone is covered with pseudomorphs of small valves, none more than 12 mm. long, which probably belong to a small race of this species. A few show traces of concentric ribbing. Many of these specimens lie with the interior of the valve facing upwards, so that the internal features of the dorsal margin can be seen, although not very plainly. Tooth-like protrusions of the margin, like those found in Pleuromya, seem to be absent. It is remarkable that Pteromya talei, so abundant at its particular horizon in England, has not been reported from other countries or, so far as can be ascertained, described under any other name. Localities and horizons of material examined. West Hatch, near Taunton, Somerset (type-locality); Puriton, Somerset; Selworthy, Somerset; Thurlbeer, Somerset; Patchway (Hewett’s Quarry), Purdown, Horfield, Eilton, Redland, and Stoke Gifford, all near Bristol; Aust Cliff, Gloucestershire (small forms) Penarth, Glamorgan; Craycombe, near Pershore, Worcestershire; Pinhay Bay, near Lyme Regis ( Beds H 1 and H 2 of Lang). All Lower Hettangian, Pre-Planorbis Beds. Binton, Warwickshire (‘Guinea Bed’ and ‘Lirestone’); Bickmarsh, east of Cleeve Prior, Warwick- shire (‘Potstone’); Leavening, Yorkshire. All Lower Hettangian, Planorbis Zone. Addesley, Shropshire (presumably from a boring); horizon uncertain. Pteromya wilkesleyetisis, sp. nov. Plate 80, figs. 3-5 Diagnosis. Shell large for the genus (length of largest valve 63 mm.), ovate, moderately C 1460 Q q 590 PALAEONTOLOGY, VOLUME 6 inequilateral to subequilateral, ratio of length to height varying from about 3:2 to 1:1, posterior outline variable; posterior diagonal ridge present in some specimens (left valves in the material studied), absent in others. Earlier growth-stages ornamented with concentric folds of low amplitude; greater part of surface smooth except for irregular growth-lines and rugae. Type specimens. The left valve (Geological Survey, no. 108103) represented in Plate 80, fig. 4 is selected as holotype. There are six paratypes, including nos. 108100-2, 108104, in the same collection. Remarks. The specimens upon which this species is based are crushed isolated valves preserved in grey shale from the cores of a borehole. The holotype (PI. 80, fig. 4) is a left valve, 63 mm. long and about 40 mm. high, in which the umbo is situated well anterior to mid-length, the posterior end of the shell is low and truncated, the postero-dorsal margin is feebly convex and gently inclined, and a very obtuse posterior diagonal ridge separates the flank from a moderately wide, slightly concave postero-dorsal area. A second left valve (no. 108102) appears to have been less elongated, with the length and height almost equal and the umbo probably almost median in position, but the shell is broken away anteriorly. Its postero-dorsal margin is steeply inclined, forming an obtuse, rounded-ofl' angle with the low, flattened posterior margin. A very obtuse but distinct diagonal ridge is present. The remaining specimens all seem to be right valves. No. 108101 is of about the same size and proportions as the holotype. Although the postero- dorsal margin of this shell is broken away, if a diagonal ridge had been present at least its posterior end should be visible, but there is no trace of it. Nos. 108100 (pi. 80, fig. 5) and 108104 (fig. 3) are rather smaller right valves, both with a long, feebly convex posterior margin and a very high, round-off postero-dorsal angle. Neither has a diagonal ridge, a ridge parallel to the posterior margin and terminating along the ventral margin in the second specimen evidently being the result of crushing. It is improbable that much significance should be attached to the fact that the only valves in which a ridge is present are left ones, as the amount of material is so small. It is, however, interesting to note that earlier workers thought (incorrectly, as shown above) that the ridge is confined to the left valve in Pteromya crowcombeia. Although all the specimens are crushed, there is a distinct suggestion that, as in P. crowcombeia, the right valves of P. wilkesJeyensis are more inflated than the left ones. The material is too limited and friable for investigation of the internal characters of the valves to be possible, but external characters suggest strongly that this species is congeneric with Pteromya tatei and P. crowcombeia. Its relatively large size distinguishes it from both of these species and its concentric ribbing terminates at an earlier stage of growth than in any specimens of either. Tate (1876, p. 406) referred to the presence at a Yorkshire locality (it is not clear which this was) of ‘very large shells, seemingly adult forms of that species [i.e. the species which he records as P. crowcombeia] but which have a strong resemblance to Ceromya infra-Iiassica Peters’. It is conceivable that these shells may have belonged to the species now described. C. infra-Iiassica is, however, a Gresslya. Locality and horizons of material examined. Wilkesley, Cheshire, Geological Survey borehole (Grid Reference SJ 628415), from depths ranging from 467 ft. 5 in. to 503 ft. 3 in. The base of the Lower Lias was encountered at 517 ft. 9 in. The holotype (no. 108103) and a figured right valve paratype (no. 108104) are from Pre-Planorbis Beds at the respective depths of 487 ft. 0 in. and 503 ft. 3 in., while the remaining specimens are from the overlying Planorbis Subzone. L. R. COX: RHAETIC-HETTANGIAN BIVALVE 591 Pteromya sp. Pteromya sp. nov.? Melville, 1956, pp. 76, 94, pi. 6, fig. 15. Melville has recorded the discovery of a single small right valve of a Pteromya, about 15 mm. long, in a core from a borehole in Gloucestershire. It is of interest as coming from a horizon higher than that of any other known specimen undoubtedly referable to the genus. As Melville remarks, the specimen seems to differ very little from P. crow- combeia and, if its species is new, ‘the available material is inadequate for a full descrip- tion’. A posterior ridge is clearly seen. Locality and horizon. Borehole at Stowell Park, Gloucestershire (National Grid Reference SP 084118). Hettangian, Angulata Zone, Subzone of Alsatites laqitens (the top subzone of the Alsatites liasicns Zone of the recent scheme of Dean, Donovan, and Howarth). GENERAL CONCLUSIONS The conclusion that ‘‘Pleuromya’’ tatei is congeneric with Pteromya crowcombeia is based mainly on external characters. In both forms the range of variation in the general outline of the shell is very similar. The posterior and dorsal margins meet in an obtuse angle in some specimens, but form a continuous, strongly convex curve in others. In both forms a posterior diagonal ridge may be either present or absent in both left and right valves. The statement made by Richardson and Tutcher that the ridge is confined to the left valve in P. crowcombeia, but is present in both valves in P. tatei is not in ac- cordance with the facts. The lack of bivalved specimens of the first species makes it impossible to say if, in the same shell, the ridge was either present or absent in both valves. In the single available bivalved specimen of P. tatei it is absent in both valves on the part of the shell which remains, but the specimen is imperfect. Some earlier authors, as mentioned above, misidentified P. tatei as P. crowcombeia. It is now suggested that these should be regarded as distinct species of the same genus. Besides differing greatly in size, they differ in the degree of inequality of the two valves. In P. crowcombeia the right valve is more strongly convex than the left to a pronounced extent, whereas in P. tatei the difference in convexity is much less marked. In P. langportensis, a species of intermediate geological age, the inequality of the valves is easily observable and is more marked than in P. tatei. The conclusion that tatei is not a Pleuromya is based mainly on its hinge structure as revealed by transverse sections, but also on external characters, for species of Pleuromya with a posterior ridge are unknown. If we accept the further conclusion that external characters indicate that it is congeneric with crowcombeia, type species of Pteromya, it is evident that the transverse sections illustrated in text-fig. 1 throw new light on the affinities of that genus. As has been pointed out, the structures revealed are very similar to those present in Ceratomya and Gresslya, and it is to be inferred that, as in those genera, the ligament in Pteromya was elongated, undivided, and subinternal, extending posteriorly from the beaks between the thickened margin of the right valve and the reffected margin of the left valve, as indicated in text-figs. Ab, c, which represent the condition in Gresslya. The family Ceratomyidae should, therefore, be extended to include Pteromya in addition to Ceratomya and Gresslya. It may be pointed out that in Gresslya there is commonly a tendency for the posterior and dorsal margins to form an 592 PALAEONTOLOGY, VOLUME 6 uninterrupted curve and for a postero-ventral angle to be present, just as in many valves of Pteromya. Some authors have suggested that Pteromya belonged to the Corbulidae. Not only is the shell wall much thinner and the maximum size of specimens greater than in species TEXT-FIG. 3. Successive transverse sections through shell of Corbiila (Bicorbiita) gaUicci Lamarck, no. L. 66203, left and right as oriented, x2: a, just posterior to beaks, showing chondrophore pro- jecting from left valve; b, almost through beaks; c, just anterior to beaks (both b and c showing tooth- like processes of the two valves); d, well anterior to beaks, beyond positions of tooth-like processes. TEXT-FIG. 4. Transverse sections through hinge-line of Corbida and Gresslya, showing position of ligament, left and right as oriented, a, Corbida (Bicorbida) exarata Deshayes, Eocene, France; partial transverse section just posterior to the beaks revealed by oblique longitudinal section, showing chondrophore projecting from left valve and ligament (dotted), X2. b, c, Gresslya pingids Agassiz, Toarcian, France; two parallel sections a little posterior to the beaks, showing thickened margins of both valves and ligament (dotted), x 3-3. Note the reflected margin of left valve, its cross-section recalling a tobacco-pipe. (After Douville, 1907.) of that family, but the hinge structure differs considerably. Text-figs. 3a-d represent a series of transverse sections passing through a species of Corbida (Bicorbida) near the beaks, while text-fig. 4n indicates the position of the ligament. The sections which pass just behind the beaks (text-figs. 2>a, 4a) reveal a narrow, elongated chondrophore projecting from the margin of the left valve far into the umbonal cavity of the right. The ligament was not extended longitudinally, but was situated between this chondrophore and the interior of the umbonal region of the valve (text-fig. 4a). More anteriorly (text- figs. 3b, 3c), below and just in front of the beaks, the sections do not pass through the L. R. COX: RHAETIC-HETTANGIAN BIVALVE 593 chondrophore, but between interlocking teeth projecting from the respective margins. Beyond these teeth (text-fig. M) the margins of the two valves are in simple contact. It will be seen that cross-sections passing through the two valves of Pteromya posterior to the beaks (text-fig. 1) differ considerably from the corresponding sections, through Corbula valves. Pleuromya differs from GressJya, Ceratoniya, and Corbula in possessing a short, external, opisthodetic ligament attached, in each valve, to a flat-topped ridge (or nymph) extending back from the beak, the length of the ligament being about 5 mm. in specimens of average size. In text-fig. 5b, c cross-sections of these nymphs appear as hook-like upturnings of the marginal region (the fact that one lies at a slightly higher level than the TEXT-FIG. 5. Successive transverse sections through dorsal half of shells of Pleuromya imifonitis (J. Sowerby), Bajocian, England, left and right as oriented; a-c, no. 50931, X 1-5; cl-f, no. LL. 30752 (a smaller shell), x2-3: a, 5-6 mm. posterior to beaks, valve margins sharp; b, 4-2 mm. posterior to beaks, passing through nymphs for attachment of ligament; c, I -3 mm. posterior to beaks, at anterior end of ligamental nymphs; d, just posterior to beaks, where valve margins are notched; e, below beak, showing thick, tooth-like protrusion from left valve margin;/ 0-5 mm. anterior to beaks, valve margins sharp and contiguous. Other is due to the fact that the two valves have been slightly sheared in the specimen sectioned). Immediately behind the beaks (text-fig. 5d) the space between the two mar- gins widens owing to the presence of a notch in each. Below the beaks (text-fig. 5e) is a thickened protrusion from the margin of the left valve the top of v/hich fits into the anterior end of the notch in the right valve (although this relationship cannot be seen in the section). More anteriorly (text-fig. 5/) the margins are thin, symmetrical, and almost contiguous. It is evident that the structures shown in text-fig. 1, observable in speci- mens of Pteromya, are quite unlike those seen anywhere along the hinge-margin of Pleuromya. The following are emended diagnoses of the family Ceratomyidae and of the genus Pteromya drawn up in accordance with the conclusions reached in the present paper. Family ceratomyidae Arkell, 1934 Diagnosis. Shell ovate, longer than high, inequilateral, moderately to strongly infiated, some specimens inequivalve; valve margins closed or with a narrow posterior gape; 594 PALAEONTOLOGY, VOLUME 6 shell-wall rather thin; beaks slightly to strongly prosogyrous; no demarcated lunule or escutcheon; ligament opisthodetic, subinternal, located between the reflected and thickened postero-dorsal margin of the left valve and the overlapping margin of the right valve, which has a subinternal thickening projecting into the cavity of the valve; true hinge-teeth absent, replaced in some forms by thickenings or protuberances of the dorsal margins ; pallial line variable ; surface of shell with variously oriented plications or unornamented; minute pustules present in some forms. Genus pteromya Moore, 1861 Diagnosis. Subovate, very slightly to strongly inequilateral, not strongly inflated, slightly to moderately inequivalve, right valve the more gibbose and with its umbo the more elevated; valve margins not gaping; umbones broad, protruding very little; postero- dorsal and posterior margins forming a continuous, strongly convex curve or else meet- ing in an obtuse angle, in which case a very obtuse posterior ridge may be present; hinge-structure as defined for the family; adductor scars and pallial line not yet clearly observed, the latter probably without a sinus; surface with concentric undulations or ridges, or merely with coarse growth-threads; no surface pustules as yet observed. Range : Lower Rhaetic to Lower Hettangian. REFERENCES ARKELL, w. J. 1933. The Junissic System in Great Britain. Oxford, xii + 681, 41 pi. BLAKE, J. F. 1872. On the infralias in Yorkshire. Quart. J. geol. Soc. Land. 28, 132-146. ETHERIDGE, R. 1865. On the Rhaetic or Avicula contorta Beds at Garden Cliff, Westbury-on-Severn, Gloucestershire. Proc. Cotteswohl Nat. Fid. CL 3, 218-234, 1 pi. JEANNET, A. 1913-8. Moiiographie geologique des Tours d’Ai et des regions avoisinantes (Prealpes vaudoises). Mat. Carte geol. Suisse, 64, 1-701, 3 pis. LEVALLOis, J. 1864. Les couches de junction (Grenz Schichten) du Trias et du Lias dans la Lorraine et dans la Souabe. Bull. Soc. geol. Fr. (2), 21, 384-440, pi. 6. MELVILLE, R. V. 1956. Stiatigraphical palaeontology, ammonites excluded, of the Stowell Park Bore- hole. Bull. geol. Surv. U.K. 11, 67-139, pi. 2-8. MOORE, c. 1861. On the zones of the Lower Lias and the Avicula contorta Zone. Quart. J. geol. Soc. Loud. 17, 483-516, pi. 15, 16. PHILLIPS, J. 1871. Geology of Oxford and the Valley of the Thames. Oxford, xxiv + 523, 17 pi. QUENSTEDT, F. A. 1856-8. Der Jura. Tubingen. 1-842, 100 pi. RICHARDSON, L. 1911. The Rhaetic and contiguous deposits of west, mid, and part of east Somerset. Quart. J. geol. Soc. Lond. 67, 1-74, pi. 1-4. RICHARDSON, L., and TUTCHER, J. w. 1916. On Pteromya crowcombeia Moore and some species of Pleuromya mA Volsella from the Rhaetic and Lower Lias. Proc. Yorks, geol. Soc. 19, 51-58, pi. 8, 9. STOLiczKA, F. 1870-1. Cretaceous fauna of southern India. 3. Pelecypoda. Palaeont. indica xxii-f 537, 50 pi. TATE, R. 1876. Class Lamellibranchiata. In tate, r. and blake, j. f.. The Yorkshire Lias. London, 357-412, pi. 11-14. VAUGHAN, A. 1903. The lowest beds of the Lower Lias at Sedbury Cliff. Quart. J. geol. Soc. Lond. 59, 396-402. and TUTCHER, j. w. 1903. The Lower Lias of Keynsham. Proc. Bristol Nat. Soc., n.s., 10, 3-55. VOKES, H. E. 1945. Supraspecific groups of the pelecypod family Corbulidae. Bull. Amer. Mus. nat. Hist. 86, 1-32, pi. 1-4. WOODWARD, H. B. 1893. The Lias of England and Wales (Yorkshire excepted). The Jurassic Rocks of Britain {Mem. geol. Surv. U.K.), 3, xii-|-399. L. R. COX: RHAETIC-HETTANGIAN BIVALVE 595 WOODWARD, H. B., ct al. 1876. Geology of east Somerset and the Bristol Coal-fields. Mem. geol. Surv. U.K. 1-271. WRIGHT, T. 1860. On the Zone of Aviaila contorta, and the Lower Lias of the south of England. Quart. J. geol. Soc. Loud. 16, 374-41 1. L. R. cox Department of Palaeontology British Museum (Natural History), Cromwell Road, Manuscript received 26 January 1963 London, S.W.7 THE PALAEONTOLOGICAL ASSOCIATION COUNCIL 1963 President Professor T. Neville George, The University, Glasgow, W. 2 Vice-Presidents Professor O. M. B. Bulman, Sedgwick Museum, Cambridge Dr. L. R. Cox, British Museum (Natural History), London Dr. W. H. C. Ramsbottom, Geological Survey Office, Leeds Treasurer Dr. T. D. Ford, Department of Geology, The University, Leicester Secretary Dr. C. H. Holland, Department of Geology, Bedford College, London, N.W. 1 Editors Mr. N. F. Hughes, Sedgwick Museum, Cambridge Dr. W. S. McKerrow, University Museum, Oxford Dr. Gwyn Thomas, Department of Geology, Imperial College of Science, London, S.W. 7 Dr. I. Strachan, Department of Geology, The University, Birmingham, 15 Other members of Council Dr. C. G. Adams, British Museum (Natural History), London Dr. F. M. Broadhurst, The University, Manchester Dr. W. J. Clarke, British Petroleum Company, Sunbury on Thames Dr. W. T. Dean, British Museum (Natural History), London Dr. C. Downie, The University, Sheffield Dr. R. Goldring, The University, Reading Dr. M. R. House, The University, Durham Mr. M. Mitchell, Geological Survey and Museum, London Dr. J. W. Neale, The University, Hull Professor F. H. T. Rhodes, University College, Swansea Dr. A. J. Rowell, The University, Nottingham Dr. R. J. G. Savage, The University, Bristol Dr. C. D. Waterston, Royal Scottish Museum, Edinburgh Mr. C. W. Wright, London Overseas Representatives Australia: Professor Dorothy Hill, Department of Geology, University of Queensland, Brisbane Canada: Dr. D. J. McLaren, Geological Survey of Canada, Department of Mines and Technical Surveys, Ottawa India: Professor M. R. Sahni, Department of Geology, Panjab University, Chandigarh New Zealand: Dr. C. A. Fleming, New Zealand Geological Survey, P.O. Box 368, Lower Hutt West Indies and Central America: Dr. L. J. Chubb, Geological Survey Department, Kingston, Jamaica Eastern U.S.A. : Professor H. B. Whittington, Museum of Comparative Zoology, Harvard Univer- sity, Cambridge 38, Mass. Western U.S.A.: Dr. J. Wyatt Durham, Department of Paleontology, University of California, Berkeley 4, Calif. PALAEONTOLOGY VOLUME 6 • PART 3 CONTENTS Cambrian trilobites from the Purley Shales of Warwickshire. By J. D. D. smith and D. E. wiHTE 397 Some Calamitean plants from the Lower Carboniferous of Scotland. By M. CHAPHEKAR 408 The ostracod species Orthonotacythere inversa (Cornuel) and its allies in the Speeton Clay of Yorkshire. By p. kaye 430 A revision of the brachiopod family Leptocoehidae. By A. amos and A. J. BOUCOT 440 The spines and diffuse fascioles of the Cretaceous echinoid Echinocorys scutata Leske. By D. G. Stephenson 458 The anatomical structure and systematic position of Pentablastus (Blastoidea) from the Carboniferous of Spain. By K. A. joysey and A. breimer 471 Devonian goniatites and stratigraphical correlations in Western Canada. By M. R. HOUSE and a. e. h. redder 491 On growth stages in Branchiosaurs. By D. m. s. watson 540 Observations on the palaeoecology and ammonite sequence of the Frodingham Ironstone (Lower Jurassic). By a. hallam 554 A new heteromorph ammonite from the Lower Cretaceous of Yorkshire. By J. C. DOYLE 575 A new Upper Cretaceous Ophiuroid from Austraha. By s. w. skwarko 579 The Rhaetic-Hettangian bivalve genus Pteromya Moore. By L. R. cox 582 PRINTED IN GREAT BRITAIN AT THE UNIVERSITY PRESS, OXFORD BY VIVIAN RIDLER, PRINTER TO THE UNIVERSITY VOLUME 6 • PART 4 Palaeontology DECEMBER 1963 PUBLISHED BY THE PALAEONTOLOGICAL ASSOCIATION LONDON Price £3 THE PALAEONTOLOGICAL ASSOCIATION The Association was founded in 1957 to further the study of palaeontology. It holds meetings and demonstrations, and pubhshes the quarterly journal Palaeontology. Membership is open to individuals, institutions, libraries, &c., on payment of the appropriate annual subscription: Institute membership . . . . £5. 5s. (U.S. $15.50) Ordinary membership . . . . £3. 7>s. (U.S. $9.50) Student membership . . . . £2. Is. (U.S. $6.50) There is no admission fee. Student members will be regarded as persons receiving full-time instruction at educational institutions recognized by the Council; on first applying for membership, they should obtain an application form from the Secretary or the Treasurer. Subscriptions are due each January, and should be sent to the Treasmer, Dr. T. D. Ford, Department of Geology, The University, Leicester, England. Palaeontology is devoted to the pubhcation of papers (preferably illustrated) on all aspects of palaeontology and stratigraphical palaeontology. Four parts are published each year and are sent free to aU members of the Association. Members who join for 1964 will receive Volume 7, Parts 1 to 4. AU back numbers are stiU in print and may be ordered from B. H. BlackweU, Broad Street, Oxford, England, at the prices shown below (post free) : Vol. 1 (for 1957-8) in 4 parts at £2 or U.S. $6.00 per part. Vol. 2 (for 1959) in 2 parts at £2 or U.S. $6.00 per part. Vol. 3 (for 1960) in 4 parts at £2 or U.S. $6.00 per part. Vol. 4 (for 1961) in 4 parts at £2 or U.S. $6.00 per part. Vol. 5 (for 1962) in 4 parts at £3 or U.S. $9.00 per part. Vol. 6 (for 1963) in 4 parts at £3 or U.S. $9.00 per part. A complete set, Volumes 1-6, consists of 22 parts and costs £52 or U.S. $156. Manuscripts on aU aspects of palaeontology and stratigraphical palaeontology are invited. They should conform in style to those already pubUshed in this journal, and should be sent to Mr. N. F. Hughes, Sedgwick Museum, Cambridge, England. A sheet of detaUed instructions for authors wiU be suppUed on request. ERRATUM Palaeontology, Volume 6, part 2, page 356, line 27 For ‘Genus palynomorphytes gen. nov.’ /•por/ ‘Genus palynomorphites gen. nov.’ \: » . i- 1 / \ ■ 1/ ■'I \ k ! " ■■'s / ' • ' .J 1 • l>. j >K WlV" ' ■ : • \ ^ \ i :■ r ‘V j 'vs'l - ,'•■■. ,V<.'\ V ^ ■ \ ' "' l A \i H.*. ^v-.c’.- '\ >■ ,r3 ''n^yM ■/:■>„ P ,/j I/.' I tJ-' i Wi'hiii '■c , CRETACEOUS AMMONITES FROM BATHURST ISLAND, NORTHERN AUSTRALIA by C. W. WRIGHT Abstract. Sixteen species of ammonites are described from the Lower Albian, Cenomanian and Turonian. They include new species of Sciponoceras, Scapbites, Chimbuites and Acanthoceras. Others are well-known wide- ranging species. Chimbuites is now securely dated to the Cenomanian. I AM grateful to Dr. M. F. Glaessner for arranging for me to study the extremely interest- ing collection of ammonites made by Dr. B. Daily in Bathurst Island in 1954, on the occasion of a National Geographic Society expedition to Melville Island (see text-fig. 1). Dr. Daily kindly prepared the locality map (text-fig. 1) and the stratigraphical columns (text-fig. 3). Not only are there several nevy species, but many specimens are exceptionally well preserved. New details can, in some cases, be observed in species known for over a cen- tury. The collection extends our knowledge of the geographical range of certain wide- spread Cenomanian species. The occurrence of the regionally endemic Chimbuites with typical Euomphaloceras allows us to date the former genus with certainty for the first time. The numbers given in the systematic descriptions are the register numbers in the collection of the Department of Geology, University of Adelaide. SYSTEMATIC DESCRIPTIONS Family hamitidae Hyatt Hamites (Hamites) (?) sp. Material. Two decaying pyritic fragments, F. 15880-1. Description. The two small fragments belong to the early stages of an open, irregularly coiled Hamites-Cke form. The section is nearly twice as high as wide (4 mm. by 2-5 mm.). The ribs are rather distant (four in a distance just greater than the whorl height), rounded, prorsiradiate and nearly as strong on the dorsum as on the venter. Occurrence. Both specimens were collected from a clilT, within 6 feet from the base of a measured section, at Numungumpi on the central south coast of Bathurst Island, in association with Parengonoceras (see below). Hamites (Stomohamites) simplex d ’Orbigny Plate 81, fig. \a-c 1842 Hamites simplex d’Orbigny, 550, pi. 134, figs. 12-14. 1956 Hamites simplex d’Orbigny; Sornay. Material. Five specimens, F. 15846-50. Palaeontology, Vol. 6, Part 4, 1963, pp. 597-614, pi. 81-89.) C 1713 R r 598 PALAEONTOLOGY, VOLUME 6 Description. Coiled probably in two more or less straight shafts after an initial open spiral. Whorl section oval to nearly circular. The more or less radial ribs are distant, high and narrow on the internal cast, but on the shell they are broader and lower and as wide as the interspaces. The aperture has the three prominent oblique ribs characteristic of Stomohamites. There are four and a half ribs in a distance corresponding to the (costal) whorl height. Remarks. Sornay (1956), refiguring d’Orbigny’s types, explains that the whorl section is not always so oval as indicated by d’Orbigny. His photographs of the lectotype and paratypes show also that, except on the bend, the ribs are more radial than in d ’Orbigny ’s figure. Taking these points into account, and the fact that the appearance of the ribs on the shell and on the internal cast differs, it is justifiable to refer the Bathurst Island speci- mens to d’Orbigny’s widespread species. The one figured in Plate 81, fig. \a-c is excep- tionally well preserved and, for the first time, allows a description of the ribbing on the shell itself, since previously figured specimens have all been internal casts. Occurrence. Four specimens come from the 3 feet immediately below the Tapara Bed at Meadinga, in company with the two new species of Acanthoceras described below. The fifth, F. 15846, comes from an isolated beach pebble at Moonkinu. In Europe, S. simplex occurs sparsely throughout the Upper Cenomanian, but it is not clear at what point in C W. WRIGHT: CRETACEOUS AMMONITES FROM BATHURST ISLAND 599 the Lower Cenomanian it evolved from the (rather similar) species of the genus that range up from the Upper Albian. Family baculitidae Meek Sciponoceras glaessneri sp. nov. Plate 81, figs. 2, 3 Types. The holotype is F. 15852, paratypes F. 15853-4 and about fifty other specimens. Description. The shell tapers gradually; the whorl section is elUptical, at first consider- ably, but later only slightly higher than wide (e.g. 8 by 6, 10 by 8, 12 by 10-5 mm.). Constrictions are frequent, approximately one in a distance equal to the larger diameter. On the shell, the constrictions are feeble on the dorsum and inner third of the side, then deeper and wider, becoming bordered by a distinct, gently rounded, posterior collar; on both dorsum and venter they are transverse and almost straight, giving an illusion of squareness to the shell; on the sides they are flexuous and oblique, at an angle of 45° at mid-flank where the obliquity is greatest. On the internal cast, the constrictions are much wider and more distinct throughout. Between the constrictions there are frequent irregular slight, but distinct, folds, on and between which are striae, all following about the same course as the constrictions, and only a little weaker on the cast than on the shell. No apertural border has been seen. The suture has moderately frilled squarish bipartite elements. Remarks. Cenomanian baculitids from the classic areas of Europe are known mainly from rather poorly preserved fragments of internal casts, usually more or less distorted; whereas the Bathurst Island material, though fragmentary, is mostly uncrushed and retains the shell beautifully preserved. The European material badly needs redescription. However, it is clear that the present specimens represent a new species, distinct from any described European forms. It differs primarily in its more frequent constrictions, nearly straight and transverse on dorsum and venter, its more distinct intermediate ribs, and its rather broadly elliptical whorl section. S. roto Cieslinski (1959, p. 39) has an almost circular whorl section at all stages, has one constriction in a length equal to about three diameters, and is almost smooth between constrictions. The Lower Cenomanian S. baculoides (Mantell), as interpreted by uncrushed specimens from the Isle of Wight [with the same close ventral ribbing near the aperture as MantelTs syntypes (1822, pi. 23, figs. 6, 7) and Sowerby’s syntypes of Baciilites obliquatus (a synonym) and the same type of apertural margin as in Sowerby’s figure (1828, pi. 592, fig. 2)], has a more compressed whorl section, relatively straight oblique constrictions strongly projected on the venter, where they form a narrow chevron, and one constriction in a distance equal to just under two major diameters. The later S. sp. nov. with a different aperture, as figured by Crick (1896, p. 77, text-fig.), is otherwise rather like baculoides, but has somewhat more sinuous and deeper constrictions. Considerable confusion exists about the nomenclature of the European Cenomanian baculitids. MantelTs and Sowerby’s types were rather small fragments. D’Orbigny (1842, pi. 138, figs. 6-11) gave a restored, idealized and, in some respects, obviously incorrect figure, based on fragments that are probably mostly true S. baculoides. In 1875, however, Geinitz (p. 281) established a new species, B. subbaculoides, to include, besides 600 PALAEONTOLOGY, VOLUME 6 a variety of ventral European specimens, d’Orbigny’s B. baculoides, which Geinitz held to be distinct from Mantell’s species. He believed the latter to be identical with the Upper Turonian form described by Fritsch as B.faujasi var. bohemica and now known as Sciponoceras bohemicwii ; evidently he confused the dorsal ribbing present near the aperture of baculoides with the regular ventral ribbing of bohemicum. In any case, the only specimen he figured as subbaculoides is probably the Lower Turonian Sciponoceras gracile Shumard. Occurrence. S. glaessneri occurs with the two new species of Acanthoceras described below, and is thus later than, at any rate, some of the Enghsh specimens of S. baculoides, which occur in the lower Scldoenbachia varians Zone with abundant Mantelliceras, and probably slightly earlier than English specimens of S. sp. nov. and S. roto which occur with typical species of Acanthoceras. The types listed are all from the beach platform outcrops at Meadinga on the south-east corner of Bathurst Island. Sciponoceras sp. Material. Several small fragments, the best, F. 15883, in the matrix of a Cliimbiiites (no. F. 158 16). Description. The section is very slightly higher than wide, the shell, and apparently the cast also, almost smooth, except for sparse feeble ventral folds on the shell, of which the stronger correspond to rather deep oblique constrictions on the cast. The distance between the constrictions is a little less than twice the major diameter. Remarks. This form is distinct from S. glaessneri, but the specimens are not good enough to indicate its relations with contemporary European species. Occurrence. In the loose beach boulders at Pouphmadourie, south-east coast of Bathurst Island. Family turrilitidae Hyatt Hypoturriiites gravesianus (d’Orbigny) Plate 81, fig. 5; text-fig. 2 1842 Tiarilites gravesianus d’Orbigny, p. 596, pi. 144, figs. 3-5. 1857 Tiirrilites gravesianus d’Orbigny; Sharpe, p. 62, pi. 25, fig. 7; pi. 26, figs. 15, 16. 1953 Hypoturriiites gravesianus (d’Orbigny); Dubourdieu, p. 44. Material. Five specimens, F. 15809, 15834-7. EXPLANATION OF PLATE 81 All figures are of natural size. Fig. 1. Hamites (Stoniohamites) simplex d’Orbigny; Cenomanian, Moonkinu. F. 15846. a. Side; b, venter; c, side view of cast taken from natural impression. Fig. 2. Sciponoceras glaessneri sp. nov. ; holotype, F. 1 5852. Cenomanian, Meadinga. a. Side; b, venter. Fig. 3. Sciponoceras glaessneri sp. nov.; paratype, F.15853. Internal cast, a. Side; b, venter. Fig. 4. Tiirrilites costatus Lamark; F. 15838. Cenomanian, between Moonkinu and Meadinga. Fig. 5. Hypoturriiites gravesianus (d’Orbigny); F. 15835. Cenomanian, between Meadinga and Toyun- gimpi. a. Side; b, top, showing spines. Fig. 6. Scaphites dailyi sp. nov.; holotype, F. 15806. Cenomanian, Moonkinu. a. Side; b, venter of shaft; c, venter of hook. i Palaeontology, Vol. 6 PLATE 81 WRIGHT, Cretaceous ammonites in# V . y 5»' C. W. WRIGHT: CRETACEOUS AMMONITES FROM BATHURST ISLAND 601 Description. This well-known species is characterized by an apical angle of about 30°, a pentagonal whorl section, ten to twelve, or rarely a few more, very large spines in a row along the middle of the exposed side, and rows of smaller tubercles, each four to five times as numerous as the large spines, close together at the base of the side. Only two rows of these smaller tubercles are generally visible on an internal cast, but the lower and larger ones each support two small chisel-ended tubercles on the test. The Australian material is interesting in that it includes external moulds of the spines, indicating their original shape and length (PI. 81, fig. 5a, b). Text-fig. 2 shows diagram- matically the whorl section, and the size and direction of the spines. Several specimens, e.g. F. 15809, have one or two more than the typical twelve to a whorl. TEXT-FIG. 2. Diagrammatic and restored whorl section of Hypotiinilites gravesiamis (d’Orbigny), Cenomanian, between Meadinga and Toyungimpi, showing the attitude and size of the spines. Remarks. The species is readily distinguished from H. tubercnlatus (Bose) by its greater apical angle and its fewer and larger major spines, as well as by the notably asymmetric first lateral saddle of the suture (see Sharpe, 1857, pi. 26, figs. 14, 15). Occurrence. H. gravesiamis occurs both with the two new species within and immediately below the Tapara Bed between Meadinga and Toyungimpi, and also (F. 15809) in one of the loose doggers found f mile east of Pouphmadourie. Typically, in Europe, it is a Lower Cenomanian species, but occasional members of some Lower Cenomanian species of Hypoturrilites do occur in the Upper Cenomanian. Turrilites costatus Lamarck Plate 81, fig. 4 1801 Turrilites costatus Lamarck, p. 102. 1842 Turrilites costatus Lamarck; d’Orbigny, p. 598, pi. 145. 1857 Turrilites costatus Lamarck; Sharpe, p. 66, pi. 27, figs. 1-5, 15, 16. Material. Eight specimens, F. 15838-45. Description. An acute angled species with a gently rounded outer face of the whorls, and with twenty-two to twenty-four well-rounded ribs, more or less oblique, particularly towards the aperture. The ribs are interrupted a little below the middle of the side and again near the base, so that there is an appearance of a rib, a large rounded tubercle, and a small weak tubercle. The Australian material is perfectly characteristic. 602 PALAEONTOLOGY, VOLUME 6 Occurrence. In and around the Tapara Bed between Moonkinu and Meadinga as well as on the beach platform just south of the Bathurst Island Mission. In Europe the species is most commonly found in the upper part of the Lower Cenomanian, below the horizon with abundant typical species of Acanthoceras. Family scaphitidae Meek Scaphites dailyi sp. nov. Plate 81, fig. 6«-c Material. The holotype, F. 15806, and two paratypes, F. 15807-8. Description. An involute, inflated, globose species (but most Scaphites species are even- tually found to include both inflated and compressed forms) with a short shaft and closely adpressed hook. An inward bulge at the beginning of the shaft covers the umbilicus of the spire. There are no tubercles. On the spire rather prominent primary ribs branch once or twice, giving rise to rather sharp well-spaced ribs that cross the flattened venter transversely. On the last part of the spire all the ribs weaken or disappear, and the last quarter whorl has only indistinct folds and flne striae. On the shaft and hook there are strong oblique primary ribs branching rather irregularly into two to four secondaries, with some intercalated ribs. Just before the constricted aperture the ribs again disappear. Affinities and differences. In size and general appearance S. dailyi compares well with S. striatus Mantell (1822, p. 119, pi. 22, figs. 9 and 11 only = lectotype here designated, BM no. C. 4799a), which is very close in turn to the Upper Albian 6'. simplex Jukes- Brown (cf. Spath, 1937, p. 505, text-fig. 177). However, the present form is readily distinguished from both these species by the bulge that covers the umbilicus, by the some- what coarser ribbing, by the smooth area on the latter part of the spire, and by the more branched ribs on the shaft and hook. The widespread S. obliquus Sowerby has no umbi- lical bulge and has more regularly spaced primaries on the shaft which branch at a regular distance from the umbilical margin into three to six very fine secondaries. The only previously described species with an umbilical bulge comparable with that of 5’. dailyi is S. basseae Collignon from the Cenomanian of Madagascar (Collignon, 1928) and Algeria (Sornay, 1955). S. basseae, however, has more or less regular even fine ribs, single or bifurcating, over the whole of the visible part of the spire, the shaft, and the hook. It is probable that S. dailyi is closely related to, and perhaps derived directly from, S. basseae. Oceurrence. S. dailyi occurs in the loose beach boulders with Euomphaloceras at Moon- kinu. It is therefore of middle to upper Upper Cenomanian in age. Family binneyitidae Reeside Borissiakoceras (?) sp. Plate 89, fig. 5 Material. A single crushed fragment, E. 15851. Description. So far as can be seen, this is a fragment of a small, high whorled, flat-sided, involute ammonite, with more or less flat venter and rounded shoulders. It has dense C. W. WRIGHT; CRETACEOUS AMMONITES FROM BATHURST ISLAND 603 feeble low falcate ribs. The shell is well preserved with a marked pink and green metallic lustre. Remarks. This specimen was at first determined doubtfully as a Falciferella, for in general build, ribbing and the peculiar lustre of the shell it shows considerable resemblance to Falciferella rnilbournei Casey, the type species from the English Middle Albian. Examina- tion of the figures of the Queensland Upper Albian Falciferella described by Brunn- schweiler (1959), however, suggested to me a connexion with certain Cenomanian species of Borissiakoceras from the United States. In fact, the Bathurst Island fragment com- pares well in all its preserved features with some specimens in my collection, given to me by the United States Geological Survey, from the base of the Cody Shale, Sheridan County, Wyoming (U.S.G.S. locality 21364), and identified as B. reesidei Morrow. This and other species of Borissiakoceras and its allies have recently been fully described by Cobban (1961). Study of this poor fragment from Bathurst Island has thus had the interesting result of linking the peculiar family Binneyitidae, ranging from Cenomanian to Santonian, whose origin was unknown, with the recently discovered Falciferella, which is known to be derived from the Haploceratacean family Aconeceratidae (see Casey, 1961, pp. 118, 1 19, 137-8). Indeed Falciferella is very close to Borissiakoceras', the suture of F. breadeni Brunnschweiler (1959, pi. 1, fig. 6), for example, resembles in a high degree that of B. reesidei Morrow (1935, pi. 50, fig. 5). Clearly Falciferella should be regarded as an early member of Binneyitidae, although F. rnilbournei Casey retains in simplified forms the Aconeceratine type of suture. Occurrence. Borissiakoceras (?) sp. is found with Acanihoceras tapara immediately below the Tapara Bed at Meadinga. Family hoplitidae Douville Chimbuites mirindowensis sp. nov. Plate 82, figs. 1, 2 Material. The holotype, F.1581 1, and eight paratypes, F. 158 10, 15812-16. Description. Moderately evolute, high whorled, widest at the umbilical shoulder, sides only slightly convex, converging to the broadly rounded and slightly flattened venter, umbilical wall steep with well-rounded shoulder. Ribs prorsiradiate, sinuous, well spaced, rather high and narrow at first, but becoming broader and flatter on later whorls, parti- cularly on the venter where they form broad projected folds. The primary ribs arise as low blunt bullae on the umbilical shoulder. They may remain single or branch at the bulla or higher up the side. Slightly weaker ribs are intercalated irregularly. The density of ribbing varies considerably, paratype F. 158 10, for example, having nearly half as many ribs again as the holotype. Owing to the irregular branching, the spacing of the ribs on the venter is uneven. In some specimens the shell is astonishingly well preserved. The surface is very finely rugose and has close set radial rows of shallow pits. Frequent crimped marks are visible, perhaps scars from the edge of the mantle. Remarks. C. mirindowensis resembles the type species C. sinuosocostatus Casey and Glaessner (in Glaessner, 1958) in many respects, for example in whorl section, sinuosity 604 PALAEONTOLOGY, VOLUME 6 of ribs and suture. It is distinguished chiefly by the ribbing. There is much less differentia- tion between primaries and secondaries in the present species. If a primary branches at the umbihcal bulla the rear rib is the stronger; this may branch again three-quarters of the way up the side, but if not it is slightly more prominent on the venter than its neighbours. Occurrence. All the specimens come from the loose boulders with Euomphaloceras on the beach between Mirindow and Mirialampi Points, F. 158 13 and F. 158 16 at Poupli- madourie, the remainder at Moonkinu, and are therefore of middle or upper Upper Cenomanian age. This suggests that C. sinuosocostatus Casey and Glaessner from New Guinea, which was originally recorded as an Aptian Deshayesites and was later described by analogy with related genera as probably Albian, is really Upper Cenomanian. Family engonoceratidae Hyatt Parengonoceras attenuation spinosum (Sommermeier) Plate 89, fig. 3o, b 1910 Kiiemiceras attemiatum-typiciim var. spinosa Sommermeier, p. 347, pi. 9, fig. 2. ? 1947 Kiiemiceras attemiatiim (Hyatt), Krause var. spinosa Sommermeier; Knechtel, p. 93, pi. 16, figs. 3, 4. 1956 Kiiemiceras atteuiiatnm spinosum Sommermeier; Benavides-Caceres, p. 449, pi. 46, figs. 11-12. Material. Two pyritic specimens, F. 15877 and 15878, and two small fragments, F. 15879 and 15882; four other fragments have decomposed. Description. Very involute. To a diameter of 20 to 25 mm. the venter is very narrow and the sides are flat and smooth, except for feeble umbilical swellings extending as faint traces of ribs leading to faint ventrolateral clavi. Thereafter the whorl thickness increases, the venter widens, the umbilical swellings become prominent spines (six to a whorl), the ribs and ventrolateral clavi become a httle more distinct and feeble swellings appear two-thirds of the way up the flank. The suture is characterized by numerous, more or less equal, elements; most of the saddles are squarish and bipartite, with bluntly rounded folioles; the lobes are rather more frilled than the saddles, have slender lobules and are generally tripartite. Remarks. P. attenuation is the smoothest and most compressed of a group of closely related species. P. attenuation spinosum is distinguished from several similar feebly orna- mented sub-species only by the precocious appearance and spinosity of the umbihcal tubercles. Occurrence. This subspecies and its associates are widespread in both hemispheres in the Douvilleiceras mamniiUatiim Zone at the top of the Lower Albian. The Bathurst Island specimens, from within 20 feet of the base of a cliff section at Numungumpi on the south coast, indicate this horizon; but, as they occur at the base of a bed of glau- EXPLANATION OF PLATE 82 All figures are of natural size. Both specimens from the Cenomanian of Moonkinu. Fig. 1. Chimbiiites mirindowensis sp. nov. ; holotype, F. 15811. a. Side; b, opposite side; c, venter. Fig. 2. Chimbuites mirindowensis sp. nov.; paratype, F. 158 10. Palaeontology, Vol. 6 PLATE 82 WRIGHT, Cretaceous ammonites C. W. WRIGHT: CRETACEOUS AMMONITES FROM BATHURST ISLAND 605 conite sand which contains Turonian CoUignoniceras cf. wooUgari 25 feet higher and as the Turonian rests conformably on the Cenomanian in bores nearby, they must be derived. Albian with pyritic fossils outcrops near Darwin, about 50 miles south of Bathurst Island. Family acanthoceratidae Flyatt Acanthoceras tapara sp. nov. Plate 83, Plate 84, figs. 1,2; Plate 85, fig. 2; Plate 86, fig. 1 Material. The holotype, F. 15830, and paratypes, F. 1 5822-33. Description. Moderately evolute, about one-third of the previous whorl being covered. The whorl section is slightly narrower to sUghtly wider than high (cf. PI. 84, fig. 16 and PI. 85, fig. 2b), rounded in intercostal, polygonal in costal section. There are twenty-six to twenty-nine ribs, irregularly long and short, the short branching from or intercalated between the primaries. The latter arise on the vertical umbilicial wall, are raised into a long umbilical bulla, at which the width of the whorl is greatest, and bulge slightly at mid flank, after which they become broad and flat until the prominent lower ventrolateral spine ; thereafter they are still more flattened until the clavate upper ventrolateral tubercle ; on the venter the ribs are almost invisible and there is only a weak and low elongated siphonal tubercle. The secondary ribs resemble the primary except that they first appear on the lower part of the flank with no umbilical tubercle. The holotype is a complete adult at a diameter of 290 mm. The other specimens are smaller, but are mostly incomplete. Remarks. This abundant form is one of those transitional species between Calycoceras of the newboldi group and Acanthoceras which are difficult to place. To a diameter from 60 to 80 mm. the feeble ribs, flat sides and venter and the clavate upper ventrolateral tubercles resemble the features of many species of Acanthoceras, but at greater diameters the ribs become more prominent and, in some individuals at least, the whorl section becomes more rounded, recalling Calycoceras. The persistence of the ventrolateral tubercles, however, with the broad weak ribs between the lower and upper ventrolateral tubercles and the flat venter with weak ribs, suggest reference to Acanthoceras rather than to Calycoceras, despite the aspect of the lateral ribbing on the later whorls. The number of ribs, the thinness and height of the lateral ribs at larger diameters and the very high rounded clavi of the upper ventrolateral row suffice to distinguish this species from other dQscxihQd Acanthoceras. ^to\iczk.2C s Ammonites morpheas {1^65, p\. 38, fig. 1) is probably the most closely related of described forms, although the figured specimen has only two, instead of three, ventral tubercles; however, the asymmetry of the whorl section, tubercles and suture show that the specimen is malformed and that the venter should be trituberculate. Ammonites morpheas is rather more densely ribbed and reverts at a smaller diameter to the rounded whorl section and ribbing without ventral tubercles reminiscent of many Calycoceras. Occurrence. Acanthoceras tapara is abundant in and around the Tapara Bed, which can be recognized in the coastal cliffs or on the beach platform over a great part of the distance between Moonkinu and Toyungumpi. Until the detailed phylogeny of Acantho- ceratidae is a good deal clearer than it is now, it is impossible to place a new species accurately in the succession on only intrinsic evidence. 606 PALAEONTOLOGY, VOLUME 6 The name is derived from the native name for the pearly nautilus. Dr. Daily reports that when he was collecting these pearly ammonites from the Tapara Bed the natives recognized their affinities with the pearly nautilus and said ‘ Him Tjapara ’. According to the local legend, when the moonman, also called Tjapara, dies, his head is washed up on the beach as the pearly nautilus. Acauthoceras mirialampiense sp. nov. Plate 84, fig. 3 ; Plate 85, fig. 1 Material. The holotype, F.15817, and paratypes, F.15818-21. Description. The whorl section is markedly wider than high, the sides flat and parallel. There are twenty to twenty-two slightly flexuous, nearly rectiradiate ribs. To a diameter of about 35 mm. the ribs are low and feeble, and more or less regularly long and short; the long ribs have fairly prominent umbilical bullae; all have slightly clavate upper ventrolateral tubercles; the siphonal tubercles are long but weak. After about 35 mm. diameter the ribs are mainly primaries, but occasional short secondaries are intercalated; the primaries arise well down the umbilical wall, soon become high and bar-like, are raised into a prominent unbilical tubercle, have a low bulge at mid-flank, bear a slightly clavate lower ventrolateral tubercle, then broaden rapidly and become low until the high rounded upper ventrolateral clavus. Affinities and differences. This species has fewer, less irregular and more distant ribs, and a squarer and more depressed whorl section than has A. tapara; the lower ventrolateral tubercles are as prominent though not so clavate as the upper ones. A. mirialampiense stands in much the same relationship to A. tapara as does Calycoceras spinosum (Koss- mat) to C. newboldi (Kossmat). It has some resemblance to A. deciduum Hyatt, but that species has a much higher and less inflated whorl section. Etheridge’s (1907, p. 18) rather poor specimens of Acanthoceras from Cape Gambier, Melville Island seem to belong to a species similar to, but not identical with, the present one. Occurrence. The same as A. tapara. Acanthoceras sp. cf. quadratum Crick Plate 86, fig. 2; Plate 87, fig. 1 Cf. 1907 Acanthus quadratum Crick, p. 192, pi. 13, figs. 2, la. Material. F. 15805. Description. The single, much worn, specimen is not well enough preserved for certain specific determination. It is very evolute, with a depressed whorl section and a broad flat EXPLANATION OF PLATE 83 Acanthoceras tapara sp. nov.; Holotype, F. 15830. Cenomanian, between Meadinga and Toyun- gimpi. Side X 0-75. (See also Plate 86, fig. 1.) EXPLANATION OF PLATE 84 All figures are of natural size. All specimens are from Cenomanian, between Meadinga and Toyun- gimpi. Fig. 1. Acanthoceras tapara sp. nov.; paratype, F. 15825. a. Side; b, venter. Fig. 2. Acanthoceras tapara sp. nov.; paratype, F. 15833. a. Side; b, venter. Fig. 3. Acanthoceras mirialampiense sp. nov.; paratype, F.15818. Side. Palaeontology, Vol. 6 PLATE 83 WRIGHT, Cretaceous ammonites Palaeontology, Vol. 6 PLATE 84 WRIGHT, Cretaceous ammonites C. W. WRIGHT: CRETACEOUS AMMONITES FROM BATHURST ISLAND 607 venter. There are ten or more primary ribs with large bullate umbilical tubercles, and on the inner whorls there are as many intercalated ribs that appear on the umbilical wall, but are not raised into umbilical tubercles. On the inner whorls each rib has more or less equally spaced and equal sized lower and upper ventrolateral and siphonal tubercles. On the outer whorls the lower and upper ventrolateral tubercles amalgamate to form a single large blunt tubercle. Remarks. So far as can be seen, this specimen closely resembles A. quadratum Crick from False Bay, Zululand, and similar specimens from England. Occurrence. The specimen comes from a beach pebble at Moonkinu on the south-east coast. Comparable English specimens come from a horizon below that of Euomphalo- ceras such as also occur in the Bathurst Island pebbles, but the A. cf. quadratum may be derived from a slightly different horizon from the Eumophaloceras. Equally, the species may well persist. Euomphaloceras cwmingtoni (Sharpe) Plate 88, fig. 2; Plate 89, fig. 1 1855 Ammonites cuimiiigtoni Sharpe, p. 35, pi. 15, fig. 2. 1864 Ammonites meridionalis Stoliczka, p. 76, pi. 41, fig. \a-c. 1897 Acanthoceras cnnningtoni var. cormita Kossmat, p. 11 (18), pi. 5 (16), fig. \a-c. 1907 Acanthoceras meridionate (Stoliczka); Pervinquiere, p. 278, pi. 15, figs. 2-6. 1937 Cinnhngtoniceras cwmingtoni (Sharpe); Collignon, p. 40. 1944 Cimningtoniceras Iwltkeri Erni, p. 470, pi. 11. 1951 Euomphaloceras enomphahmi (Sharpe); pars, Wright and Wright, p. 29. 1953 Acanthoceras ? eulessannm Stephenson, p. 201, pi. 47, fig. 5, pi. 48, figs. 3, 4. Material. F. 15800, 15803. Description. Very evolute, whorl section depressed, wider than high. Umbilical wall high and sloping. Sides vertical or converging slightly to the venter. Venter flat, or with flat siphonal area and sloping ventrolateral shoulders. Twelve to sixteen primary ribs rise on the umbilical slope, form large round to bullate tubercles at the umbilical shoulder, run radially up the side, and end in a prominent ventrolateral spine directed upwards or outwards. On the earlier whorls the venter is wide and more or less flat and is crossed by tuberculate ribs, among which may be rather deep constrictions. In most cases a broad flat rib runs from the large ventrolateral tubercle to a low clavate one, from which, in turn, run two rounded ribs, on each of which there is a rounded siphonal tubercle, in addition there are regularly intercalated ribs between the ventrolateral tubercles, and they bear three equally spaced rounded tubercles or a single siphonal one only. Occasionally, however, there are two rounded tubercles instead of the single clavate one, or other irregularities. On the outer whorls the siphonal tubercles, the transverse ribs and alternate ventral tubercles tend to disappear; the ventrolateral tubercles and the next inner ones tend to amalgamate, with the result that the costal section may show a depression over the siphon. The suture includes a broad squarish bifid first lateral saddle and a long narrow first lateral lobe extending as far back as the external lobe, more or less bifid and divided by a long narrow element of no great importance. 608 PALAEONTOLOGY, VOLUME 6 Remarks. Several species have been established for relatively large specimens of Euom- phaloceras that differ, if at all, only in minor detail of ventral ribbing and tuberculation, or in the shape and size of the ventrolateral horns on the outer whorl. Comparison of the Bathurst Island specimens with the holotype of E. cwmingtoni (Sharpe) and with the figures of meridionale Stoliczka, conmtum Kossmat and eulessanum Stephenson has helped to clarify the relationships of these forms. There is in fact very Httle material available of this group, not more, perhaps, than two or three dozen reasonably sized specimens. There is, therefore, the usual danger of allowing too little variabihty and of erecting too many species, each based on a handful of individuals. In this genus anyhow, I would take little account of minor variations in the size and direction of the ventrolateral horns on the outer whorl or in the prominence of tubercles in relation to ribs on the venter in the middle growth stages. Moreover, the specimens figured in the hterature show apparent differences due to varying types of matrix and to different techniques of illustration. The holotype of cwmingloni, at the earliest stage visible, shows three siphonal tubercles to each ventrolateral; corresponding to two of these siphonal tubercles, there is on each side an elongated tubercle, exactly as in the holotype of Kossmat’s var. cormita. Stephen- son’s eulessanum has the same type of ventral ornament, but has shghtly more ribs, and rather less prominent ventrolateral horns and umbilical tubercles on the outer whorls. In Stoliczka’s meridionale, however, there are two separate rounded tubercles in the place of the single elongate one and the transverse ribs are more prominent and persis- tent; Pervinquiere separated his specimens of this species from Stoliczka’s type on the ground that the latter had narrow ridges connecting the tubercles of the successive trans- verse rows, but this is clearly not an important feature and is perhaps merely a matter of preservation. Pervinquiere’s meridionale var. tuberculata has larger and more regularly spaced tubercles, fewer in proportion to the ventrolateral ones. I would regard meridionale Stoliczka, including africana Pervinquiere, and tuberculata Pervinquiere as subspecies of cwmingtoni Sharpe; neither cornuta Kossmat nor eulessanum Stephenson can, it seems to me, be separated from cwmingtoni sensu stricto. Cunningtoniceras holtkeri Erni from New Guinea was stated to differ from cwmingtoni in size and number of ribs, but Sharpe’s figure is reduced and at comparable diameters the holotype of the two species are similar in rib density as in other characters. There are two good specimens of this group from Bathurst Island. They are assignable to the subspecies c. cwmingtoni (F. 15800) and c. meridionale (F. 15803), although the EXPLANATION OF PLATE 85 All figures are of natural size. Both specimens are from Cenomanian between Meadinga and Toyungimpi. Fig. 1. Acanthocerus mirialampiense sp. nov., holotype, F.15817. a. Side; b, venter. Fig. 2. Acanthoceras tapara sp. nov; paratype, F. 15823. a. Side; b, aperture. EXPLANATION OF PLATE 86 Fig. 1. Acanthoceras tapara sp. nov.; holotype, F.15830. Cenomanian between Meadinga and Toyoungimpi. Apertural view, xO-75. (See also Plate 83.) Fig. 3. Acanthoceras sp. cf. qiiadratumCv\c]s',¥. 15805. Cenomanian, Moonkinu. Venter, natural size. Fig. 3. Collignoniceras sp. cf. wooUgari (Mantell); F. 15869. Turonian, Ticklitipinapitti. Side, natural size. Palaeontology, Vol. 6 PLATE 85 WRIGHT, Cretaceous ammonites Palaeontology, Vol. 6 PLATE 86 WRIGHT, Cretaceous ammonites C. W. WRIGHT: CRETACEOUS AMMONITES FROM BATHURST ISLAND 609 latter specimen shows ventral ornament of the c. cwmingtoui type for a short space, just as the enlarged horns begin to appear. In 1951 (p. 29) my brother and I considered Sharpe’s Ammonites cwmingtoui (with a diameter of about 150 mm.) to be merely the adult of his A. eiiomphalus (diameter 22 mm.). We based our opinion on several topotypes of euomphalus, larger than the holotype, which appeared to link that species with the still larger holotype of cwming- toni. Sinee then, we have studied a considerable number of euomphalus (over twenty in our own collection) and it has become clear that we were wrong to synonymize the two species. EuomphaJoceras euomphalus (Sharpe) is rather variable at all growth stages. Small examples (e.g. my coll. 2081 1, diameter 20 mm.) can be found which, in whorl section and ventral ornament, are probably indistinguishable from E. cunningtoni of similar size. At greater diameters, however, two distinct types of ventral ornament seem to be com- mon in topotype material. Some specimens have strong transverse ventral ribs, frequent ventral constrictions, weak siphonal tubercles, and high pointed ventral tubercles on some of the ribs. The larger of Pervinquiere’s two specimens of Acanthoceras giltairei (1907, pi. XV, fig. 9a, b) is just such an example. Other topotypes have rather strong siphonal tubercles and paired ventral tubercles of equal strength on all the ventral ribs, there being three or four tuberculate ventral ribs to each ventrolateral tubercle: this multituberculate stage ends rather suddenly at a diameter of about 40 mm. and the ventral ornament then becomes like that of the other type with only two transverse rows of tubercles to each ventrolateral tubercle. At a greater, but still undetermined, diameter alternate rows of ventral tubercles disappear and, thereafter, the ornament comprises large distant umbilical tubercles and simple ventrolateral horns. The suture in E. euomphalus is characterized in the earlier stages at least by a long, narrow first lateral saddle with converging, not parallel, sides and a very wide shallow first lateral lobe divided by a short stout element (cf. Pervinquiere, 1907, fig. 108 on p. 286, sub A. giltairei). After a diameter of about 40 mm. the suture gradually approxi- mates to that of E. cunningtoni (that is, the normal Acanthoceras type), but a fragment of euomphalus (my eoll. 23353) at a diameter of about 100 mm. still shows a mueh shal- lower and wider first lateral lobe than that of E. cunningtoni although at this size with the amalgamation of the ventrolateral and ventral tubercles into large horns the first lateral saddle has the typical massive squarish shape of E. cunningtoni and Acanthoceras. Occurrence. Both Bathurst Island specimens come from beaeh pebbles at Moonkinu, on the south-east coast. No exact dating within the Upper Cenomanian can be given for the rare specimens from England, France, North Africa and Texas, but all the evidence, stratigraphical and intrinsic, suggests that the speeies is later than the horizons with abundant typieal Acanthoceras and earlier than the horizon with E. euomphalus. Euomphaloceras Jonsdalei (Adkins) Plate 87, fig. 2; Plate 88, fig. I ; Plate 89, fig. 2 1928 Acanthoceras lonsdalei Adkins, p. 244, pi. 26, fig. 5, pi. 27, fig. 3. 1955 Euomphaloceras lonsdalei (Adkins); Stephenson, p. 62, pi. 6, figs. 6-20. Material. F. 1580 1 -2, 15804. 610 PALAEONTOLOGY, VOLUME 6 Description. Rather evolute, whorl section squarish, only a little wider than high, sides parallel or slightly convergent. Venter gently arched, becoming flat with age. Up to sixteen primary ribs appear at the umbilical seam and strengthen on the shoulder, where most of them form large bullae, although some are feebler than the rest. On the flanks, the ribs are prominent and radial. They bear large outwardly directed, more or less conical, lower ventrolateral tubercles, clavate upper ventrolateral ones and rounded to subdued clavate siphonal tubercles. To a diameter of about 80 mm. there are, on the venter, one or two distinct intercalated ribs to each primary, weakening with age ; most have rather feeble siphonal and some of them upper ventrolateral tubercles. At greater diameters the intercalated ribs become feebler and disappear, the siphonal tubercles on the primary ribs weaken and the lower and upper ventrolateral tubercles tend to amalgamate. Remarks. E. lonsdalei is distinguished from E. cwmingtoni, as defined above, by its less depressed whorl section, greater number of ribs and less pronounced multipfication of ribs and tubercles on the venter. E. eiiomphahis is even more extreme in these respects than E. cimningtoni. E. alvaradoense (Moreman) has higher whorls and feebler ribbing while its ventral tubercles disappear at a still earlier stage. Occurrence. The Bathurst Island specimens come from the beach pebbles at Moonkinu on the south-east coast. In Texas, E. lonsdalei occurs in the basal Eagle Ford beds, a little later than the Tarrant, in which occurs E. cwmingtoni. Family collignoniceratidae Wright & Wright Collignoniceras cf. wooUgari (Mantell) Plate 86, fig. 3 ; Plate 89, fig. 4 1822 Ammonites wooUgari Mantell, p. 197, pi. 21, fig. 16. 1855 Ammonites wooUgari Mantell; Sharpe, p. 27, pi. 11, figs. 1, 2. 1946 Prionotropis wooUgari Meek (? non Mantell); Haas, p. 150, pis. 11-16, text-figs. 1-14, 23-78, 80-83. 1959 Collignoniceras wooUgari (Mantell); Matsumoto, p. 105, text-figs. 55-57. Material. F. 15856-76. Description. The specimens are all more or less crushed internal casts or impressions, of inner whorls only; the largest has a diameter of only about 30 mm. They therefore show insufficient features to allow one to decide to which of the various subspecies described by Haas (1946) they belong, or, indeed, to be quite certain that they do not belong to some other species of the genus. EXPLANATION OF PLATE 87 All figures are of natural size. Both specimens are from the Cenomanian of Moonkinu. Fig. 1. Acanthoceras sp. cf. qnadratum Crick; F. 15805. Side. Fig. 2. Euomplialoceras lonsdalei (Adkins); F. 15802. a. Side; b, venter. EXPLANATION OF PLATE 88 All figures are of natural size. Both specimens are from the Cenomanian of Moonkinu. Fig. 1. Enomphaloceras lonsdalei (Adkins); F. 15801. Apertural view. Fig. 2. Enomphaloceras cimningtoni cwmingtoni (Sharpe); F. 15800. a. Side; h, venter; c, aperture. Palaeontology, Vol. 6 PLATE 87 WRIGHT, Cretaceous ammonites Palaeontology. Vol. 6 PLATE 88 WRIGHT, Cretaceous ammonites C. W. WRIGHT; CRETACEOUS AMMONITES FROM BATHURST ISLAND 611 A TEXT-FIG. 3. Stratigraphic columns, measured by Dr. B. Daily, 1954. A. Stratigraphic column of the Cretaceous sequence between Ticklitipinapitti and Pipiyanyamili Creek, Bathurst Island, Australia. All units except the lowermost 10 feet of glauconitic clay are found in the measured cliff section at Pulliamandera Creek. B. Stratigraphic column of the Cretaceous sequence exposed in coastal clilfs between Mirindow Point and the Bathurst Island Mission. That part of the section marked x-y was measured at Meadinga. The stratigraphic position of all fossils found only in situ and mentioned in the text are located on columns A and B. The fossils described from beach boulders and dated as middle Cenomanian are believed to be derived by wave action from beds immediately below the oldest known out- cropping beds which are located on beach platforms in the Moonkinu and Pouplimadourie areas. 612 PALAEONTOLOGY, VOLUME 6 They are all moderately evolute with a more or less rectangular whorl section and a high crenulate keel. The narrow, rather high, ribs are mostly of the same strength, but in parts of some specimens they are alternately strong and weak. They have fairly prominent thin tubercles on the umbilical margin, small, but usually distinct, lower ventrolateral and rather stronger ventrolateral tubercles. At varying diameters the fine close ribbing of the earlier whorls become coarser and more distant. Remarks. Haas (1946) has shown how wide is the variation of the ornament on the inner whorls of this species. Matsumoto (1959, p. 105), however, points out that some of Haas’s varieties may belong rather to C. liyatti (Stanton). Occurrence. Central south coast of Bathurst Island, F. 15856-64 from cliffs at Pullia- mandera, F. 15865-76 from cliffs at Ticklitipinapitti. C. woollgari is widespread in the northern hemisphere and characterizes a horizon in the middle part of the Turonian, above that of Mammites nodosoides. SUMMARY OF STRATIGRAPHICAL AND PALAEOGEOGRAPHICAL CONCLUSIONS Albian. The only fully determined Albian ammonite in the collection, Parengonoceras attenualum spinoswn (F. 15877-82, text-fig. 3), is one of a widespread group of forms characteristic of the top of the Lower Albian. These species of Parengonoceras are abun- dant in many Tethyan faunas, for example around the Mediterranean, in Iran, and in South America. Stray examples have even been found in England. The occurrence in Bathurst Island might be taken to indicate an open sea connexion north-westward, hnk- ing with the old world Tethys. Krause (1902) described from western Borneo some pseudoceratitic ammonites as Knemiceras pinax; his specimens, however, belong to an Upper Albian Engonoceras and were associated with a small Mortoniceratid. Cenomanian. Two distinct Cenomanian faunas are represented on Bathurst Island. That which occurs in situ in and near the Tapara Bed includes the following; Stomohamites simplex (d’Orb.) Sciponoceras glaessneri sp. nov. Hypotiirrilites gravesianus (d’Orb.) Turrilites costatus Lamarck. Borissiakoceras (?) sp. Acanthoceras tapara sp. nov. Acanthoceras mirialampiense sp. nov. EXPLANATION OF PLATE 89 All figures are of natural size. Fig. 1. Euomphaloceras cwmingtoni meridionale (Stoliczka); F. 15803. Cenomanian, Moonkinu. Apertural view. Fig. 2. Euomphaloceras lonsdalei (Adkins); F. 15801. Cenomanian, Moonkinu. Side. Fig. 3. Parengonoceras attenuation spinosum (Sommermeier) ; F. 15877. Derived Lower Albian, Numungumpi. a. Side; b, aperture. Fig. 4. Collignoniceras sp. cf. woollgari (Mantell); F. 15866. Turonian, Ticklitipinapitti. Fig. 5. Borissiakoceras (?) sp.\ F. 15881. Cenomanian, Meadinga. Palaeontology, Vol. 6 PLATE 89 WRIGHT, Cretaceous ammonites C. W. WRIGHT: CRETACEOUS AMMONITES EROM BATHURST ISLAND 613 Of the previously described species, none is of much value for accurate dating, although Twrilites costatus has a rather restricted range in the Lower Cenomanian, at any rate in England; it first appears, with Scaphites equalis Sowerby, above the horizon of abundant Mautelliceras mautelli (Sowerby), ventuoreuse Dienerand allies and seems to be converted to a successional species, T. acutus Passy, by the period of abundant Acanthoceros of the typical square-whorled type. The two new species of Acanthoceras from Bathurst Island are morphologically intermediate between Calycoceras of the newboJdi group and typical square-whorled Acanthoceras. C. newboldi and its allies, however, persist long after the first appearance of Acanthoceras, and too much stress cannot therefore be laid on the intrinsic evidence of these undescribed species for purposes of dating. I would have been inclined to place the Tapara Bed fauna at about the middle of the Cenomanian, but the stratigraphical evidence, such as it is, indicates that this fauna is slightly younger than that of the beach boulders. On this slender evidence the Tapara Bed fauna would be well up in the Upper Cenomanian. Despite the fact that there are three new species, the fauna has links with that of part of the Utatur Beds of Southern India and includes widespread species of Stomohamites and Tiirrilitidae. The fauna found in beach boulders includes: Sciponoceras sp. Hypotiirr Hites gravesianus (d’Orb.) Scaphites dailyi sp. nov. Chimbuites inirindowensis sp. nov. Acanthoceras cf. quadra turn Crick Euomphaloceras cunningtoni (Sharpe) EuomphaJoceras lonsdalei (Adkins) The two species of Euomphaloceras indicate a horizon about the middle of the Upper Cenomanian. They, the Hypoturrilites and the Acanthoceras, are widespread forms, sug- gesting open sea connexions, either north and east around the Pacific or north and west through the Tethys to Africa, Europe and beyond. Euomphaloceras cunningtoni is known from New Guinea and southern India, and comparable forms occur in Japan. Scaphites dailyi is related to a species that occurs in North Africa and Madagascar. Chimbuites, however, is otherwise known only from New Guinea and is apparently a regional endemic form. Turonian. The CoUignoniceras are of a type abundant and widespread in Europe and North America and known also in western Asia, South America and Japan. Thus again an open sea connexion northwards or north-westwards is suggested at this date. REFERENCES BRUNNSCHWEiLER, R. 1959. Ncw Aconcceratinae (Ammonoidea) from the Albian and Aptian of Australia. Bur. Miu. Res., Geol. & Geophysics, Australia, Bull. 54, 1-19, 1 pi. CASEY, R. 1954. Falciferella, a new genus of Gault ammonites, with a review of the family Acone- ceratidae in the British Cretaceous. Proc. Geol. Ass., Lond. 65, 262-77, pi. 7. 1961. The ammonoidea of the Lower Greensand. Palaeontogr. Soc. {Mouogr.} Pt. II, 45-118, pi. 11-25; Pt. HI, 119-216, pi. 26-35. and GLAESSNER, M. F. 1958. in: glaessner, m. f. New Cretaceous Fossils from New Guinea. Rec. S. Australian Mus. 13, 199-226, pi. 24-26. s s C 1713 614 PALAEONTOLOGY, VOLUME 6 ciESLiNSKi, s. 1959. The Albian and Cenomanian in the northern periphery of the Swiety Kryz Moun- tains. Inst. Geol. Prace, Warsaw, 28, 1-95, 8 pi. (Polish with English summary.) COBBAN, w. A. 1961. The ammonite family Binneyitidae in the western interior of the United States. J. Paleont. 35, 737-58, pi. 87-89. COLLIGNON, M. 1928. Paleontologie de Madagascar. XV. Les Cephalopodes du Cenomanien pyriteux de Diego Suarez. Aim. Paleont. 17, 139-60, pi. 15(1)-19(5). CRICK, G. c. 1896. On the aperture of a baculite from the Lower Chalk of Chardstock, Somerset. Proc. Malac. Soc. 2, 77-80. DACQUE, E. 1939. Die Fauna der Regensburg-Kelheimer Oberkreide. Ab/i. Paver. Akad. Wiss. (Miinchen), N.F. 45, 1-218, 17 pi. ERNi, A. 1944. Fin Cenoman-Ammonit, Ciinningtoniceras holtkeri nov. spec, aus Neuguinea, nebst Bemerkungen liber einige andere Fossilien von dieser Insel. Eclog. geol. Helv. 37, 468-75, pi. 11. ETHERIDGE, R. (fil.). 1907. Official contributions to the Palaeontology of South Australia No. 22. The Cretaceous fossils of Maclear Creek, Cape Gambler, Melville Island. Suppl. to Parly. Pap. No. 55, of 1906, Adelaide. GEINITZ, H. B. 1875. Das Elhthalgebirge in Sachsen. Erster Theil. Der imtere Qiiader. Pt. VIII, 277- 319, pi. 61-67, Cassel. HAAS, o. 1946. Intraspecific variation in, and ontogeny of, Prionotropis woollgari and Prionocylus wyomingensis. Bull. Ainer. Mas., Nat. Hist. 86, Art. 4, 141-224, pi. 11-24. KOSSMAT, F. 1895-8. Untersuchungen liber die Slidinische Kreideformation. Beitr. Pal. Geol. Ost. Ung. ii. O/-. 9 (1895), 97-203 (1-107), pi. 15-25 (1-11); 11 (1897), 1-46 ( 108-53), pi. 1-8 (12-19); 12 (1898), 89-152 (154-217), pi. 14-19 (20-25). KRAUSE, p. G. 1902. Die Fauna der Kreide von Temojah in West-Borneo. Saintnl. Geol. Reiclis-Miis. Leiden (1), 7, 1-28, pi. 1-2. MANTELL, G. 1822. Fosslls of the Soiitli Downs. London, 1-320, 43 pi. MATSUMOTO, T. 1959. Upper Cretaceous ammonites of California. Part II. Mem. Fac. Sci. Kyushu Univ., D, Geol., Spec. Vol. 1, 1-172, 41 pi. MORROW, A. L. 1935. Cephalopods from the Upper Cretaceous of Kansas. J. Paleont. 9, 463-73, pi. 49-53. ORBiGNY, A. d’. 1840-2. Paleoiitologle frangaise; Terrains cretaces, /, Cephalopodes. Paris, 1-662, 148 pi. PERViNQUiERE, L. 1907. Etudes de Paleontologie tunisienne; I. Cephalopodes des terrains secondaires. Carte geol. Timisie, 1-438, 27 pi. SHARPE, D. 1853-7. Description of the fossil remains of Mollusca found in the Chalk of England. Palaeontogr. Soc. [Monogr.\ 1-68, 27 pi. SORNAY, J. 1955. Ammonites nouvelles du Cretace de la Region des Monts du Mellegue (Constantine). Serv. Carte geol. Algerie, Mem. Pal. 18, 1-40, 2 pi. • 1956. Hamites simplex d'Orhigny. Pal. Universalis, 'n.s., no. 18. SOWERBY, J. DE c. 1823-46. The Mineral Conchology of Great Britain. London, vol. 4, pi. 384 to vol. 7. STOLiczKA, F. 1864-6. Ammonitidae, with revision of the Nautilidae. In; The fossil Cephalopoda of the Cretaceous Rocks of Southern India. Pal. Indica, sers. 1 and 3, 1, 41-216, pi. 26-94. WRIGHT, c. w. 1957. In: arkell, w. j. et al. Treatise on Invertebrate Paleontology, Pt. L, Mollusca 4, Cephalopoda, Ammonoidea. Kansas, 1-490, 558 figs. and WRIGHT, E. V. 1951. A Survey of the fossil Cephalopoda of the Chalk of Great Britain. Palaeontogr. Soc. [Monogr.], 1-40. Manuscript received 20 December 1962 C. W. WRIGHT 37 Phillimore Gardens, London, W. 8 GROWTH GRADIENTS AMONG FOSSIL MONOTREMES AND MARSUPIALS by N. G. STEPHENSON Abstract. Although a few groups appear to be characterized by relative stability of size, growth gradients, whether isometric or allometric, have played an important role in the evolution of Australian monotremes and marsupials. Intra-specific size reduction, resulting from isometric change, has been a widespread phenomenon from Pleistocene to Recent. Its implications were not recognized by early taxonomists, whose undue emphasis on minor size dififerences as a criterion of speciation, and even failure to compare fossil and recent forms, have led to nomenclatorial confusion in some families. Further study of the available fossil and recent material in such groups suggests cases of conspecificity, and the consequent priority of either a palaeospecific or neospecific name. In certain marsupials such as the Diprotodontids and the wombats, there is evidence of a trans-specific increase in size, involving allometric as well as isometric changes, and leading to gigantism and subsequent extinction. Evidence of the progressive stages by which such gigantism was achieved in the Diprododontinae is given, and a new genus and species of this subfamily of the Diprotodontidae is described. Growth gradients, whether isometric or allometric, may be evidenced in phytogeny as well as in ontogeny. In the phytogeny of mammals this phenomenon has long been recognized, and a progressive phyletic increase in body size is known to have occurred at some stage in the ancestry of most eutherian orders. Similarly, the earliest fossil mar- supials are relatively small types, and gigantism, which reached its peak in the Pleisto- cene, was not apparent at first. Such samples of progressively increasing body size in lines of descent are commonly referred to under Cope's Rule and are numerous. On the other hand, there are few quoted cases of a graded trans-specific decrease of body size. The rarity of such broad-scale dwarfism gives a false impression of the role size reduction has played in mammalian evolution. Although on a lesser scale, and at an intra-specific level, the phenomenon of size reduction is very much in evidence when one compares Pleistocene and Recent mammals. As Hooijer (1951) has stated, it is fairly common to find Pleistocene animals, in all parts of the globe, both on continents and islands, larger than the living individuals belonging to the same species. No extinction nor even migration is involved; it is evolution in situ. The means and modes of the various metrical characters are shifting in the course of time, thus producing temporal dines or chronoclines. The quaternary decrease in the intraspecific size of many mammals is little understood. The size decrease is quite appreciable, and may be as much as 20-25 per cent. Such isometric changes occurring within one and the same species produce many problems in the equating of fossil and recent forms and in the sorting out of geographical and temporal races. Whatever the factor or factors involved, palaeontologists giving first descriptions of certain Australian fossils last century and in the first half of this century understood little, if anything, of this phenomenon. They commonly ascribed to distinct species fossil forms which merely exhibited slight size differences, but which were other- wise morphologically indistinguishable from living forms at the specific level. Any systematic study of the present-day Australian monotremes and marsupials should [Palaeontology, Vol. 6, Part 4, 1963, pp. 615-24, pi. 90. 616 PALAEONTOLOGY, VOLUME 6 therefore take due cognizance of the validity of various fossil species that have been described. If growth gradients are isometric, conspecificity may be recognized and taxo- nomic revision may be required. In view of these problems, the following account of selected monotremes and mar- supials is provided and suggestions concerning their generic and specific nomenclature are made. Failure in the past to equate fossil and recent species, when this would have been possible merely by using the available data, has in itself left a legacy of taxonomic problems. Article 1, International Code (1961), makes it quite clear that zoological nomenclature is the system of scientific names applied to taxonomic units of animals known to occur in nature, whether living or extinct. This study is based on a comparison of both fossil and recent material in the British Museum (Natural History) and in various Australian museums and university depart- ments. In some instances, material is reidentified, and in one case a description is given of a new genus and species. SIZE CRITERION IN TAXONOMY The taxonomic confusion which may arise, even at the generic level, when size is used as a criterion, is exemplified by certain monotreme species of the family Tachyglossidae. The genus Tachyglossus was first designated in 1811 by Illiger. This is the oldest generic name available for the spiny anteater of the Australian mainland, Tasmania, and New Guinea, because Cuvier’s (1798) Echidna, which is preserved in the vernacular, was preoccupied by Forster (1788) for a genus of eels within the family Muraenidae (Echidnidae). By the end of 1876 Tachyglossus was the valid generic name not only for mainland and Tasmanian species of the spiny anteater, but also for a new species from New Guinea, which Peters and Doria had described in that year under the name of Tachyglossus bruijnii. This generic name should also have been applied by Kreflft when he described the first fossil species in 1868. Krefft’s material, the proximal end of a humerus, was scanty and was described under the name Echidna owenii. In 1884 Owen examined the cast of a more perfect example of a humerus, which he described under the name Echidna ramsayi, but which is now regarded as being identical with Krefft’s species. This first fossil species, of which more material has subsequently been found, is still more correctly known as Tachyglossus owenii (Krefft, 1868). In 1877 Gill, in referring to the newly described New Guinea form, wrote; ‘This has very lately (December 3, 1876) been described by Messrs. W. Peters and G. Doria as a new species of the genus Tachyglossus, under the name T. bruijnii. It nevertheless differs markedly from the T. hysirix {= T. aculeatus) and T. setosus of Australia in the much more elongated and nearly uniformly or very gradually attenuated and decurved rostrum, as well as in the contour of the palate, etc. So great are these differences that . . . the newly discovered form may therefore be appropriately contrasted under the name Zaglossus bruijnii, with the previously known Tachyglossus hystrix and Tachyglossus setosus.' Although these relatively minor morphological differences of rostrum curvature and palate contour might well have been questioned as hardly constituting generic distinc- tiveness, Gill’s proposal of a separate generic name for the large New Guinea species N. G. STEPHENSON: FOSSIL MONOTREMES AND MARSUPIALS 617 has been generally accepted by subsequent writers. For example, Laurie and Hill (1954) list, apart from TachygJossiis knvesii Ramsay, which occurs in south-east New Guinea, three species of the genus Zaglossus Gill. These are Z. bruijnii (Peters and Doria), Z. bartoni (Thomas), and Z. bubuensis Laurie, and various subspecies. Whether, in 1877, Gill was aware of a fossil species, then already described, which was even larger than this largest of living species with which he was concerned, is not known. In proposing a new generic name, he made no attempt to accommodate the fossil species. Apart from large size, in which it exceeded Z. bruijnii, other diagnostic features of generic significance to Gill were, and still are, unknown in T. oweuii. Because of its size, however, this fossil form probably had a proportionally long beak with which to reach the ground. Whatever the justification of Gill’s description of the species bruijnii under the new generic name of Zaglossus, which in any case had priority over Acanthoglossus (= Pro- echidna) suggested by Gervais (November 1877), there seems to be little validity in the reference of later fossil species, in none of which are details of rostrum and palate suffi- ciently well known, if at all, to any genus other than TachygJossus (Illiger, 1811). Dun (1895) described an imperfect skull and atlas vertebra of a fossil species from Gulgong, New South Wales, under the name Echidna (Proechidna) robusta. He claimed that this species was much larger and more robust than the living forms, and also than T. Owenii. Dun further stated that these vertebral and cranial remains seemed to show a greater resemblance to Proechidna than Echidna, although he commented on the uncer- tainty of Oldfield Thomas (1888) as to the distinction of these two genera. It is clear that Dun was very uncertain as to how this new fossil species should be generically designated. On the evidence available. Dun’s species should more correctly be referred to as Tachy- glossus robust us (Dun). Further confusion in the fossil nomenclature was caused in 1914 when Glaubert described a fossil species from material collected at the Mammoth Cave, one of a series of limestone caverns in Western Australia. Glaubert’s material consisted of an atlas vertebra, clavicles and episternum, pelvic girdle, two femora, a tibia and a radius, and was referred by him to a new species, Zaglossus (Proechidna) hacketti. Again, no adequate reason was given for assigning this species to a genus other than TachygJossus, and the criterion of size was inadequate. The other minor diflerences in limb bones and girdles observed by Glaubert appear merely to mark distinction at the sub-specific or specific level. One important feature which distinguishes the large New Guinea anteater from other living species, including TachygJossus Jawesii Ramsay, from Port Moresby, is the reduced condition of its digits. There are normally only three claws on each limb, but the reduced first and fifth digits, both before and behind, are represented in the skeleton by phalanges and, according to Oldfield Thomas (1888), in some cases by functionally developed claws. However variable in extent it may be, this reduction of first and fifth digits of fore and hind limbs in the so-called three-toed anteater, ZagJossus, seems a more satisfactory basis for generic or subgeneric distinction amongst living forms, if such is desired, than any other feature so far proposed. Nevertheless, it was a character not commented upon by Gill (1877) in his original claim of generic status for ZagJossus bruijnii. There is, furthermore, no information as to whether digital reduction is solely a recent trend characteristic of the New Guinea species, or whether it had already been embarked upon 618 PALAEONTOLOGY, VOLUME 6 by any of the fossil species. Under these circumstances it seems preferable at present to sustain the fossil forms as species of the genus Tachyglossus. PRIORITY OF A P AL AEOSPECIFI C NAME In 1838 Richard Owen described a fossil marsupial from the Wellington Caves, New South Wales, under the name of Dasyiirus laniarius. He stated that it resembled the Tasmanian Devil, a species now confined to Tasmania, but that it differed in being one- third larger, and in having the canines, or laniaries, of proportionately larger size. At this stage, no other criterion than that of size was used. By 1877 Owen (pp. 105-6) was recognizing his fossil species as belonging to the genus Sarcophilus Cuvier and Geoffroy Saint-Hilaire, 1837. Some of the material from the Wellington Caves he assigned to the same species as the existing Tasmanian form, but other specimens from the same locality he included in his fossil species, Sarcophilus laniarius (Owen, 1838). With regard to these latter specimens, he stated that: ‘Besides the difference in size, the following modifications of structure are noticeable. The larger and deeper digital pit on the inner or palatal side of the penultimate upper molar; the better and broader ossification of the medial border of the palatal vacuity. In the under jaw the relatively broader and deeper symphysis; the wider interval between the two predental foramina, and the more backward position of the hinder one, beneath the fore part of the antepenultimate molar. ’ Lydekker (1887) did not question the validity of Owen’s defined fossil species. He stated that Sarcophilus laniarius ‘presents a considerable excess in size’ over the existing Tas- manian form, ‘ but may probably be regarded merely as the ancestral form of the latter, which, from inhabiting a continental area, attained superior dimensions’. Lydekker further suggested that the smaller specimens from the Wellington Caves, assigned by Owen to the same species as the existing form, were probably females of the fossil species. Lydekker's measurements for Sarcophilus laniarius have been checked, e.g., 5-6 cm. for the length of the check-teeth series in specimen B.M. Cat. No. 42555; 3-95 cm. for the length of space occupied by the three teeth in specimen B.M. Cat. No. 42559. The corresponding measurements of a male specimen of the recent Tasmanian species are 4-9 cm. and 3-6 cm. respectively. Such comparative measurements do not support Lydekker’s contention of a fossil species presenting ‘a considerable excess in size’ over the existing Tasmanian form. A comparison of recent and fossil specimens in the collections of the British Museum, including Owen’s type material, suggests that there is insufficient evidence for the recogni- tion of a fossil species as distinct from the existing form. In nomenclatorial revision, the palaeospecific name, Sarcophilus laniarius (Owen, 1838) would take precedence over S. harrisii (Boitard, 1841) if, as seems to be the case, S. ursinus (Harris, 1808) is already invahdated as a name for the living Tasmanian Devil. Certain corrections to the legends of original illustrations should perhaps be noted here. In Plate 31, accompanying Owen’s letter which is pubhshed in Mitchell’s (1838) Three Expeditions into the Interior of Eastern Australia, Fig. 6 is described as the ‘left ramus, lower jaw, with last grinders’, and Fig. 7 as ‘anterior part of the right ramus of lower jaw’. Both are ascribed to Sareophilus laniarius, although Owen expresses doubt, N. G. STEPHENSON: FOSSIL MONOTREMES AND MARSUPIALS 619 because of the spacing of teeth, as to whether the specimen illustrated in Fig. 7 is really the lower jaw of iS. laniarius, or of some extinct marsupial carnivore of an allied but distinct species. By 1845 Owen had seen a specimen of Thylacinus spelaeus, and conse- quently revised his identification of the anterior portion of the right lower jaw, illustrated in Fig. 7, Plate 31 (1838). This is now ascribed to a fossil species of Thylacinus, namely T. spelaeus. Owen ’s correction may be carried further to include the specimen illustrated in Fig. 6 of the same plate. This is also a portion of the lower jaw of Owen’s T. spelaeus, namely, a part of the left mandibular ramus, viewed from the outer side, and illustrating the last tooth in the jaw. Probably the portions of the left and right lower jaws, illustrated in Figs. 6 and 7, both belong to the same animal. In Fig. 2, Plate 32, Owen illustrates the smaller bones in the foot of a Dasyurus, by which name Sarcophilus was at that time known to him. These small bones have now been cleared of matrix and some of them have been reassembled. These include what are undoubtedly the second, third, fourth, and fifth right metacarpals, the proximal phalanx of the fourth digit, and a carpal bone. The bones are clearly those of a small wombat. PRIORITY OF A NEOSPECIFIC NAME The Thylacine of Australia, commonly known as the Tasmanian or Marsupial Wolf, or Tiger, was first described by Flarris (1808) under the name of Diclelphis cynocephala. The generic name Thylacinus was established by Temminck (1827) and the Thylacine, which recently has been restricted to Tasmania, and which may already be extinct there, is known as Thylacinus cynocephalus (Harris). Despite the restriction of the living Thylacine to Tasmania, fossil material of this genus has been recorded from various localities on the Australian mainland. Owen (1845) distinguished fossil material under a new specific name, Thylacinus spelaeus. In 1877 Owen {Extinct Mammals of Australia, p. 106) used the name T. major for the first time. According to Lydekker (1887), this name was given inadvertently for T. spelaeus', the mandible represented in PI. V, Fig. 8, being drawn from the last three molars of Sarcophilus laniarius added to the hinder part of a mandible of T. cynocephalus. Owen distinguished T. spelaeus on the following criteria: in one specimen (No. 1548, Cat. Royal College of Surgeons, 1845), the depth of the lower jaw below the first pre- molar was nine-twelfths of an inch, whereas that of the corresponding part of the jaw in the existing Thylacine was seven-twelfths of an inch; in another specimen (No. 1549, Cat. Royal College of Surgeons, 1 845), the penultimate molar of the right side of the lower jaw had a small accessory cusp on the inner side of the large middle compressed cusp; an upper canine, erroneously described in 1877 as T. major, was proportionately larger in comparison with the lower one than it is in T. cynocephalus. Such slight differences and size distinctions of individual specimens used by Owen appear to be too inadequate as a basis for full specific recognition. Under these circum- stances, it seems best to recognize only one species of Thylacinus, namely, the neo- species T. cynocephalus (Harris). SIZE STABILITY IN PHYTOGENY In 1900 Baldwin Spencer described Wynyardia a fossil marsupial from marine sediments near Wynyard, northern Tasmania. This was the first and for many years the 620 PALAEONTOLOGY, VOLUME 6 only reputedly Tertiary marsupial from Australasia, and is still the oldest Tertiary mar- supial of more precisely known age from this region (Woods, 1962). In the complete absence of teeth and of bones of the feet of this fossil, Spencer was led to the conclusion that it was representative of a new extinct series of forms which were more nearly allied to ancestral Polyprotodonts than are any of the existing Dipro- todont forms. He regarded Wynyardia as intermediate between the former and the latter, and as indicative of a stage in the development of Australian marsupials when the ancestors of recent Diprotodontia were beginning to diverge from their original Poly- protodontid stock. In 1930 Wood Jones re-examined Wynyardia and, in reviewing the skull, observed that there were singularly few noteworthy characters to differentiate it from that of a modern phalanger. In general outline, the proportions of the skull preserved in the specimen resembled very strikingly the similar parts of a rather large skull of the Tas- manian form of the phalanger, Trichosurus vulpecidus. From his re-examination of parts of the postcranial skeleton, Wood Jones deduced that Wynyardia was a sturdily built creature, considerably heavier than the living Trichosurus, and probably differing from that animal in habit and in bodily poise. His illustrations, however, show only slight differences, and, on the whole, the evidence points to relatively little change from the Tertiary to Recent, involving a size decrease so slight as to be possibly within the limits of subspecificity. Certainly, there is no evidence of phylogenetic gigantism in this or any other form that is comparably arboreal. Stirton (1957) described a new genus and species, Perikoala palankarinnica, possibly Oligocene, from the Palankarinna fauna, east of Lake Eyre, South Australia. This early phascolarctine is of roughly comparable size to the modern Phascolarclos and to Trichosurus. Again, in this ancestry, there is evidence only of size stability. Owen (1838, p. 361) described a fossil phalanger from the Wellington Caves, but recog- nized that a comparison with the bones of Trichosurus vulpeculus, known to him as Phalanagista vulpina, but not available in his osteological collections, would be neces- sary to establish whether or not his fossil material was specifically different from the modern form. Owen’s original material is available for examination at the British Museum (Natural History) and appears to be specifically identical with recent skulls of T. vulpeculus. Gill (1957) concluded that Wynyardia bassiana is a valid fossil, having the same age as the TurriteUa bed in which it was found, and with a fluorine index of nearly 500 times as great as that of Trichosurus vulpeculus living in the same area in modern times. The issue of the inclusion of Wynyardia in the Phalangeridae seems settled. Rather the main question today is whether its generic distinction from Trichosurus is really justified. With one incomplete fossil skeleton in which only part of the skull is represented, and that completely lacking teeth, this question remains unanswered. GRADIENTS LEADING TO GIGANTISM IN THE DIPROTODONTINAE Richard Owen (1838) gave the first description of Diprotodon optatus, which fossil species is now known as the largest Australian marsupial. Subsequently, three further species, Diprotodon minor Huxley, 1862, D. longiceps McCoy, 1876, and D. bennetti Owen, 1877 were described, although Owen (1877) himself expressed the view that N. G. STEPHENSON; FOSSIL MONOTREMES AND MARSUPIALS 621 D. minor was founded on teeth of a species of Nototheriiim. Whatever the validity of these species of Diprotodon, they certainly indicate forms which were smaller, but at the same time reasonably close to D. optatus in size. De Vis (1888) commented on the fact that D. minor was but a fourth smaller than D. optatus, and from the measurements given by McCoy (1876) and Owen (1877) D. longiceps and D. bennetti were apparently not less than this. No precise information has been available regarding the progressive stages whereby gigantism could have been achieved in the Diprotodontinae, nor has irrefutable evidence existed of gradients leading to Diprotodon optatus from very much smaller forms. With the exception of the notothere MeniscoJophus mawsoni Stirton, 1955, the fragmentary remains of smaller diprotodontids so far found have not been referred beyond the sub- familial level. None appear to have been ascribed to the subfamily Diprotodontinae. Owen (1877, p. 274) stressed the fact that the first incisors of Diprotodon were scalpri- form, ever-growing tusks, with enamel continued to the widely open bases. On the other hand, the incisors of Nototheriiim were teeth of limited growth, each with a well-defined, partly enamelled crown, and a non-enamelled root. The incisors are therefore very important in enabling a diprotodontine to be distinguished from a notothere. There is now evidence that gigantism has been progressive in the Diprotodontinae. Stephenson has given evidence (in press) to support the view that the scalpriform upper incisors known as Sceparnodon ramsayi actually belong to a small member of the Diprotodontinae, and that they were erroneously restored on to the front part of a skull of Phascolomis gigas by Stirling (1913). An even smaller diprotodontine was found at Wellington Caves by Mr. J. Mahoney, University of Sydney, in 1954. This specimen consists of part of the left premaxilla with the complete first and second incisors im- planted, and also part of the alveolus of the third incisor of the left side (Plate 90, figs. 1 and 4). This fragment is indicative of a diprotodontine related to Diprotodon, and belonged to an animal which appears to be the smallest of the series yet to be discovered. In contrast to Diprotodon optatus, which in bulk fully equalled a large rhinoceros, this animal would have been about the size of a Shetland pony. The first incisor is strongly curved and increases in width from its tip to base. The length of the tooth is approximately 5 inches, but of this only about 1 inch would project beyond the alveolus. The enamel of this incisor is continued to the base of the tooth and bears the same relationships to the dentine as in other diprotodontine teeth. The worn surface at the tip bears a transverse notch which no doubt was caused by the lower incisor. The second incisor is the small, typically peg-like tooth found in such a position in other diprotodontids, but it is proportionately larger than the corresponding tooth in Diprotodon. It is well worn, but remarkable in exhibiting on the inner and outer sides of its crown small patches of enamel. This enamel does not extend back to the root, and is limited to the sides of the crown. In possessing enamel, this tooth resembles those of nototheres in which enamel appears normally to be present on the crowns of the second and third upper incisors. There is no previously recorded case of enamel occurring on the diprotodontine second and third incisors which, as in nototheres, are also teeth of limited growth. In describing a specimen of Diprotodon optatus, Owen (1877, p. 199) stated that probably, by reason of the age of the individual, and the extent of tooth worn away, the original enamelled crown had gone, and both and I3 were represented only 622 PALAEONTOLOGY, VOLUME 6 by their cylindrical cement-covered portion. However, there is no evidence from the specimens in the Australian Museum that enamel is ever developed on the second and third upper incisors of Diprotodon. The reidentification of Sceparnodon ramsayi Owen as a member of the Diprotodon series makes it desirable that this more recently discovered diprotodontine, which is even further removed from Diprotodon than is Sceparnodon, should be ascribed to a new genus and species. The type specimen is not unique. Gill (1955) appears to have erroneously ascribed a portion of an identical incisor (Plate 90, fig. 2) to the very much larger diprotodontine, Diprotodon minor Huxley. This specimen, described by Gill, is a right upper incisor and is a very much worn tooth. In the Queensland Museum there is a left upper incisor, F.1651, which is only slightly worn. There are also two casts of an identical upper incisor, L.1292 and F.5056 (Plate 90, fig. 3), in the Australian Museum collection, although there is no trace of the original which had been in an old collection in Sydney. SYSTEMATIC DESCRIPTION Genus diarcodon gen. nov. Type species. Diarcodon parvus sp. nov. The diagnostic characters of tne genus are those of the genotypic species until other species have been described. Diarcodon parvus sp. nov. Plate 90, figs. 1-4 Holotype. Left premaxilla, with first and second upper incisors implanted and entire. Australian Museum Collection, Reg. No. F. 50099. Paratype. Portion of a well-worn, right upper incisor. National Museum of Victoria, Reg. No. P.16155. Diagnosis. First upper incisor small, flattened, arcuate; tapering slightly from root towards tip, though more markedly so beyond the alveolus ; dorsal outline slightly con- cave; root open; enamel extending back to root, covering dorsal and part of lateral surfaces. Second upper incisor peg-like, with enamel on crown; small, but strongly developed and proportionately larger in relation to first upper incisor than in Diprotodon. Description. Total length of I^, 120 mm.; greatest width, at root, 30 mm.; depth varying from 15 mm. near root to 10 mm. near tip; length of tooth normally exposed beyond EXPLANATION OF PLATE 90 Fig. 1. Diarcodon parvus, gen. et sp. nov. Lateral view of holotype showing first upper incisor, second incisor with patch of enamel, and part of alveolus of third incisor. Australian Museum Collection, Reg. No. F.50099. xL2. Fig. 2a, b. First upper incisor (right) of Diarcodon parvus. Paratype, Nat. Mus. Viet. Reg. No. P.16155, a. Dorsal view; b, mesial view, x 0-6. Fig. 3a, b. First upper incisor (left) of Diarcodon parvus. Cast, A.M. Reg. No. F.5056. a. Dorsal view; B, mesial view. xO-6. Fig. 4a, b. Upper incisors of Diarcodon parvus. Holotype, A.M. Reg. No. 50099. A, Dorsal view, first upper incisor; b, as in Fig. 1, but mesial view. xO-6. Palaeontology, Vol. 6 PLATE 90 STEPHENSON, Diarcodon parvus N. G. STEPHENSON: FOSSIL MONOTREMES AND MARSUPIALS 623 alveolus, 25 mm. ; greatest width of exposed tooth, 25 mm. I., well worn, with patches of enamel on outer and inner sides of crown; total length of [-2, approximately 50 mm.; length of I, normally exposed beyond alveolus, 10-15 mm. ; greatest width of Ig, at base of crown, 15 mm.; depth, 12 mm. Part of the alveolus of I3 indicates a tooth of smaller size than I,,; estimated length of I3, 30 mm. I wish to express my thanks to Mr. A. J. Mahoney, University of Sydney, to Mr. E. D. Gill, National Museum of Victoria, and to Mr. H. O. Fletcher, Australian Museum, for the loan of material. My thanks are also due to the University of Sydney for travelling expenses enabling me to visit the Queensland Museum, and to the Department of Photographic Illustrations, University of Sydney, for Plate 90. REFERENCES BOITARD, p. 1841. Le Jardin des Plantes. Paris. CUVIER, G. L. c. F. D. 1798. Tableau elementaire de Vlustoire naturelle des animanx. Paris. CUVIER, G. F. and saint-hilaire, e. g. Histoire naturelle des manvniferes. Paris. DE VIS, c. w. 1 888. Note on the genera Zygomatiirus and Nototherinm. Proc. roy. Soc. Qd. 5, 1 1 1-16. DUN, w. s. 1895. Notes on the occurrence of monotreme remains in the Pliocene of New South Wales. Rec. geol. Siirv. N.S.W. 4, 118-26. FORSTER, J. R. 1788. Enchiridion Historiae Naturali inserviens, quo termini et delineationes ad Avium, Piscium, Insectorum et Plantarum adumbrationes intelligendas et concinnandas, secundum methodum systematis Linnaeani continentur. Halae. GERVAis, F. L. p. 1877-8. Osteograplue des Monotremes vivants et fossiles. Paris. gill, e. d. 1955. The range and extinction of Diprotodon minor Huxley. Proc. roy. Soc. Viet. 67, 225-8. 1957. The stratigraphical occurrence and palaeoecology of some Australian Tertiary marsupials. Mem. nat. Mus. Viet. 21, 135-203. GILL, T. 1877. General summary of scientific and industrial progress during the year 1876. Verte- brate Zoology. Ann. Rec. Sci. hid. 1876, clxx-clxxi (appeared 5 May 1877). GLAUERT, L. 1914. The Mammoth Cave. Rec. W. Aust. Mus. 1, 244-8. HARRIS, G. p. 1808. Description of two new species of Didelphis from Van Diemen’s Land. Trans. Linn. Soc. Land. 9, 174-8. HOOUER, D. A. 1951. Pigmy Elephant and Giant Tortoise. Sci. Mon. N.Y. 72, 3-8. HUXLEY, T. H. 1862. Oil the premolar teeth of Diprotodon, and on a new species of that genus. Quart. J. geol. Soc. Land. 18, 422-7. ILLIGER, c. 1811. Prodromus systematis mammalium et avium additis terminis zoographicis utriusque classis. Berlin. INTERNATIONAL COMMISSION FOR ZOOLOGICAL NOMENCLATURE. 1961. International Code of Zoological Nomenclature. London. JONES, F. w. 1930. A re-examination of the skeletal characters of Wynyardia bassiana, an extinct Tasmanian marsupial. Pap. roy. Soc. Tasm. 1930, 96-115. KREFFT. G. 1868. On the discovery of a new and gigantic fossil spccics of Ec/hW/w in Australia. Ann. Mag. nat. Hist. 1, 113-14. LAURIE, ELEANOR M. o. and HILL, J. E. 1954. List of Land Mammals of New Guinea, Celebes and Adjacent Islands. London. LYDEKKER, R. 1887. Catalogue of the Lossil Mammalia in the British Museum (Natural History). Part 5. London. m‘coy, F. 1876. Prodromus of the Palaeontology of Victoria. Decade ‘X. Melbourne. OWEN, R. 1838. Letter in t. l. mitchell. Three Expeditions into the Interior of Eastern Australia. London. • 1845. Descriptive and illustrated catalogue of the fossil organic remains of Mammalia and Aves contained in the Museum of the Royal College of Surgeons of England. London. 624 PALAEONTOLOGY, VOLUME 6 OWEN, R. 1877. Researches on the Fossil Remains of the Extinct Mammals of Australia. 2 vols. London. 1884. Evidence of a large extinct monotreme (Echidna ramsayi Ow.) from the Wellington Breccia Cave, New South Wales. Phil. Trans. 175, 273-5. PETERS, w. and doria, g. 1876. Descrizione di una nuova specie di Tachyglossiis proveniente dalla Nuova Guinea settentrionale. Ann. Mas. Star. nat. Genova, 9, 183-7. SPENCER, B. 1900. A description of Wynyardia bassiana, a fossil marsupial from the Tertiary beds of Table Cape, Tasmania. Proc. zool. Soc. Land. 1900, 776-94. STIRLING, E. c. 1913. On the identity of Phascolomys (Phascoloniis) gigas Owen, and Sceparnodon ramsayi Owen. Mem. roy. Soc. S. Aust. 1, 128-77. STiRTON, R. A. 1955. Late Tertiary marsupials from South Australia. Rec. S. Aust. Mus. ■ A new koala from the Pliocene Palankarinna fauna of South Australia. Rec. S. Aust. Mus. 13, 73-81. TEMMiNCK, c. J. 1827. Monographies de Mamnmlogie. Tome I. Paris. THOMAS, o. 1 888. Catalogue of the Marsiipialia and Monotremata in the collection of the British Museum (Natural History). London. WOODS, j. T. 1962. Fossil marsupials and cainozoic continental stratigraphy in Australia: a review. Mem. Od Mus. 14, 41-49. N. G. STEPHENSON Department of Zoology, University of Sydney, Manuscript received 3 December 1962 Sydney, New South Wales, Australia ‘HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES OF THE WENLOCK SHALES (SILURIAN) OF WENLOCR, ENGLAND by C. DOWNIE Abstract. In the examination of twenty-five assemblages obtained from samples from the Wenlock Shales of Wenlock, several thousands of ‘hystrichospheres’ (acritarchs) and about 100 spores were found. These are allo- cated to sixteen genera and over sixty-five species and varieties are recognized. The restricted ranges of some of the species and the marked variation in the proportion of others present allow these assemblages to be grouped into three main types. One assemblage type is confined to the Lower Wenlock Shales (Buildwas Beds), one to the Middle Wenlock Shales (Coalbrookdale Beds), and one to the Upper Wenlock Shales (uppermost Coal- brookdale and Tickhill Beds). Some account of the occurrence of ‘hystrichospheres’ (acritarchs) in the Wenlock Shales has already been given by the author (Downie 1959, 1960). The latter paper was concerned with the description of a few novel forms from the lowest part of the Wenlock Shales; the former dealt fully with an assemblage from the middle part. The present paper is more comprehensive, ‘including the description of hystrichosphere’ assemblages from twenty-five samples spaced throughout the Wenlock Shales and an account of the differences in the character of these assemblages as they follow each other through the rock succession. New facts about the affinities of the hystrichospheres have necessitated their reclassi- fication and proposals put forward by Downie, Evitt, and Sarjeant (1963) have been adopted in this paper. The name ‘hystrichosphere’ is no longer appropriate for the forms dealt with here and is replaced by the term ‘acritarch’ (Evitt 1963): formal classification into orders is replaced by a less rigid supra-generic grouping not based on type genera. Location of the samples. The thickness of the Wenlock Shales below Wenlock Edge is difficult to estimate accurately. Pocock et al. (1938) estimate it to be about 1,000 feet; measurements made by the author west of Much Wenlock indicated 950 feet and this figure is used in this account. The difficulty in measuring the thickness is partly due to the absence of a continuous section through the formation, indeed some parts of the sequence are rarely exposed and in order to get samples more or less evenly spaced throughout the sequence the sampling localities are spread for quite a considerable distance along the strike (text-fig. 1). Because of this and the difficulty in accurately measuring the thickness there is some uncertainty about the precise position of the samples in the sequence. Their positions (text-fig. 1 ) are based on direct field measure- ments when circumstances permitted, and in other instances on geometric calculations. Although the possible errors for some of the samples in the middle of the sequence may be about 20 feet plus or minus, there is no doubt about the relative positions of any of the samples, and the assemblages obtained from them can be placed in the correct chronological order. In general the spacing of the samples is about 100 feet. Closer spacing would have [Palaeontology, Vol. 6, Part 4, 1963, pp. 625-52, pis. 91-92.] 626 PALAEONTOLOGY, VOLUME 6 WENLOCK LIMESTONE Tickhill Beds Coalbrookdale BedsI Buildwas Beds -Ws|4a,b WSlSa.b ■ WsjSa.bx W5)a Ws|Za,b Ws|l Q,b,c WS/l2 ws|io ws|ll - wsjba.b - WsLc.b BS/2a,b ,BS|3 ■BSila.b HUGHLEY SHALE Scale I" ■. aoott. SAMPLE LOCALITiES & STRATIGRAPHIC POSITION TEXT-FIG. 1. Map and section showing the location of the samples collected. been ideally more desirable but the results show that except near the base and near the top, where the spacing is already fairly close, the character of the assemblages changes only slightly. At a number of localities two or even three samples have been taken at vertical intervals of about 1 or 2 feet in order to check that the nature of the microflora was not varying erratically over short periods. In every case the results confirmed the relative constancy of the character of the assemblages over these short ranges. C. DOWNIE: ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 627 The localities from which samples were collected are as follows: BS/la Soft grey shale, 4 feet above base of Wenlock Shales, Harley Brook, 100 yards downstream from bridge at Domas. G.R. 33/596008. BS/lb As for BS/la, but 5 feet above base of Wenlock Shales. BS/3 As for BA/la, but 6 feet above base of Wenlock Shales. BS/2a Calcareous nodule, c. 55 feet above base of Wenlock Shales, Harley Brook, 150 yards up- stream from bridge at Domas. G.R. 33/594006. BS/2b As for BS/2a, but 56 feet above base of Wenlock Shales. WS/la Somewhat weathered shale, roadside exposure where railway crosses road from Ticklerton to Eaton. G.R. 32/497901. WS/lb One foot above WS/la. WS/lc One foot above WS/lb. WS/A Calc, grey mudstone, section in lane behind Eaton Church. G.R. 32/501900. WS/2a Three feet below WS/A. WS/2b One foot above WS/2a. WS/4a Grey mudstone, lane section above Eaton Church, 105 feet below Wenlock Limestone. WS/4b One foot below WS/4a. WS/5a Calc, grey mudstone, lane section above Eaton Church, 89 feet stratigraphically below WS/4a. G.R. 32/505899. WS/5b As for WS/5a. WS/5c One foot below WS/5a. WS/6a Grey mudstone, small stream section half mile east of old mill at Sheinwood. G.R. 33/624028. WS/6b One foot above WS/6a. WS/7a Grey mudstone, small stream section 250 yards east of old mill at Sheinwood. G.R. 33/618027. WS/7b One foot above WS/7a. WS/8a Calcareous grey mudstone, road section at Harley Hill, 30 feet below base of Wenlock Lime- stone. G.R. 33/610003. WS/8b One foot above WS/8a. WS/10 Grey mudstone, stream section in Whitwell Coppice, 500 yards above junction with Sheinton Brook. G.R. 33/6120021. WS/11 Grey mudstone, stream section in Whitwell Coppice, 70 feet stratigraphically below WS/10. G.R. 33/619020. WS/12 Grey mudstone, road section 200 yards south of Rushbury Church. G.R. 32/514918. Treatment of the samples and general character of the residues. The samples were pre- pared by first removing the carbonates in dilute hydrochloric acid and then dissolving the silicates in hydrofluoric acid. Subsequent oxidization was generally omitted. The Wenlock Shales consist of a relatively uniform series of more or less calcareous grey mudstones which readily break down when treated with hydrochloric acid followed by hydrofluoric acid. The residues which are obtained contain abundant acritarchs; only one sample was checked quantitatively : this yielded about 5,000 acritarchs per gramme of rock. Accom- panying the acritarchs was a smaller number of spores, chitinozoa, scolecodonts, grapto- lite fragments, and organic debris of various kinds. Only the acritarchs and spores are dealt with in this paper. DISTRIBUTION OF THE SPORES There is as yet no generally agreed principle whereby spores are to be separated on an objective morphological basis from certain of the acritarchs and in the Lower Palaeozoic taxonomic procedure has not been stabilized. In this paper only those microfossils which are indubitably the reproductive organs of higher plants are considered as spores, the 628 PALAEONTOLOGY, VOLUME 6 presence of triradiate marks on one surface being the main determining character. But a number of other forms are present in the Wenlock Shales which, although lacking this character, in other ways resemble the spores of plants more highly organized than the algae. They are nevertheless grouped with the acritarchs because of the difficulty of distinguishing acritarchs from spores without triradiate marks. Spores in the sense defined above are not numerous; altogether just over a hundred were observed, all belonging to one species with a single exception. Although they are never common the spores appear to range throughout the Wenlock Shales and may form as much as 4 per cent, of the whole assemblage. Table 15 shows how the propor- tion present varies and it will be seen that they tend, in general, to be a little more common in the higher than in the lower beds. All the spores observed, except one, belong to the species Piinctatisporites? dilutus Hoffmeister, previously described only from the Lower Silurian of Libya (Hoffmeister 1959). The Wenlock forms (Plate 92, fig. 13) agree almost exactly with those from Libya and many show the darkening at the equator described by Hoffmeister. In no instance, however, is this sufficiently clearly set off from the rest of the exine to justify attributing any of the specimens to Ambitisporites avitiis which accompanies P.? dilutus in the Libian assemblage. The only other spore with triradiate markings was found in WS/12 and was a form of Lophotriletes somewhat similar to L. minuscida Naumova but with thinner exine and with an ornament of fine irregular verrucae. The distribution of the spores is shown in Table 3. SYSTEMATIC DESCRIPTIONS AND DISTRIBUTION OF THE ACRITARCHS Acritarchs always form the greater part of the assemblages. They vary greatly in shape and have been divided into a large number of species and genera: these are grouped according to their more obvious morphological features. Group ACRITARCHA Evitt 1963 Subgroup SPHAEROMORPHITAE Downie, Evitt, and Sarjeant 1963 (1) Smooth spheres (leiospheres) These fossils have in the past been classified with the hystrichospheres (Eisenack 1958) but they are a polyphyletic group, few if any having any affinity with the dino- flagellates. Although most are likely to be planktonic organisms of various kinds or their reproductive stages, others may well be spores of more highly organized plants. Several thousands were observed in the Wenlock Shales where they form from 20 to 99 per cent, of the total assemblage, but generally about one-third (Table 15). It is not known what significance, if any, is to be attached to these fluctuations in the propor- tions of leiospheres present. The increases may be due to chance ‘flooding’ of the sample from a local source or to secular changes in the composition of the plankton. The complete dominance of samples from locality WS/1 by leiospheres was so remark- able that its reality was checked in three separate rock samples from the locality. All were similar in that they yielded no spores or spiny acritarchs. Unfortunately no C. DOWNIE: ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 629 unweathered samples could be got at this locality and the effects of weathering cannot be eliminated as a factor in producing these unusual samples. However, it may be signi- ficant that both the preceding and succeeding samples show a much higher content of leiospheres than average. There is no notable variation in the species of leiospheres present throughout the succession. By far the most common is Leiosphaeridia weiilockia Downie, which is always at least twice as common as all the others considered together and may be twenty times as common. In extending observations of this species throughout the Wenlock Shales no need has arisen to alter the range of variation stated in Downie (1959) but it should be noted that pi. 3, fig. 4, in that paper shows a common and distinctive variant, probably due to preservation, having its surface distorted into numerous swellings and its interior occupied by a number of pyrite grains. It is doubtful if more than one or two of the many hundreds of specimens examined have a pylom and since the wall may sometimes be a little thicker or thinner than the 1 p stated in the diagnosis it is very doubtful if Protoleiospliaeridium orbicidatwn Staplin is a distinct species. Leiosphaeridia cf. microsystis (Eisenack) Downie 1959, although rare by comparison with L. weidoekia, is generally quite common and has been found to occur throughout the Wenlock Shales. Its size range is now known to be from 40 to 75 microns which brings it into conformity with Eisenack’s typical L. microcystis', there is, however, good reason for maintaining a distinction since the Wenlock forms never appear to have a pylom but open by splitting cleanly along a great circle (Downie, 1959, pi. 12, fig. 1). Next in importance is Leiosphaeridia sp. cf. Protoleiosphaeridiwn diaphamini Staplin which probably also ranges throughout the Wenlock Shales generally in smaller but occasionally in slightly greater numbers than L. cf. microcystis (see Table 1 ). They closely resemble the species described by Staplin (1961) but are larger, 50 to 80 microns as against 40 to 50 microns for the Devonian forms. There are a few other forms present but these are rare not being represented by more than one or two in populations of about a thousand. They appear to be of little interest. The distribution of the leiospheres is shown in Table 1 . TABLE 1. Recorded occurrences of Leiosphaeridia Locality BS/l oa BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 1 1 WS/2 WS/5 WS/8 WS/4 L. wenlockia .... 687 61 71 178 146 45 35 90 148 1,202 122 213 326 L. cf. microcystis .... 80 38 10 31 2 8 5 19 26 37 4 1 1 61 L. cf. diaphanum .... 15 4 5 5 3 6 1 43 9 7 3 L. spp. ..... 5 2 I 1 4 2 4 3 (2) Ornamented leiospheres In every assemblage there is a number of microfossils resembling the leiospheres described above but ornamented with solid spines and tubercles which are short in rela- tion to the size of the body. They are provisionally placed in the genus Lophosphaeridium Timofiev and in general form about 3 per cent, of the whole assemblage. Seven species are considered here; others exist but are too rare to be of any impor- tance. Of these seven species, four, L. sp. cf P.papillatum Staplin, L. triangidarwn sp. nov., L. granulosum (Staplin), and L. citrinum sp. nov., range throughout the Wenlock Shales, Tt C 1713 630 PALAEONTOLOGY, VOLUME 6 L. granulosiun apparently extending into the Devonian. Of the three other species, one, L. microspinosum (Eisenack), has been found only in the lower and middle part of the Wenlock Shales, and, by Eisenack, and in the Upper Llandovery of Estonia. The exis- ence of a similar form in the Upper Ludlow (Eisenack 1955) suggests, however, that the species may range right through. The other two species are probably characteristic of the upper levels of the Wenlock Shales, L. piloswn sp. nov. was found only in samples from locality WS/4, and L. sp. cf. P. cryptogranulosum Staplin was confined to this locality also. The distribution of these forms is shown in Table 2. TABLE 2. Recorded occurrences of Lophosphaeridium Locality BS/l BS/3 BS/2 WS/7 WS/6 WS/ll WS/10 WS/12 WS/1 WS/2 WS/5 WS/8 t D L. cf. cryptogramilosiim . 11 L. granulosum .... 63 1 4 8 25 3 3 6 2 24 13 8 13 L. citrinum ..... 12 3 4 7 8 1 7 L. cf. papillatum .... 2 1 5 2 I 1 5 2 2 12 L. triangulatwn .... 1 1 1 2 L. piloswn ..... 8 L. microspinosum .... 2 7 6 2 2 8 SYSTEMATIC DESCRIPTIONS Genus lophosphaeridium Timofiev 1959 Type species. Designated here: L. rariim Timofiev 1959. Remarks. Timofiev did not designate a type species for his genus, the first named of his two species is selected here. The genus is characterized by its solid tubercules which distinguish it from Leiosphaeridia. Some forms occurring in the Wenlock with short solid spines with rounded or capitate ends are also included here provisionally. Lophosphaeridium granulosum (Staplin) Synonym. Protoleiosphaeridiiim grannlosiim Staplin 1961. Remarks. Several hundred individuals were seen from samples ranging throughout the Wenlock Shales. They appear to be identical to the Devonian species, although they may range up to 40 microns in size. Lophosphaeridium sp. cf. P. cryptogranulosum Staplin Remarks. This form, of which eleven examples were found in samples from locality WS/4, resembles in most respects Protoleiosphaeridium cryptogranulosum from the Devonian, but being generally about 25 microns in size, it is consistently larger. Its out- line is always subquadrate and it is quite possible that though no triradiate marks were seen, that this form is in fact the spore of a higher plant. Lophosphaeridium citrinum sp. nov. Plate 92, fig. 3 Diagnostic. Vesicle ellipsoidal, lemon yellow in colour. Ornament of capitate spine C. DOWNIE: ‘HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 631 (pilae). Body size about 40 by 30 microns; spine length 1 to 2 microns, spacing 1 to 2 microns. Holotype. Slide 5, position 080.450. Locality. Wenlock Shales, WS/4a. Remarks. The species is not uncommon, over forty individuals having been found, and it appears to range throughout the Wenlock Shales. It is clearly a leiosphere since most of the examples show a pylom whose diameter is about 20 per cent, of the test length. Lophosphaeridium pilosum sp. nov. Plate 92, fig. 2 Diagnosis. Vesicle thin-walled elhpsoidal subcircular. Ornament of closely spaced short solid spines. Body colour lemon-yellow. Body size 35 to 40 microns, spine length 3 to 4 microns, spacing 1 micron. Holotype. Slide 9, position 000.630. Locality. Wenlock Shales. WS/4a. Remarks. This species, of which eight examples are known, has been found only in samples from locahty WS/4. It resembles P. microsaetosum Staphn but is more ellip- soidal and has a denser pile of spines. It also bears some resemblance to L. citrinum sp. nov., but the ornament is saetose and generally denser. Lophosphaeridium sp. cf. P. papillatiim Staplin Plate 92, fig. 12 Remarks. This species is known from over thirty specimens spread throughout the Wenlock Shales. The body is a pale yellowish-green colour and circular to subtriangular in outline. It may well be a spore from one of higher plants but no triradiate marks have been seen. Its ornament consists of short solid spines 1 to 2 microns in length and 1 to 2 microns apart. It resembles both P. papillatiim and P. microsaetosum of Staphn but is rather smaller than either, the body measuring from 18 to 28 microns. Furthermore, the subtriangular outline is distinctive. Lophosphaeridium triangulatum sp. nov. Plate 92, fig. 1 Diagnosis. Vesicle subtriangular in outline, pale yellow-green in colour. Ornament of solid spines up to 5 microns long and 5 microns apart. Body size from 18 to 25 microns. Holotype. Slide 8, position 050.620. Locality. Wenlock Shales, WS/4a. Remarks. This species is rare, only five examples having been found, but it probably ranges throughout the Wenlock Shales. It resembles closely L. cf papillatum, differing only in size and spacing of the spinose ornament. Like L. cf papillatum it may well be a spore. 632 PALAEONTOLOGY, VOLUME 6 Lophosphaeridium microspinosiim (Eisenack) Plate 92, fig. 11 Synonyms. Hystrichosphaeridium microspinosiim Eisenack 1954. Baltisphaeridiiim microspinosiim Downie 1959. Remarks. This species with its ornament of short solid spines and pilae must be grouped with the forms discussed above since it differs markedly from the typical forms of Balti- sphaeridiiim with their relatively long hollow spines. Twenty-eight examples have been found in the Wenlock Shales most of them in the lower horizons and none above locality WS/A, about 300 feet from the top. This is in accord with the occurrence of Eisenack’s type material in the Upper Llandovery, but the occurrence of a similar form in the Upper Ludlow suggests that it may eventually turn up in the Upper Wenlock. Subgroup PTEROMORPHiTAE Dowiiic, Evitt, and Sarjeant 1963 Only eleven specimens of acritarchs with a characteristic equatorial flange were found, all belonging to the genus Pterospermopsis which appears to range throughout the Wenlock Shale, but is probably more common in the upper part (see Table 3). Three types were found, the most common being P. cf. ouondagaensis Deunff (Downie 1959). The others are a small colourless form from locality WS/2 and a form with a quadrate outline from WS/12. TABLE 3. Recorded occurrences of spores, Tasmanites, a.nd Pterospermopsis Locality BS/l BS/3 BS/2 WSI7 WS/6 WS/ll 1 WS/IO WS/12 WS/1 WS/2 WS/5 WS/8 i WS/4 Pimctatisporitesl dilutus . Lophotriletcs sp. .... Tasmanites spp. .... Pterospm. cf. ouondagaensis Pterospm. spp. .... 1 3 1 7 3 22 1 3 1 1 4 1 10 30 6 1 6 2 12 5 1 41 1 1 Subgroup HERKOMORPHITAE Downic, Evitt, and Sarjeant 1963 The only acritarch genus present with flanges dividing the surface into polygonal fields is Cymatiosphaera which is fairly common throughout the Wenlock Shales form- ing up to 4 per cent, of the whole assemblage. The commonest species, C. wenlockia TABLE 4. Recorded occurrences of Cymatiosphaera spp. Locality BS/l BS/3 i 1 BS/2 WS/7 WS/6 WS/ll WS/10 WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 C. octoplanus .... 2 2 2 1 1 8 1 3 C. cubus ..... 2 1 1 3 1 4 9 3 2 2 C. prismatica .... 1 3 C. pavimenta .... 2 14 1 1 2 C. wenlockia .... 3 2 3 2 2 1 3 7 1 7 Other spp. ..... 4 1 1 2 2 1 1 2 Downie, and the closely related C. eubus Deunff and C. octoplana Downie, appear to be present through the whole sequence. Among the other species present C. pavimenta C. DOWNIE: ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 633 (Deflandre) and C. prismatica DeunfF have been found only in the middle part of the sequence. The latter is, however, rare, and was, furthermore, first described from the Devonian so that it has little stratigraphical signihcance. Other forms found are too rare to merit description at present. Subgroup NETROMORPHiTAE Downie, Evitt, and Sarjeant 1963 This group of acritarchs with distinctly elongate vesicles includes the genera Leiofusa, Deimffia, and Domasia. The leiofusids are represented by L. fiUfera Downie and L. tumida Downie which differ only in the proportions of the body. Since publishing the first account of these forms it has been found that there is complete gradation between the two and that intermediate forms predominate at some horizons (see Table 5). The stouter bodied TABLE 5. Proportion of different types of leiofusids expressed as percentage total number of leiofusids present Locality L. filifera L. tumida spherical body breadth/length + 65«o narrow body breadth/length 20-35 °o wide body breadth/length 35-65°b WS/4 18 70 12 WS/8 50 29 21 WS/5 90 10 WS/2 70 20 10 WS/1 WS/12 84 16 WS/10 86 7 7 WS/11 92 4 4 WS/6 100 wsp 100 BS/2 100 BS/3 100 BS/1 forms are not, however, always present, being confined to locality WS/11 and higher levels where they increase in importance at the expense of the slim-bodied forms, and eventually become dominant. These leiofusids, of which several hundred were seen, tend as a whole to become abundant in the upper levels and in sample WS/4b form over 14 per cent, of the total assemblage. A difficult taxonomic problem is set by the few small leiofusids that occur in the Buildwas Shale. Four examples were found, all measuring about 40 microns in overall length. They are morphologically almost identical to small specimens of L. tumida which, however, usually exceeds, often greatly, 100 microns in length. In the Buildwas Shale these small leiofusids are associated with Deimffia and Domasia and in body size, shape, and texture all three are identical (text-fig. 2). There seems little doubt that they are closely related and that these small leiofusids are homoeomorphs of L. tumida more closely related to Deimffia and Domasia than to L. filifera. Apart from slight, and TEXT-FIG. 2. The stratigraphic distribution and possible relationships of some acritarchs. a, Deimffia moiiospiuosa; b, Deunffia brevispinosa\ c, Leiofusa cf. tumida\ d, Domasia bispinosa; e, Domasia trispinosa; f, Domasia eloiigata; g, Veryhachiiim elongatiarr, h, V. trispinosum; i, V. europaeiim var. wenlockium; j, V. fonnosum var. nov. ; k, Leiofusa tumida. The arrows indicate where morphological transition is probable. 635 C. DOWNIE: ‘HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES possibly unreliable, differences in the texture of the wall there appears to be no morpho- logical character to distinguish them from L. tuinida, but because of their associations they are referred to here as L. cf. tumida. Deunffia and Domasia are closely related and were found only in the Buildwas Beds at the base of the Wenlock Shale. There they were quite common, especially at locality BS/2 where they form over 22 per cent, of the whole assemblage. The morphological relationship of the forms present and their distribution is shown in text-fig. 2 and it is clear that those from BS/2 are quite distinct from the assemblages from BS/1 and BS/3. Deimffia monospinosa has not been found at BS/3 but it probably occurs throughout the Buildwas Beds. Domasia bispinosa and D. trispinosa were found only in the two lowest localities, BS/1 and BS/3. Deimjfia brevispinosa and D. ramuscuJosa were confined to BS/1 while Domasia elongata and Deunffia fur cat a were found only in samples from BS/2. The distribution of these forms is shown in Table 6. TABLE 6. Recorded occurrences of Domasia, Deunffia, and Leiofusa Locality BS/1 BS/3 BSI2 WS/7 WS/6 |ws/ll WS/10 WS/12 WS/1 |wS/2 WS/5 WS/8 WS/4 Domasia elongata .... 19 D. trispinosa ..... 5 3 D. bispinosa ..... I 4 Deunffia monospinosa 19 7 Dnf. brevispinosa .... 11 Dnf. ramitsculosa .... 7 Dnf. furcata ..... 63 . . Leiofusa fUifera .... 5 3 8 1 23 13 5 28 38 11 108 L. tumida ..... 1 1 1 3 3 16 L. cf. tumida .... 2 2 Subgroup POLYGONiMORPHiTAE Downie, Evitt, and Sarjeant 1963 (1) Acritarchs with polygonal bodies The veryhachids in the Wenlock Shales are mostly small but throughout the succes- sion they are generally abundant (Table 8). Of the eight species recognized, Veryhachium rhomboidium Downie, V. trispinosum (Eisenack), and V. europaeum Stockmans and Wil- liere var. wenlockium (Downie) are the most common and range all through the sequence. V. rhomboidium is quite distinct from the other veryhachids in the Wenlock but is prob- ably closely related to Baltisphaeridium longispinosum var. parvum with which it is com- monly associated. V. trispinosum and V. europaeum var. wenlockium are probably the three- and four-spined varieties of the same organ or organism; the five-spined variety being V.formosum Stockmans and Williere, which is less common and confined to locality WS/6 and higher horizons (text-fig. 2). Apart from the appearance of the five-spined forms in the Coalbrookdale Beds there is little change in the composition of this complex throughout the sequence (Table 7). Among the other forms present, V. cf. balticum has been found only twice and is too rare to be of any importance. V. elongatum sp. nov., V. bulbosum (Deflandre), and V. trisphaeridium sp. nov., however, appear to have restricted ranges; V. elongatum being confined to the lower Coalbrookdale Beds, V. bulbosum to the lower and middle Coal- brookdale Beds, and V. tripartitum to the Tickhill Beds. The distribution of these forms is shown in Table 8. 636 PALAEONTOLOGY, VOLUME 6 Veryhachium rhomboidium Downie Remarks. This species is now considered as a form species with homoeomorphic repeti- tions at widely separated horizons (Wall and Downie 1963). The description of the Silurian forms (Downie 1959) needs no alteration, but it should be noted that rare examples with shorter or longer spines than stated in the diagnosis do occur. TABLE 7. Changes in the V. Irispinosum-V. europaeum-V. formosum complex Locality BS/I BS/3 BS/2 'NSn WS/6 WS/11 WS/IO WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 Percentage of forms with: 3 spines {V. trispiuosum) . 19 15 30 31 35 13 9 3 13-4 23 31 28 4 spines {V. europaeum) . 81 85 70 69 63 83 91 97 86 64 63 68 5 spines (K formosum 2 4 0-6 13 6 4 TABLE 8. Recorded occurrences of Veryhachium Locality BS/1 e/sa BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 V. rhomboidium .... 45 24 10 15 6 1 4 5 6 30 8 V. europaeum var. wenlockium . 17 19 19 31 38 58 41 29 144 25 69 39 V. formosum var. nov. 1 3 1 5 6 2 V. trispinosum .... 4 3 8 14 21 9 4 1 22 9 34 16 V. elongatum ..... 15 1 V. bulbiferum .... 1 3 3 1 2 V. cf. balticum .... 1 1 F. iripartitum .... 2 4 V. spp. 5 * 1 4 1 Veryhachium europaeum Stockmans and Williere var. wenlockium (Downie) Synonym. V. tetraedron var. wenlockium Downie 1959. Remarks. The reasons for transferring this form to V. europaeum are discussed in Wall and Downie 1963. A small number of specimens whose spines are shorter than allowed for in the original diagnosis have been found. Veryhachium trispiuosum (Eisenack) Remarks. The forms from the Wenlock Shale allocated to this species are closely akin to V. europaeum var. wenlockium, differing only in the number of spines. Both are prob- ably only parts of a variable group of organs or organisms belonging to one inter- breeding species. The Wenlock specimens of V. trispiuosum are smaller than the typical forms of Eisenack and have slightly more constricted spine bases. In these respects they resemble more closely V. reductum Jekhowsky, which, however, has shorter spines. V. trisidcum Deunff differs in having a markedly inflated test with a tendency to have an inequilateral outline. The Wenlock forms are therefore retained in V. trispiuosum. C. DOWNIE; ‘HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 637 Veryhachium elongatum sp. nov. Plate 92, fig. 10 Diagnosis. A species of Veryhachium with three simple spines about equal in length to the body. The body is elongate, length about three times the width, subtriangular in outline. Dimension. Body length 7 (10) 16 microns, width 4-7 microns; spinebody length ratio 10 to 2-5. Holotype. Slide 1, position 280.515. Locality. Wenlock Shales, WS/7b. Remarks. This small species is associated in samples from WS/7 with V. trispinosum which it resembles except for its elongated body. This body is, however, distinctly tri- angular which distinguishes it from Domasia elongata which it also resembled in size, spine number, and length. It is indeed intermediate in morphology between Veryhachium of the V. trispinosum-V. trisulcum group and Domasia of the D. trispinosa-D. elongata group and it is interesting to note that it follows D. elongata immediately in the sequence (text-fig. 2). Altogether over fifteen individuals were found in samples from WS/7 and a single one from a sample from WS/1 1 . Veryhachium trisphaeridium sp. nov. Plate 92, fig. 7 Diagnosis. A species of Veryhachium with three simple spines and an equilateral thin walled subtriangular test. The spines are equal in length or longer than the sides of the test. The test is constricted midway between the spine bases forming a Y-shaped furrow clearly visible when the spines are more or less in the same plane. Dimensions. Body size about 15 microns; spinebody length radio, 10 to 1-7. Holotype. Slide 2, position 198.623. Locality. Wenlock Shales, WS/8b. Remarks. In size and shape this species resembles V. trispinosum from the Wenlock Shales, differing in the presence of the grooves which divide the test into three rather bulbous apical parts. It is not a common form and the six known examples all come from the Tickhill Beds at the top of the Wenlock Shales. (2) Acritarchs with body formed by fusion of the process bases This group includes forms belonging to the genera Estiastra and Pulvinosphaeridium. Estiastra, hitherto represented by a single species from the Upper Llandovery of the Baltic, occurs at two distinct levels in the Wenlock Shales. E. barbata sp. nov. occurs in considerable numbers but only in the lowest horizons of the Buildwas Beds, E. granu- lata sp. nov. is less common and was found only from locality WS/4 in the Tickhill Beds (Table 1 3). Pulvinosphaeridium has been found only as a rare constituent in samples from locality WS/A, but its distribution in the rocks of the Baltic area shows it to range from the Upper Llandovery to the Ludlow (Eisenack 1959). 638 PALAEONTOLOGY, VOLUME 6 Estiastra barbata sp. nov. Plate 92, fig. 8 Diagnosis. A species of Estiastra 80 to 1 30 microns across with four to eight processes. The surface of the processes ornamented with varying density by small hollow spines up to 2 microns in length. Holotype. Slide 3, position 135.452. Locality. Wenlock Shales, BS/3. Remarks. This species resembles the type species E. magna Eisenack in overall shape but it is much smaller and is ornamented with small spines. Estiastra grauulata sp. nov. Plate 91, fig. 8 Diagnosis. A species of Estiastra 100 to 150 microns across with eight to twelve processes. The walls are thin, generally crumpled, and ornamented with a fine ornament of small closely spaced granules. Holotype. Slide 9, position 120.615. Locality. Wenlock Shales, WS/4a. Remarks. This species resembles E. magna Eisenack but is smaller, paler in colour, and has a finely granular surface. Pulvinosphaeridium oligoprojectum Downie Remarks. No more specimens of this species have been found in the Wenlock Shales since it was first described (Downie 1959) but subsequently discussion with Professor Eisenack suggests it is a junior synonym of P. puJvinellum Eisenack which ranges from the Llandovery to the Ludlow. Subgroup ACANTHOMORPHiTAE Downie, Evitt, and Sarjeant 1963 (1) Acritarchs with relatively large spherical bodies and hollow spines (Baltisphaeridium) (a) Spines smooth and generally simple. This group embraces two large variable species of acritarchs, varieties of which are likely to appear at many horizons. The long process forms in the Wenlock Shales are allocated to varieties of B. longispinosum (Eisenack), the forms with shorter processes to B. brevispinosiim (Eisenack). Both of these species range throughout the Wenlock Shales and are nearly always present in considerable numbers (Table 9). Variation in the numbers present and in their relative proportions does not appear to have any stratigraphical significance in the Wenlock EXPLANATION OF PLATE 91 Figs. 1-7. Baltisphaeridiian spp. 1, B. graniiiatispiuosum sp. nov.; holotype, x520; locality BS/3. 2, B. longispinosum var. parvuin var. nov. ; holotype, X 550; locality BS/3. 3, B. microcladum sp. nov. ; paratype, x600; locality BS/3. 4, B. longispinosum var. paucispinosmn var. nov.; holotype, x550; locality WS/5. 5, B. arbuscidiferum sp. nov.; holotype, x700; locality WS/4. 6, B. ravum sp. nov.; holotype, X 400; locality BS/2. 7, B. granuiatispinosum sp. nov. ; a form with markedly cladate pro- cesses, X 575 ; locality BS/3. 8, Estiastra granulata sp. nov. ; holotype. x 500; locality WS/4. Palaeontology, Vol. 6 PLATE 91 DOWN IE, Silurian ‘hystrichospheres' C. DOWNIE: ‘HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 639 Shales, except that the forms of B. longispinosum are more abundant in the earlier horizons. Baltisphaeridium longispinosum (Eisenack) var. parvum var. nov. Plate 91, fig. 2 Synonymy. Baltisphaeridium longispinosum (Eisenack); Downie 1959 (pars), pi. 10, figs. 2, 6. Micrhystridium stellatum Deflandre var. inflatum Downie 1959, pi. 11, fig. 12. Diagnosis. A variety of Baltisphaeridium longispinosum having a more or less spherical body 15 to 35 microns in diameter, bearing spines longer than the test radius and five to twenty-five in number, the spines are hollow, closed at the tips, with a smooth surface and rarely branching. Holotype. Slide 3, position 100.560. Locality. Wenlock Shales, BS/3. Remarks. This species is generally abundant, over 300 individuals distributed throughout the Wenlock Shales were seen. It is, however, rather more common in the Buildwas Shales. The variety resembles the typical Ordovician form in general appearance but it is TABLE 9. Recorded occurrences of Baltisphaeridium spp. belonging to group {a) Locality BS/I BS/3 (N CQ WS/7 WS/6 WS/11 WS/IO WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 B. longispinosum var. parvum . 101 37 48 10 10 16 9 2 12 26 18 18 B. longispinosum var. paucispinosuni . 1 4 6 B. brevispinosum var. wenlockensis 23 49 35 40 53 17 43 5 77 11 18 17 B. brevispinosum var. granuliferiim . 1 2 6 2 2 8 1 B. brevispinosum var. naniim 2 3 2 4 5 1 1 1 3 2 1 consistently smaller and a pylome has never been observed. Furthermore, branching spines rarely occur, broad processes are absent and in general these appear more flexuous than those of the typical Ordovician forms. Some examples from the Wenlock show morphological transitions to other species occurring there. For example some with six regularly disposed processes differ from V. rhomhoidium only because the test is ellipsoidal and not polygonal. Fortunately this difficulty rarely arises and only in samples from locality BS/1 where there were 12 of these intermediaries among 101 specimens of B. longispinosum and 40 of V. rhomboidium did it assume any importance. Some small forms of B. longispinosum var. parvum with more markedly flaring spine bases are difficult to separate from Micrhystridium stellatum Deflandre. It can in fact be done only on an arbitrary basis and for this reason M. stellatum var. inflatum is con- sidered synonymous with B. longispinosum var. parvum. In the Ordovician Eisenack found that it was convenient to make an arbitrary distinc- tion between B. longispinosum and forms with short processes. The same is the case in the Wenlock and the distinction between B. longispinosum var. parvum and B. brevi- spinosum var. wenlockensis is based mainly on the length of the processes. Forms near the boundary (processes equal to the radius in length) are, however, less common than typical forms except in samples from BS/3 where they predominate. 640 PALAEONTOLOGY, VOLUME 6 Boltisphaeridium longispinosum (Eisenack) var. paucispinosum var. nov. Plate 91, fig. 4 Diagnosis. A variety of B. longispinosum with a more or less spherical body 15 to 35 microns in diameter, bearing spines longer than the test diameter. Spines hollow, un- branched and closed at the tips, less than five in number. Holotype. Slide 1, position 210.448. Locality. Wenlock Shales, WS/5c. Remarks. This form of which only eleven examples have been found occurs mainly near the top of the Coalbrookdale Beds. It may be connected with B. eoplanktonicum but the intermediate forms with some branched and some unbranched spines are rare. Six were found, five of which were from locality WS/5, the only place where this supposed transi- tion is substantially supported. Baltisphaeridium brevispinosum (Eisenack) var. wen/ockensis Downie Remarks. This variety is one of the commonest in the Wenlock Shales and one of the most variable. The body size ranges from 15 to 40 microns, its shape from spherical to ellipsoidal, the process length from one-fifth to three-quarters of the diameter and number from about twelve to forty in optical section; their tips may be sharply pointed, hair-like, rounded or capitate. B. brevispinosum var. granuliferum Downie belongs to the same plexus, but B. brevispinosum var. nanum Deflandre appears to be less closely connected with its widely spaced relatively short processes. All of these varieties range throughout the Wenlock Shales and there is no change in the emphasis of the variation that has any stratigraphic significance. TABLE 10. Recorded occurrences of Baltisphaeridium spp. belonging to group {b) Locality BS/l BS/3 BS/2 WS/7 WS/6 ws/n WS/IO WS/12 WS/i WS/2 WS/5 WS/8 WS/4 B. granulatispinosum B. robustispinosum .... 26 14 2 6 22 3 n 9 26 2 9 1 43 2 60 71 73 {b) Spines, branched or unbranched, ornamented with small tubercles. This group includes two species, B. granulatispinosum sp. nov. and B. robustispinosum Downie. The former is the more common and over 350 examples have been found. They are distributed throughout the Wenlock Shales but are clearly more common in the upper part. Some rather distinctive varieties are included in this species but they do not appear to be restricted in their stratigraphical distribution and are relatively rare. B. robustispinosum is represented by only twenty specimens and is entirely absent in the samples from the upper part of the Wenlock Shales, most of them being found in the lower part of the Coalbrookdale Beds (Table 10). Baltisphaeridium granulatispinosum sp. nov. Plate 91, figs. 1,7; text-fig. 3c Synonyin. B. longisposum (Eisenack); Downie 1959 (pars), pi. 10, fig. 1. Diagnosis. A hollow spherical, ellipsoidal or rarely polygonal test with hollow spines C. DOWNIE; ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 641 about equal or greater in length than the test diameter. Spines simple, rarely branching, but terminating in nearly every instance in short hnger-like digitations. Invariably the surface of the spines along all their length is ornamented with granules about 1 micron in size. Dimensions. Body size 18(30)36 microns; ratio spine lengthbody size 0-8(1 ■2)2-0; number of processes 5(8)12. Holotype. Slide 3, position 030.540. Locality. Wenlock Shales, BS/3. Remarks. Typical members of this species have spines unbranched except at the tip where they fray into small finger-like processes. Varieties with the tips closing to a point are very rare (10 in 364) as are those with markedly cladate spines (14 in 364). Baltisphaeridium robustispinosum Downie Remarks. This species is distinguished from B. gramdatispiuosum by its broader stouter processes and generally coarser ornament of granules. In addition the spine tips terminate by constricting suddenly to a hair and never break up into small digitations. Since the species was originally described on the basis of two specimens (Downie 1959) others have been found which indicate that it would be desirable to extend the original diagnosis to include forms whose body size ranges from 20 to 40 microns with six to twelve spines visible in optical section, their length varying from 20 to 100 per cent, of the text diameter. TABLE 11. Recorded occurrences of Baltisphaeridium spp. belonging to group (c) Locality BS/I BS/3 BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 B. meson group (including B. oligo- fitrcatum and B. brevifurcatum B. dilatispinosum .... 1 6 3 1 1 5 1 (c) Forms with dark spherical bodies and pale-coloured short processes. Only a few speci- mens of this type have been found but they are of some interest since they are very similar to the forms described from the Upper Visby Marl (Upper Llandovery) of the Baltic by Eisenack (1954). From there he described three intergrading species, H. brevi- furcatum, H. intermedium (^ B. meson), and //. oligafurcatum (PI. 92, fig. 9). Examples of all three have been found in the Wenlock Shales together with a new species B. dilatispinosum which resembles B. piriferum (Eisenack) from the Upper Visby Marl. These hystrichospheres are too few in number to make any definite assertions about their stratigraphic range (Table 1 1) but so far B. dilatispinosum has been found only at locality BS/3 and the others are mainly concentrated in the Buildwas Beds. Baltisphaeridium meson (Eisenack) Remarks. Since allocating certain forms from localities WS/A and WS/2 to B. cf. meson (Downie 1959) because of their smaller size, other speeimens have been found indieating that they are inseparable from the typical Baltic forms. 642 PALAEONTOLOGY, VOLUME 6 Baltisphaeridium dUatispinoswn sp. nov. Plate 92, fig. 4 Diagnosis. Body spherical, walls stout and dark, processes pale, relatively short, dis- tally inflated. The distal part bears numerous small spines. Dimensions. Body size 50 to 60 microns; process length 10 to 12 microns, width 6 to 8 microns, number in optical section fourteen to twenty-two; spines on processes 1 to 1-5 microns, 2 microns apart. Holotype. Slide 3, position 190.480. Locality. Wenlock Shales, BS/3. Remarks. Three individuals of this species were included in statistical counts of the assemblages but a number of others were seen. All were confined to samples from locality BS/3. The general shape of the test and the processes shows a close resemblance to B. pirifemm (Eisenack) which does not, however, have any small spines ornamenting the bulbous processes. {d) Forms with smooth branching spines. A great variety of acritarchs with branching processes occur in the Wenlock Shale. Some have already been discussed in previous sections; the remaining ones are considered here. They include B. ramusculosum (Deflandre), B. eoplanktonicum (Eisenack), B. digitatum (Eisenack), B. ravum sp. nov., B. cladum sp. nov., B. arbuscidiferum sp. nov., and B. microcladiim sp. nov., as well as some other varieties so far too rare to merit description. In all about 300 examples were found, their abundance decreases markedly upwards. B. eoplanktonicum, originally described from the Ludlow by Eisenack, was repre- sented by forty-five specimens and occurred in similar proportions throughout the Wen- lock Shales. B. ramusculosum was more common with 150 examples and was more numerous in the lower horizons, especially near the base of the Coalbrookdale Beds. Since it was first described by Deflandre from the Wenlock it has been recorded from the Devonian by DeunflF. B. cladum (6 examples) and B. ravum (34 examples) are restricted to the Buildwas and lowest Coalbrookdale Beds. B. microcladum (26 examples) has a similar distribution but isolated examples have been found as high as locality WS/5. B. arbusculiferum (12 examples), on the other hand, has been found only in the Tickhill Beds. B. digitatum (5 examples) has been found only at high levels but it is rare and it clearly has a wider range since it has been found by Eisenack in the Llandovery of the Baltic as well as in the Wenlock Limestone. Their distribution is shown in Table 12. EXPLANATION OF PLATE 92 Figs. 1-3, 11-12. Lophosphaeridium spp. 1, L. triangulatum sp. nov.; holotype, X900; locality WS/4. 2, L. pilosum sp. nov.; holotype, x660; locality WS/4. 3, L. citrimim sp. nov.; holotype, x660; locality WS/4. 11, L. microspinosum (Eisenack), x540; locality BS/3. 12, L. sp. cf. P. papiltatum Staplin, X 900; locality WS/4. Figs. 4-6, 9. Baltisphaeridium spp. 4, B. dilatispinosum sp. nov.; holotype, x500; locahty BS/3. 5, B. cladum sp. nov.; holotype, X 1,000; locality BS/3. 6, B. microcladum sp. nov.; holotype, X 540; locality BS/3. 9, B. oligofurcatum (Eisenack), X 540; locality BS/3. Figs. 7, 10. Veryhachium spp. 7, V. trisphaeridiiuyi sp. nov.; holotype, X 1000; locality WS/8. 10, V. elougatum sp. nov.; Holotype, X 1000; locality WS/7. Fig. 8. Estiastra barbata sp. nov.; holotype, X 500; locality BS/3. Fig. 13. Punctatisporites? dilutus Hoffmeister, X700; locality WS/4. Palaeontology , Vol. 6 PLATE 92 DOWNIE, Silurian ‘hystrichospheres’ and spores r X- r. V CLUC jm- C. DOWNIE; ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 643 Baltisphaeridium eoplanktonicum (Eisenack) Remarks. As well as typical forms of this species with 4 spines one example was found with 3, one with 5, and five with 6 spines. These forms differed from the typical ones in no other respect. TABLE 12. Recorded occurrences of Baltisphaeridium spp. belonging to group {d) Locality BS/l BS/3 BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 WS/2 WS/5 WS/8 WS/4 B. ravum ..... 3 29 2 B. eoplanktonicum .... 6 3 6 1 3 3 24 7 2 1 B. ramusculosum .... 29 19 2 36 28 6 1 2 19 2 5 1 B. cladum ..... 3 2 1 B. microcladurn .... 13 6 1 3 1 1 1 B. arbusculiferum .... 2 10 B. digitatum ..... 3 1 1 Baltisphaeridium ravum sp. nov. Plate 91, fig. 6; text-fig. 3c Diagnosis. Test spherical rather dark greyish-yellow in colour, processes long cylindrical, pale coloured, flaring at the tips and breaking up into numerous digitations. Dimensions. Text diameter 30 to 55 microns; process lengthbody size 60 to 100 per cent.; number of processes in optical section six to twelve, width of processes 3 to 4 microns, length of digitations 3 to 4 microns. Holotype. Slide 4, position 200.450. Locality. Wenlock Shales, BS/2a. Remarks. This species resembles B. ramusculosum but is larger, the contrast it shows between body and process colour and texture is another distinction. The style of branch- ing also differs (text-fig. 3c), B. ramusculosum in general having wide-angled regular bifurcations, often to the fourth order, whereas B. ravum has less frequent narrow- angled bifurcations and irregular distal digitation. TABLE 13. Recorded occurrences of Estiastra, Pulvinosphaeridium, and Polyedrixium Locality BS/l BS/3 BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 WS/2 WSI5 WS/8 WS/4 E. barbata ..... E. granulata ..... Pulv. pulvinellum .... Polyedrixium spp. .... 25 17 8 1 5 1 Baltisphaeridium cladum sp. nov. Plate 92, fig. 5 ; text -fig. 2>a Diagnosis. Test slightly ellipsoidal, several processes with stout shanks tapering to point of forking, forks fairly wide angled, short, tips irregularly bifurcate or trifurcate. 644 PALAEONTOLOGY, VOLUME 6 Dimensions. Body size 18 to 25 microns; spine number in optical section ten to fourteen, spine length about 65 per cent, of test diameter. Holotype. Slide 3, position 030.470. Locality. Wenlock Shales, BS/3. Remark. This species differs from B. ramusculosum in the slightly ellipsoidal test, the stout tapering processes and the more limited and irregular bifurcations at the tips. TLXT-FiG. 3. Illustration of the types of processes occurring in some species of acritarchs. a, Balti- sphaeridium cladum (12 ^ long); b, B. eoplanktonicinn (50 p, long); c, B. gramdatispinosum (60 p. long); d, B. arbusculiferiim (20 p long) ; e, B. ravum (40 p long) ; f, B. ramusculosum (20 p long) ; g, B. micro- cladum (20 p long). Bahisphaeridium arbusculiferiim sp. nov. Plate 91, fig. 5; text-fig. hd Diagnosis. Test subspherical to subpolygonal, processes long broad tapering, forking irregular at a moderate angle, branches often long and broad ; forking usually bifurcate up to fourth order. Dimensions. Test size 15 to 30 microns; spine number in optical section six to nine, spine length 100 to 140 per cent, of test diameter. Holotype. Slide 2, position 010.560. Locality. Wenlock Shales. WS/4b. Remarks. This species bears some resemblance to B. digitatum in the broad tapering form and mode of branching of the processes but differs in its smaller size and in the clear separation of the processes from the test. C. DOWNIE: ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 645 Baltisphaeridiiim microcladum sp. nov. Plate 91, fig. 3, Plate 92, fig. 6; text-fig. 3^ Diagnosis. Test slightly elhpsoidal, smooth or granular, spines moderately long and numerous, narrow and slightly tapering. Forking only at the tips, bifurcate, trifurcate, or quadrifurcate, branches very short and thin but second-order branching may occur. Dimensions. Test size 20 to 45 microns; number of processes in optical section fifteen to twenty, length 30 to 50 per cent, of test diameter, branches 1-0 to 2-5 microns long. Holotype. Slide 3, position 145.540. Locality. Wenlock Shales BS/3. Remarks. This species resembles Midtiplicisphaeridiuml sprucegrovensis Staplin but is smaller, has somewhat fewer processes, and shows greater variation in the style of branching. (2) Acritarchs with small spherical bodies bearing spines This group is represented by the genus Micrhystridium, of which several hundred examples were found spread throughout the Wenlock Shales. They formed a relatively small proportion of the total assemblage except in the higher levels where they became TABLE 14. Recorded occurrences of Micrhystridium Locality BS/l BS/3 BS/2 WS/7 WS/6 WS/11 WS/10 WS/12 WS/1 WS/2 WS/5 00 WS/4 M. stellatum ..... 1 1 2 12 2 5 5 49 11 27 IS M. eatonensis .... 3 14 3 M. nannacanthum .... 12 8 4 8 1 1 6 20 48 52 M. parinconspicuum 7 2 4 6 9 4 1 32 10 18 97 M. imitatum ..... 2 1 3 I 20 5 26 15 abundant forming between 10 and 20 per cent, of the sample. The species present were M. stellatum Deflandre, M. imitatum Deflandre, M. parinconspicuum Deflandre, M. nannacanthum Deflandre, and M. eatonensis Downie. They all appear to range through- out and to become more common in the upper beds. A number of other small hystricho- spheres were present but insufficiently well preserved or not present in large enough numbers to merit description. Their distribution is shown in Table 14. Subgroup PRiSMATOMORPHiTAE Downie, Evitt, and Sarjeant 1963 This group of Acritarchs with polygonal tests bearing flanges on the edges of the poly- gons is represented by the genus Polyedrixinm, and is very rare in the Wenlock Shales. Only two examples have been found, one quite unidentifiable specifically; the other, slightly better preserved, closely resembled P. centrigerum Deunff from the Middle Devonian. Hitherto Polyedrixinm has been known only from the Devonian where it appears to be common at certain levels ; the Wenlock specimens occurred near the top (Table 13). C 1713 u u 646 PALAEONTOLOGY, VOLUME 6 Class CHLOROPHYCEAE ORDER UNCERTAIN Tasmanites appears to be present in small numbers throughout the succession (Table 3), never forming more than about 1 per cent, of the whole assemblage. It is represented mainly by Tasmanites medius (Eisenack) and a similar form slightly smaller in size (Downie 1959). STRATIGRAPHIC RANGE OF WENLOCK SHALE ACRITARCHS (u) Forms confined to the Wenlock. As far as is known thirty-eight species or varieties among the forms occurring in the Wenlock Shale have not been found at other horizons. Of these C. pavimenta, B. brevispinosum var. namim, M. imitatum, M. nannacanthum, and V. bulbifernm have also been found in the Wenlock of France (Deflandre 1945). (b) Forms ranging from the Lower Silurian into the Wenlock. B. meson and L. micro- spinosnm occur in the Upper Llandovery of the Baltic (Eisenack 1954), P.? dilntiis in the Lower Silurian of Libya (Hoffmeister 1959), and M. parinconspicnum in the Middle Silurian of New York (Fisher 1953). P. pidvinellnm and B. ramscidosum seem to range throughout the Silurian. (c) Forms ranging from the Wenlock into the Ludlow. P. pidvinellnm, B. eoplanktonicum, and T. medius have been recorded from the Ludlow but not from higher horizons (Eisenack 1955). {d) Forms ranging into the Devonian. These include P. onondagaensis, C. cubus, C. pris- matica, Polyedrixium, and B. ramusculosum (Deunff 1954, 1955) and L. wenlockia, L. diaphana, L. cryptogranulosum, L. granulosum, and L. papillatum (Staplin 1961). (e) Forms with longer ranges. A small number of species have longer ranges than any of the above. B. digitatum ranges from the Caradoc to the Wenlock Limestone and L. microsystis from the Ordovician at least to the Wenlock. V. trispinosum and M. stellatum show subtle variations in morphology over a long period of time producing a number of only slightly distinct varieties. The former ranges from the Ordovician at least to the Devonian, the latter from the Wenlock to the Upper Jurassic. (/) Comparison with the Wenlock assemblage from the Montague Noire. Deflandre (1945) described sixteen forms from the Wenlock of France. Of these ten have been identified in the Wenlock of Shropshire; the other six {H. staurasteroides, H. geometricum, M. mendax, M. tenuissimum, L. retigera, and L. aurata) were not seen or could not be cer- tainly identified. Of the various Wenlock assemblages those in the Coalbrookdale Beds compare most closely with that described by Deflandre. (g) Comparison with Silurian assemblages from the Baltic. Eisenack (1954, 1955, 1959) has described ‘hystrichospheres’ from a number of horizons in the Silurian of the Baltic regions. From the Visby Marl (Upper Llandovery) he has recorded B. piriferum, B. brevifurcatum, B. meson, B. oligofurcatiim, B. gotlandicum, B. erraticum, B. polygonale, B. digitatum, B. visbyense, E. magna, L. microspinosum, and P. pidvinellnm. Of these B. brevifurcatum, B. meson, B. oligofurcatiim, B. digitatum, L. microspinosum, and P. TEXT-FIG. 4. Chart showing the stratigraphical range of some species of acritarchs. 648 PALAEONTOLOGY, VOLUME 6 piilvilielliim have been found in the Wenlock of Shropshire and forms very close to B. piriferuin, B. eiraticwn, B. visbyense, and E. magna also occur. All ten of the forms identical or similar to the Baltic species occur in the Buildwas Shales, only six in the Coalbrookdale Beds and only two in the Tickhill Beds. The greatest similarity is there- fore to be found in the earliest beds of the Wenlock Shales. From the Wenlock of the Baltic Eisenack has recorded B. polygonale, B. visbyense, B. corallinum, B. lophophorum, and P. pulvinelhmi of which only P. pulvinellum has so far been found in Shropshire. In the Beyrichia Limestone (Ludlow) Eisenack found the following species B. eoral- Hmim, B. erroticiim, B. eoplanktonicwn, B. meson, L. cf. microspinosum, L. media, P. aff. pulvinellum, Dictyotidiiim dictyotum, and D. tenuiornatum. Of these B. meson, B. eoplanktonicwn, L. medio, L. microspinosum, and P. pulvinellum probably range all through the Wenlock Shale but L. microspinosum and P. pulvinellum have not been found in the highest beds. Eorms like B. erraticum also occur {B. dilatispinosum) but they are confined to the Buildwas Shales. Nothing resembling B. corallinum or the two species of Dictyoiidium has been found, and there is therefore no close comparison with any part of the Wenlock Shales, only three species being directly comparable. COMPARISON OF ASSEMBLAGES WITHIN THE WENLOCK SHALE It has been found that several species range throughout the Wenlock Shales without much change in the numbers present. These are V. europaeum var. wenlockium, V. trispinosum, B. eoplanklonicunu all the varieties of B. brevispinosum, all the species of Cymatiosphaera, L. granulosum, L. citrinum, L. cf. papillatum, and L. diaphanum. These play no part in giving character to the different assemblages. The other forms present either have a restricted range or vary markedly in their abundance. On the basis of these latter forms the Wenlock Shale assemblages can be divided into three types. Assemblage type 1. All the samples from the Buildwas Beds (BS/la, BS/lb, BS/2a, BS/2b, and BS/3) yielded assemblages of this type. They contain the greatest number of forms like those described from the Upper Llandovery of the Baltic, in particular L. micro- spinosum, the B. meson complex {B. oligofurcatum, B. meson, and B. brevifurcatum), and B. dilatispinosum. In general these asemblages have relatively few specimens of Micrhystridium (IT to 3-9 per cent.), few leiofusids (0-2 to 1-3 per cent.), few specimens of B. granulatispinosum (1-6 to 3-9 per cent.), and few spores (0 to 0-2 per cent.) (see Table 15). On the other hand, acritarchs of the B. longispinosum type are relatively common (8-5 to 15 per cent.), as are those with branching processes (4T to 10-5 per cent.). The most striking feature, however, is the presence of Deunjfia and Domasia which are frequently abundant and quite absent from the other assemblage types. Samples from BS/la and BS/lb have the general features of assemblage type 1 but leiospheres are unusually abundant (61-6 per cent.) D. brevispinosa and D. ramusculosa are characteristic: B. microclodum is relatively abundant. B. dilatispinum, B. cladum, and B. rovum are absent. Sample BS/3 is similar to samples from BS/la and BS/lb but leio- spheres are less common (31-2 per cent.). D. brevispinosum and D. ramusculosum are TABLE 15. Occurrence of microfossils as percentages C. DOWNIE; HYSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 649 BS/1 r4 i ! ri o r4 o ^ ^ O O o ! ! so ^ ! ^ so 0001 BS/3 ^ ^ r- so r-i oo ^ qs ri C4 oo oo n r4 ri ON ri r4 o O ri o r 1 — m 1000 BS/2 ON CO SO r-1 lo 00 rn *-r> O *^1 r- r j ro ! ! !r^4r^oo6s'^cbuno i ! n — — rt 1000 WS/7 Osrnriirisnr-^ — so^j-o^oi ! r-i o cb so i r\ ! o rb so — — — Tj- 1000 WS/6 O ^ ^ ^ T* r4 Os so . . .sb"^ .sbsbo . ^ — < — m 1000 WS/11 O *^0O -+05'/^ Tj-Orn — in ‘ ^ i Os cs . • ’— ! o ri rb ^ — ri r-i 0001 WS/IO O'! r-i so lo Tj- rn in m> c> m in >n i i ri so ^ — i O ri r-i o r\ rt n[ 0001 WS/T2 — rjm sotJ-*— r-iu~i — "^O rn . 1 i*^0 .risbrb . .-^O irnori — so 1000 WS/1 M 98-9 0 001 WS/2 r-4 Tj* mrorirnCNSO ^ m rn rn m in so ! O ! lb rt o r-i ob — ' * i ri o ri so O so 0001 WS/5 ^ ^ _ O in r^oori _ .’^'O ooso'?j- .rl^ds • .^o .mrj^ 1000 WS/8 OO’-H ^ ^ _qs-^r^i/Dsooo r^O * iboo'^-^ri i !ooo-^sb^ 1— c r4 m 100-0 WS/4 1 rl »-H in so ro ro O r-i Tf 1— fo oo n obo iobob*^ ^ -H m 100-0 Micrhystridium Polyedrixium . Pulvinosphaeridium . Estrastra Baltisphaeridium Gp. (a) . Baltisphaeridium Gp. (b) . Baltisphaeridium Gp. (c) . Baltisphaeridium Gp. (ci) . Veryhachium . Leiofusa .... Deunffia .... Domasia .... Cymatiosphaera Pterospermopsis Tasmanites Lophosphaeridium . Leiosphaeridia . Spores .... Total .... 650 PALAEONTOLOGY, VOLUME 6 absent; B. dilatispitwswn, B. cladum, and B. ravum are present. Samples from BS/2a and BS/2b are similar to samples from BS/3 but D. trispinosa, D. bispinosa, E. barbata, and B. dilatispinosum are absent. B. microcladwn is rare. D. elongata and D.fiircata are characteristically present. Assemblage type 2. Samples from localities WS/2a, WS/2b, WS/6a, WS/6b, WS/7a, WS/7b, WS/A, WS/10, WS/1 1, and WS/12 yielded assemblages of this type. All of these are from the Coalbrookdale Beds. The samples from localities WS/la, WS/lb, and WS/lc although they contain only leiospheres are included here provisionally for it is not yet certain that their peculiar features are due to original differences in the plankton assemblages. However, in quoting proportions of various types present, samples from this locality are ignored. The features of this assemblage are mainly negative and it grades into assemblages of type 1 which preceed it and those of type 3 which succeed it. Assemblages of this type compare most closely with those described from the Wenlock of France. Assemblages of type 2 differ from those of type 1 in that Deimjfia, Domasia, and Estiaslra are absent, Micrhysthdium is more common (1-3 to 6-2 per cent.), leiofusids are also more common in general (0-2 to 9-5 per cent.) and L. tumida is present. Very- hachium is more common (8-9 to 29-9 per cent.) and also spores (0-2 to 5T per cent.). On the other hand, B. meson, L. microspinosum, and M. microcladum are very rare. Acritarchs with branching processes are less common, especially in samples from the higher horizons (there is a decline from 8-5 to 2-3 per cent, in their contribution to the assemblage). B. longispinosum is also less common (0-6 to 6-0 per cent.). The only forms thought to be characteristically restricted to this assemblage are V. bidbiferum and V. elongatum: the latter, however, occurs only in samples from locality WS/7 at the base of the Coalbrookdale Beds. Samples from WS/7a, WS/7b, Ws/6a, WS/6b, WS/11, and WS/10 (i.e. the Lower Coalbrookdale Beds) have the general features of assemblages of type 2, but are distinct in the relatively small proportion of leiospheres (22-3 to 46 per cent. ). On the other hand, acritarchs with branching processes and those of the B. longispinosum type are present in relatively high proportions, this and the presence of small numbers of B. ravum and B. clavum are reminiscent of the assemblages of type 1. V. elongata occurs but only in samples from locality WSpa. and WS/7b. Samples from localities WS/12, WS/2a, WS/2b, WS/A (i.e. Upper Coalbrookdale Beds) are similar to those from the Lower Coalbrookdale Beds but the proportion of leiospheres is very high (60-4 to 66-3 per cent.) ; the proportion of acritarchs with branch- ing processes is small (2-3 to 2-5 per cent.) and also that of B. longispinosum 0-6 to 1-0 per cent.); B. cladum, B. ravum, and V. elongata are absent. Assemblages from localities WS/la, WS/lb, and WS/C at a similar stratigraphic level are unique since they consist exclusively of sphaeromorphs. Assemblage type 3. Samples from localities WS/5a, WS/5b, WS/4a, WS/4b, WS/8a, and WS/8b yielded assemblages of this type. These localities occur in the Tickhill Beds and near the top of the Coalbrookdale Beds. In general this assemblage type differs from type 2 because of the larger proportion of Micrhystridum (IIT to 18-2 per cent.), leiofusids (2T to 12-2 per cent.), and B. granu- latispinosum (7-3 to 14-3 per cent.). Acritarchs with branching processes are few (1-3 to C. DOWNIE: ‘H YSTRICHOSPHERES’ (ACRITARCHS) AND SPORES 65! 2-7 per cent.) and so is B. longispinosum{\-^ to 6-2 per cent.). V. buJbifenuri, V. ehngatuw, B. cladum, B. ravwn, and L. microspinosum have not been found but B. arbuscidiferum, K trisphaeridiimi, L. cf. cryptogramdosum, L. pdosum, E. gramdata, and Polyedrixium appear for the first time. Samples from WS/5a and WS/5b have the general features of assemblage type 3 but B. orbuscidiferum, V. trisphaeridium, L. cf. cryptogramilosiim, L. pdosum, and E. gramdata have not been found. B. arbusculiferum and V. trisphaeridium appear in samples from WS/8a and WS/8b and L. tumida becomes more common (Table 5). Assemblages from WS/4a and WS/4b also contain L. cf. cryptogramdosum, L. pdosum, and E. gramdata and leiofusids intermediate in form between L. fdifera and L. tumida are abundant (Table 5). CONCLUSIONS The acritarch assemblages of the Wenlock Shales show a more or less progressive change throughout the sequence but may be divided into three main types which succeed each other in stratigraphic order. Type 1 characterizes the Buildwas Beds, type 2 the Coalbrookdale Beds, and type 3 the uppermost Coalbrookdale Beds and Tickhill Beds. Insufficient data is at present available to show over what area this succession of assemblage types persists and an assessment of their value in correlation must wait until several other Wenlock successions have been examined. At present all that can be said in this respect is that the assemblages from the Buildwas Beds have their closest parallel in the Upper Llandovery of the Baltic and those from the Coalbrookdale Beds in the Wenlock of France. All the holotypes are at present in the collections of the Micropalaeontology Labora- tory of the Department of Geology at the University of Sheffield. Acknowledgement. The author wishes to acknowledge his indebtedness to the University of Sheffield Research Fund for a grant in aid of the expense of collecting the samples. REFERENCES DEFLANDRE, G. 1945. Microfossiles des calcaires siluriens de la Montague Noire. Ann. Paleont. 31, 41-75, pi. 1-2. DEUNFF, j. 1954. Sur le microplancton du Gothlandien armoricain. C.R. Soc. geol. France, 54-55. 1955. Un microplancton fossile devonien a hystrichospheres du continent Nord-Americain. Bull. Microscop. app. 2, 5, 138-47, pi. 1-4. DOWNIE, c. 1959. Hystrichospheres from the Silurian Wenlock Shale of England. Palaeontology, 2, 1, 56-71, pi. 10-12. 1960. Deiinffia and Domasia, new genera of hystrichospheres. Micropaleontology, 6, 2, 197-202, pi. 1. EViTT, w. R., and sarjeant, w. a. s. 1963. Dinoflagellates, hystrichospheres and the classification of the acritarchs. Stanford tJniv. Geol. Papers, 7, 3, 3-16. EiSENACK, a. 1954. Hystrichospharen aus dem baltischen Gotlandium. Senckenbergiana, 34, 205-1 1, pi. 1. • 1955. Chitinozoen, Hystrichospharen und andere Mikrofossilien aus dem Beyrichia-Kalk. Ibid. 36, 157-88, pi. 1-5. 1958. Tasmanites Newton and Leiosphaeridia n.g. als Gattungen der Hystrichosphaeridia. Palaeontographica, A 110, 1-3, 1-19, pi. 1-2. 1959. Neotypen baltischer Silur-hystrichospharen und neue Arten. Ibid, A 112, 5-6, 193-211, pi. 15-17. 652 PALAEONTOLOGY, VOLUME 6 Evrrr, w. r. 1963. A discussion and proposals concerning fossil dinoflagellates, hystrichospheres and acritarchs. U.S. Nat. Acad. ScL, Proc. 49, 158-64; 298-302. FISHER, D. w. 1953. A microflora in the Maplewood and Nehaga Shales. Buffalo Soc. Nat. Hist. Ball. 21, 2, 13-18. HOFFMEiSTER, w. s. 1959. Lower Silurian Plant spores from Libya. Micropaleontology, 5, 3, 331-4, pi. 1. POCOCK, R. w., et al. 1938. Shrewsbury District. Mem. Geol. Surv. U.K. STAPLiN, F. L. 1961. Reef-controlled distribution of Devonian microplankton in Alberta. Palaeon- tology, 4, 3, 392-424, pi. 48-51. TiMOFiEV, B. V. 1959. The ancient Baltic flora and its stratigraphical significance. Trudy VNIGRl, pp. 136, 25 pi. [in Russian.] WALL, D., and DOWNiE, c. 1963. Hystrichospheres from the Permian. Palaeontology, 5, 4, 770-84, pi. 112-14. C. DOWNIE Department of Geology, University of Sheffield Manuscript received 14 November 1962 THE GENERA BROTZENIA AND VOORTHUYSENIA (FORAMINIFERA) AND HOFKER’S CLASSIFICATION OF THE EPISTOMARIIDAE by W. G. GORDEY Abstract. Perfectly preserved specimens of the Rotaliid species Brotzenia mosquensis Uhlig 1883 and B. para- stelligera Hofker 1954 from the Oxford Clay of England, show dental plates of the Voorthuysenia type. These species were assigned by Hofker, on the basis of the dental plates, to Brotzenia. Hofker’s (1954) classification is therefore invalid. The genus Voorthuysenia is considered synonymous with Brotzenia and the latter is emended. The genera Hdglandina Brotzen 1948 and Hiltermannia Hofker 1954 are briefly discussed in relation to Hofker’s classification. Hofker s classification of the Epistomariidae. Hofker (1954) considered the genus Episto- niina to be invalid, stating that the type species E. regularis (Terquem 1883) did not fit this group as it is understood at present, the type species showing neither the charac- teristic protoforaminal and deuteroforaminal apertures or dental plates (text-fig. \a-c). Hofker (1954) suggested that Terquem (1883) may have described badly damaged speci- mens, or possibly that Terquem’s specimens belonged to the genus Couorboides Hofker 1954. He therefore proposed a complete revision of the foraminifera with dental plates previously assigned to Epistomina, erecting three new genera, Brotzenia and Voorthuy- senia (Jurassic-Lower Cretaceous), and Hiltermannia (upper Lower Cretaceous), and also including Brotzen’s genus Hdglandina (Cretaceous-Recent). The key to the recogni- tion of these genera (and the basis of the classification) was the form of the dental-plate. Hofker’s classification has not been widely adopted by workers on Jurassic foramini- fera, e.g. Lutze (1960), Bielecka (1960). However, Seibold and Seibold (1960), Gordon (1962), and Lloyd (1962) have used it, but with the exception of Lloyd make no reference to the use of thin sections. However, Lloyd’s figures of B. porcellanea ( Bruckmann), parti- cularly text-fig. 7d, are of limited value since only part of the dental-plate is shown and the species could therefore belong either to Brotzenia or Voorthuysenia. Briefly, the characteristics of the tooth-plates of the genera referred to are as follows: Brotzenia. Tooth-plates in all chambers, narrow, smooth, partly attached to the proximal chamber wall, not reaching the marginal distal angle. Type species : Episto- mina spinulifera Reuss. Voorthuysenia. Plates in all chambers, large, smooth, attached entirely along the proximal chamber wall, reaching the marginal distal angle. Type species; Epistomina tenuicostata Bartenstein. Hiltermannia. Plates as in Brotzenia but occurring in the final chamber only, resorbed in earlier chambers, plates frequently pustulose (‘Hocken’). Type species: Epistomina chapmani Ten Dam. Hdglandina. Plates as in Voorthuysenia, occurring only in the final chamber, resorbed in earlier ones, plates frequently pustulose. Type species : Epistomina elegans d’Orbigny. Foraminifera with dental plates from the Oxford Clay of England. The author’s material [Palaeontology, Vol. 6, Part 4, 1963, pp. 653-7, pi. 93.] 654 PALAEONTOLOGY, VOLUME 6 was obtained from two exposures, one at Stewartby (Bedfordshire) in the Lower Oxford Clay and the other at Burton (Wiltshire) in the Upper Oxford Clay (see Arkell 1941). Approximately 250 specimens of Brotzenia mosqiiensis were obtained in several samples from the Burton section, being exceptionally well preserved, particularly those obtained from Beds 4 and 5 (of Arkell 1941); and about 300 specimens of B. parastelligera in several samples from both the Burton and Stewartby sections. As most of the specimens were not infilled, a three-dimensional view of the plates could be obtained by embed- ding them in a resin. Lakeside 70, and removing the dorsal surface with a sharp needle. This contrasts with the frosted-glass method (van Morkhoven 1958) used by Hofker, where preservation did not permit the former technique. TEXT-FIG. 1. Diagram illustrating the morphology of Brotzenia. a. Parallel section; b, peripheral view with the terminal face of the final chambers removed; c, ventral view. Abbreviations', mda — marginal distal angle; dp — dental-plate; pew — proximal chamber wall; pf — protoforamen; df — deuteroforamen. The dental-plates in all the specimens of B. ivosquensis and B. parastelligera occur in all chambers (except where damaged in preparation). They are always smooth and attached along the entire length of the proximal chamber wall (text-fig. 2a-e), and extend into the distal marginal angle of each chamber. The plates correspond exactly to the type described by Hofker (1954) for the genus Voorthuysenia. However, B. mosquensis and B. parastelligera were assigned by Hofker to the genus Brotzenia. It would thus appear that these two species, at least, are capable of possessing two distinct forms of dental plate, i.e. in the case of Hofker’s 1954 material the Brotzenia type of plate, and in the present material the Voorthuysenia type. These conclusions agree with those made by the author (1962) on material from the Oxford Clay of Skye. In view of the better preserved and more abundant material now available, the author considers that Hofker’s genera Brotzenia and Voorthuysenia are not distinct and that his classification (1954) is invahd. It is therefore proposed that the genus Voorthuysenia be considered synonymous with Brotzenia and the latter emended to include it. With regard to the author’s ornamented forms (i.e. B. mosquensis), Hofker (in litt.) states that on the basis of the plates they should be assigned to Voorthuysenia praeornata W. G. GORDEY: BROTZENIA AND VOO RTHUYSENIA 655 (Bartenstein and Brand 1951), the two species being identical externally. However, the author has examined paratypes of F. praeonuita (PL 93, figs. 1-3) and the species can be distinguished from B. inosquensis (PL 93, fig. 4n-c) in having sharply convex dorsal sutures; distinct pustulose ornament on each chamber; and less intense sutural and intersutural ornament. Furthermore, the author has compared his specimens with some TEXT-FIG. 2. a, b, Brotzenia parastelligera Hofker 1954. a. An eroded specimen showing dental plates of the ‘Voortluiyseiiia type’ in two chambers, coronatum Zone, Oxford Clay, Stewartby, Bedfordshire, England. BMNH P45156, X60. b. Parallel section showing plates in several chambers, cordatiim Zone, Oxford Clay, Purton, Wiltshire, England. BMNH P45159, x65. c-e, Brotzenia mosquemis (Uhlig 1883). Cordatiim Zone, Oxford Clay, Purton, Wiltshire, England c. Peripheral view showing the dental-plate, deuteroforamen situated on the dorsal side of the dental plate, BMNH P45157, x55. d. Parallel section showing well-developed plates in the earliest chambers, BMNH P45 1 58, X 57. e. Parallel section, earlier chambers removed in preparation, BMNH P45160, X57. of Hofker’s (1954) material, which is identical. Specimens of this species loaned by Lutze are also identical, though preservation did not permit a study of the plates. Specimens identified by Hofker and Lutze as Brotzenia parastelligera have also been examined, and are identical to the author’s specimens. The Oxford Clay material also agrees closely with the figures and descriptions of this species by Seibold and Seibold (1960), Bielecka (I960), and Lutze (1960). SYSTEMATIC DESCRIPTION Genus brotzenia Hofker 1954 Type speeies. Brotzenia spinidifera (Reuss). Emended diagnosis. Test calcareous, perforate, trochoid, dextrally or sinistrally coiled, smooth or ornamented, having latero-marginal apertures (protoforamen) (text-fig. 1) on 656 PALAEONTOLOGY, VOLUME 6 the ventral side of the test in each chamber, together with areal apertures (deutero- foramen or septal foramen) on the dorsal side of the dental plate. The dental plates are smooth, usually convex towards the ventral side, the plates being either attached along the entire length of the proximal chamber wall or only partly so ; plates occurring in all chambers. The following species are included in the genus Brotzeuia : Brotzenia mosquensis (Uhlig 1883) Brotzeuia parastelligera Hofker 1954 Brotzenia ornata (Roemer 1841) Brotzeuia spiuitlifera (Reuss 1862) Voorthuyseuia praeoruata (Bartenstein 1951) Voortiniyseuia braudi Hofker 1954 Voorthuyseuia parafavosoides Hofker 1954 Voorthuyseuia suturalis (Ten Dam 1948) Voorthuyseuia teuuicostata (Bartenstein 1951) Voorthuyseuia pachyderma Hofker 1954 Remarks. In the Cretaceous species caracolla {Epistomina of earlier authors, HdgJandina Hofker 1954) from the Speeton Clay of Lincolnshire, England, the author observed plates in chambers other than the final chamber. This throws doubt on Hofker’s defini- tion of Hdglandina, and it is possible that species assigned to this genus should be referred to Brotzenia (here emended). The distinction between HdgJandina and Hiltermannia on the basis of dental plates is similar to that between Brotzenia and VoortJniysenia. The fact that the distinctions between the latter genera cannot be sustained casts some doubt on the validity of the genus Hiltermannia. However, the species assigned to this genus have not been studied. Even if Hofker’s tooth-plate groups were proved to be valid, his classification, being dependent on the study of the dental plates, is impracticable, since preservation often prevents the plates being exposed, making it impossible to assign specimens to a genus. Vella (1961) commented on the practical difficulties of such a classification with respect to uvigerinid foraminifera, where, however, it was possible (Vella 1961, p. 468) to use external morphology to recognize the dental plate groups. Unfortunately, this is of limited value only in the group considered here since many of the species have a similar external morphology, e.g. HdgJandina caracoJJa (Roemer) and Brotzenia parasteJJigera Hofker. AHnowIedgeiueuts. The author is indebted to Dr. T. Barnard and Dr. C. G. Adams for criticism and for reading the manuscript; to Dr. J. Hofker for discussing the problem and examining some of the EXPLANATION OF PLATE 93 Figs. 1-3. Brotzeuia praeoruata (Bartenstein and Brand 1951). 1, Dorsal view. 2, Peripheral view. 3, Ventral view. Three specimens, paratypes, Bohrung Diiste 1, 510-8 — 516-3 m., Upper Valanginian, X 1 10. Dr. H. Bartenstein personal collection. Fig. Aa-c. B. mosquensis (Uhlig 1883). a. Dorsal view; b, peripheral view; c, ventral view. Cordatuiu Zone, Oxford Clay, Purton, Wiltshire, England. BMNHP45161, x 67. Fig. 5a, Zi. B. parastelligera Honker 1954. a, Dorsal view; 6, ventral view. Ma/-/ap Zone, Oxford Clay, Purton, Wiltshire, England. BMNHP45162, x52. Fig. 6a-c. B. parastelligera Hofker 1954. a. Dorsal view; b, peripheral view; c, ventral view. Cor- datum Zone, Oxford Clay, Purton, Wiltshire, England. BMNH P45163, X68. Palaeontology, Vol. 6 PLATE 93 CORDEY, Brotzenia W. G. GORDEY; BROTZENIA AND VOORTHUYSENIA 657 author's material; to Dr. J. H. van Voorthuysen for the loan of some of Hofker’s material; to Dr. H. Bartenstein for the loan of paratypes of B. praeornata and specimens of B. jnosquensis; and to Dr. G. F. Lutze for many specimens of B. parastelligera and B. mosqiiensis. Repository. All figured specimens, except those in Plate 93, figs. 1-3, are deposited in the British Museum (Natural History), London. REFERENCES ARKELL, w. ;. 1941. The Upper Oxford Clay at Purton, Wiltshire, and the zones of the Lower Oxfordian. Geol. Mag. 78, 161-72. BARTENSTEIN, H. and BRAND, E. 1951. Micropalaontologische Untersuchen zur Stratigraphic des nord-westdeutschen Valendis. Abh. Senckeiib. uaturf. Ges. 439, 1-224. BiELECKA, w. 1960. Micropalacontological stratigraphy of the Lower Malm in the vicinity of Chrzanov (southern Poland). Inst. Geol. Brace, 31, 1-155. (In Polish, with English summary, 103-55.) BROTZEN, F. 1948. The Swedish Paleocene and its foraminiferal fauna. Sverig. geol. Unders. Afh. Ser. C., 493, 3-140. CORDEY, w. G. 1962. Foraminifera from the Oxford Clay of Staffin Bay, Isle of Skye, Scotland. Senck. leth. 43, 375-409. GORDON, w. A. 1962. Some foraminifera from the Ampthill Clay, Upper Jurassic, of Cambridgeshire. Palaeontology, 4, 520-37. HOFKER, J. 1954. liber die Familie Epistomariidae (Foram.). Palaeontographica, lliS, \66-106. LLOYD, A. J. 1962. Polymorphinid, miliolid and rotaliform Foraminifera from the type Kimmeridgian. Micropalaeontology, 8, 369-83. LUTZE, G. F. 1960. Zur Stratigraphic und Palaontologie der Callovien und Oxfordien in Nordwest- Deutschland. Geol. Jb. 77, 391-532. MORKHOVEN, F. p. c. M. Van. 1958. A simple method of grinding Foraminifera. Micropalaeontology, 4, 209-10. SEiBOLD, E. and seibold, I. 1960. Foraminiferen der Bank- und Schwamm-Fazies im unteren Malm Siiddeutschlands. Nenes Jb. Geol. Paldont., Abh. 109, 309-438. terquem, o. 1883. Memoire sur les Foraminiferes du systeme oolithique. Mem. Acad. Metz, 5, 339-406. UHLiG, V. 1883. Ueber Foraminiferen aus dem rjasen’schen Ornatenthone. Jahrb. Geol. Reichsanst. 33, 735-74. VELLA, p. 1961. Upper Oligocene and Miocene uvigerinid Foraminifera from Raukamura Peninsula, New Zealand. Micropalaeontology, 7, 467-83. W. G. CORDEY Texaco Trinidad Inc., Pointe-a-Pierre, Trinidad, W.I. Manuscript received 9 January 1963 TWO NEW SPECIES OF DISCOCYCLINA (FORAMINIFERA) FROM THE UPPER EOCENE OF ASSAM, INDIA by B. K. SAMANTA Abstract. Two new species of the foraminiferal subgenus Discocyclina, Discocyclina (D.) assamica and D. (D.) eamesi are described and illustrated from Upper Eocene rocks of the Garo Hills, south-western Assam, India. Fossiliferous Eocene rocks are well developed in the Garo Hills, south-western Assam, India, and consist of three rock units, in ascending order the Tura Sandstone formation (Lower (?)-Middle Eocene), the Siju Limestone formation (Middle-Upper Eocene), and the Kopili Shale formation (Upper Eocene). The sequence is apparently conformable. Larger foraminifera occur abundantly both in the Siju Limestone forma- tion and in the Kopili Shale formation, and the subgenus Discocyclina is well represented. Along with several well-known species of Discocyclina s.s. described from southern Europe, Western India, and Indonesia, two new species of this subgenus are present, and are described here. Holotypes and other figured specimens are deposited in the collections of the Geology Department of the University of Calcutta (CUGD). Acknowledgements. The author is indebted to Dr. F. E. Fames for valuable information on Indian discocyclines and critical reading of the manuscript; and to Dr. M. Neumann for literature and helpful information on European discocyclines. SYSTEMATIC DESCRIPTIONS Order foraminifera Family discocyclinidae Vaughan and Cole 1940 Genus discocyclina Giimbel 1868 Subgenus discocyclina Giimbel 1868 Discocyclina {Discocyclina) assamica sp. nov. Plate 94, figs. 1-6 Holotype. CUGD SD46 (PI. 94, fig. 1). Material. Twenty specimens examined externally, fourteen specimens studied in equatorial section, and eleven in vertical section. No microspheric forms were observed. Diagnosis. Test very thin, flat, with a small umbo surrounded by wide, thin flange. Equa- torial chambers typically narrow and radially much elongated. Radial chamber walls well developed and mostly aligned in adjacent annuli. Low and long lateral chamber cavities are arranged in regular tiers. [Palaeontology, Vol. 6, Part 4, 1963, pp. 658-64, pi. 94-95.] B. K. SAMANTA: DISCOCYCLINA FROM THE UPPER EOCENE 659 Description. External characters. Test of medium size, very thin, waferlike, usually some- what undulating, always with a small well-defined umbo. The whole surface is covered with fine papillae which appear to be more prominent on the umbo than on the sur- rounding wide flange, where papillae of larger dimension occur along the trace of the annular walls, forming low concentric ridges, while the interannular spaces are covered with finer papillae. Extreme thinness of the test does not permit separation from the rock in undamaged state. Pillars are small and subcircular in cross-section. Each is surrounded by five to seven small, polygonal lateral chambers with straight to moderately undulating chamber walls. Internal characters. Equatorial section. The small, bilocular embryonic apparatus consists of round to quadrilateral protoconch and deuteroconch. The larger deutero- conch embraces about two-thirds to nine-tenths of the much smaller protoconch. Rarely, the outer wall of the embryonic apparatus is irregular. Rectangular to spatulate periembryonic equatorial chambers generally form a com- plete annulus surrounding the embryonic chambers. Segment formation is very rare. Periembryonic chambers are distinctly radially elongate. Often the two principal auxiliary chambers are easily distinguishable by being tangentially elongate. Periembryonic equa- torial chambers differ from the later formed equatorial chambers in having greater tangential diameter and more prominent distal arches. There are about twenty-two to thirty chambers in the periembryonic ring. From the second annulus onwards the equatorial chambers are typically narrow and radially considerably elongated. The arrangement of radial chamber walls in adjacent annuli is generally non-alternating and this arrangement continues up to the peripheral region, where alternating chamber partitions appear side by side with non-alternating ones. Frequently the chambers are arranged in radially running colonies. The annuli are irregular both in their courses and in width. The chambers are rectangular in shape. Their radial diameter increases considerably towards the periphery, so that the radial diameter of the peripheral chambers amounts to three or four times that of the chambers near the centre. Annular walls are about 6 to 10 thick while the radial chamber walls are 5 to V jj, thick. Annular stolons are regularly situated on the proximal side of the radial chamber walls. The number of radial stolons cannot be definitely ascertained. There seem to be four radial stolons in the case of alternating chambers, but the non-alternat- ing chambers generally possess two radial stolons, one each proximally and distally. Vertical section. The embryonic apparatus is small and usually flattened along the median plane, so that in some sections the roof and floor of the embryonic apparatus are parallel to each other; the length of the embryonic apparatus is about twice to four times its height. The equatorial chamber layer is very thin. The periembryonic equatorial chambers are higher than the later formed chambers. From the second annulus the height of the chambers decreases slightly for some distance; then their height remains almost constant, but near the periphery increases so that the peripheral equatorial chambers are not only much longer but also a little higher than those near the centre. Roofs and floors of equatorial chambers are 10 to 13 jj, thick. Annular walls are slightly convex outwards. Lateral chambers are arranged in regular tiers. The cavities are long and low between moderately thick roofs and floors. Between the tiers of lateral chambers, there are slightly thickened areas which resemble pillars. 660 PALAEONTOLOGY, VOLUME 6 Measurements External features Specimen no. 1 2 3 4 Diameter of test (mm.) . 9-2 9-0 11-8 8-7+ Diameter of umbo (mm.) 0-7 M 0-9 0-8 Diameter of papillae (ju.) 25 20-45 20-45 25-40 Equatorial sections Specimen no. 1 2 3 4 Diameter of test (mm.) Embryonic chambers: 6-2+ 2-5 + 2-5 + 7-0+ Diameters of initial chamber {p) . 137X94 137x150 137x143 113x 125 Diameters of second chamber {p) . 250x350 275 X 300 312x343 237 X 300 Distance across both chambers (ju) 263 288 325 250 Thickness of outer wall (/a) Periembryonic chambers : 6 6 6 8 Number .... 27 28 23 23 Radial diameter (p.) 44-69 50 44-63 50 Tangential diameter (p.) . Equatorial chambers: 38-50 25 38-50 31-44 Near centre : Radial diameter {p) 75 63 38 50 Tangential diameter ip) Near periphery : 14-19 19 25 19 Radial diameter (/x) 175 162 135 155 Tangential diameter ip) 19 17-21 16-21 16-19 Vertical sections Specimen no. 1 2 3 4 Diameter of test (mm.) 100 8-5 9-4 7-0 + Thickness at centre (mm.) . 0-675 0-6 0-775 0-6 Diameter of umbo (mm.) . 1-0 0-75 1-12 0-7 Thickness at periphery (mm.) Embryonic chambers : 0-15 0-15 0-22 0-15 Length ip) . 300 300 325 300 Height ip) ■ 150 100 140 75 Thickness of outer wall ip) Equatorial chambers : 7 7 9 9 Height at centre ip) 20 19 20 19 Height at periphery ip) . Lateral chambers : 25 25 25 25 Number .... 9 10 12 11 Length ip) . 50-63 62-75 50 63-70 Height ip) . 20 19-20 13-19 13 Thickness of roofs and floors ip) 13-20 13-19 10-13 19 Surface diameter of pillars ip) 13 15-40 15^0 15-40 Occurrence. Holotype and other figured specimens from soft, black mudstone of Kopili Shale forma- tion exposed in Simsong River section near Siju Artheka (25° 20' N., 90° 41' E.); the species also B. K. SAMANTA: DISCOCYCLINA FROM THE UPPER EOCENE 661 occurs in hard, compact, foraminiferal limestone of the same formation in the Nala section near Kamipara (25° 30' N., 90° 12' E.), 2 miles west of Tura. Age. Upper Eocene. Derivation of name. After the province of Assam. Remarks. This extremely thin species with strikingly narrow, radially elongate equa- torial chambers and mostly non-alternating radial chamber walls in adjacent rings is different from any described species of the subgenus Discocyclina. It bears some resem- blance to Discocyclina tencUa Giimbel, but the inadequate description and illustrations of Giimbel’s form does not permit closer comparison with it. The most significant point regarding this new species is that the radial chamber walls in adjacent annuli are mostly aligned. Bronnimann (1945), Vaughan (1945), and Cole (1948) were of the opinion that the radial walls of the equatorial chambers in adjacent annuli are usually aligned in Aktinocyclina, whereas in Discocyclina they generally alter- nate. Consequently, the discovery of a species of Discocyclina s.s. with usually non- alternating chamber partitions is of considerable interest. In fact this Assam form is so closely similar to Aktinocyclina radians (d’Archiac), described and figured by Bronni- mann (1945), that although these two forms can easily be distinguished from each other by external characters, they are identical in the characters of the equatorial chamber layer. If only equatorial sections are available, therefore, it is not possible to separate them. This clearly indicates that in the case of these two closely related subgenera, Discocyclina s.s. and Aktinocyclina, the arrangement of equatorial chambers in adjacent annuli has no supraspecific taxonomic value. Discocyclina {Discocyclina) camcsi sp. nov. Plate 95, figs. Holotype. CUGD SD61 (PI. 95, fig. 1). Material. Twenty specimens examined externally, seventeen specimens studied in equatorial section and thirteen in vertical section. No microspheric forms were observed. Diagnosis. Test of medium size, plano-umbonate with small, round granules uniformly distributed over the surface. Embryonic chambers small and bilocular. In tangential section, each pillar is surrounded by four to six open, polygonal lateral chambers with straight walls. In vertical section, striking regularity of the tiers of open, rectangular lateral chamber cavities between thin roofs and floors. Description. External characters. Test of medium size and usually plano-umbonate to compressed lenticular in shape. Very often the test is saddle-shaped. Numerous, small, slightly raised, round granules are evenly distributed on the surface of the test. In the peripheral region, these granules are generally arranged concentrically. In tangential section, each pillar is surrounded by a rosette of four to six open, poly- gonal, lateral chambers. Vertical walls of lateral chambers are thin and straight. The pillars are separated by a single row of these chambers. X X C 1713 662 PALAEONTOLOGY, VOLUME 6 Internal characters. Equatorial section. The embryonic apparatus is small and bilocular. The circular protoconch is generally attached to the inside wall of the deu- teroconch either tangentially or by a short peduncle. Specimens with deuteroconch embracing the protoconch, except for a small part at the base, are also present. The size of the embryonic apparatus does not vary appreciably, but in shape it shows varia- tion from perfectly circular, triangular to rectangular in equatorial section. Periembryonic equatorial chambers always form a complete annulus in which there are about twenty-three to thirty-eight radially elongate rectangular chambers. Usually they are smaller than the later formed equatorial chambers. The annuli are more or less regular in their courses and gradually increase in width towards the periphery. The radial chamber walls are well developed and in adjacent annuli usually alternate in position, but aligned chamber walls also occur frequently in the peripheral region. Equatorial chambers are rectangular to spatulate in shape. Annu- lar walls are 10 to 13 |L(, thick and the radial walls 6 to 10 ju, thick. Annular stolons are situated regularly on the proximal side of the chambers. Vertical section. The embryonic apparatus is compressed laterally and the length is two to three times the height. Length as well as height of the equatorial chambers increases gradually towards the periphery. Roofs and floors of equatorial chambers are 13 to 15 ju, thick. Annular walls are gently convex outwards. The lateral chambers form definite tiers which are strikingly regular throughout the section. The chambers have open, rectangular cavities and thin roofs and floors. Thickened areas occur between the tiers of lateral chambers and resemble pillars. They appear to be more prominent in the umbonal region. Measurements Diameter of test (mm.) Thickness at centre (mm.) . Diameter of umbo (mm.) . Thickness at periphery (mm.) Diameter of papillae (ju.) External features 1 Specimen no. 2 3 4 11-9 11-5 12-7 9-8 F8 2-0 1-7 1-5 3-7 3-2 3-3 3-1 0-6 0-65 0-5 0-4 75-125 50-100 50-100 60-125 EXPLANATION OF PLATE 94 Figs. 1-6. DiscocycUna (Discocyclina) assamica sp. nov. From Simsong River section near Siju Artheka, Garo Hills. 1, External view of holotype, CUGD SD46, X 5. 2, 5, Equatorial sections of paratypes. 2, X 44; 5, X 22. 6, A part, X 150, of the section illustrated in fig. 2, to show clearly the principle of arrangement of equatorial chambers in adjacent annuli. 3, Vertical section of paratype, x30. 4, Enlargement of umbonal part of fig. 3, X 60. EXPLANATION OF PLATE 95 Figs. 1-4. Discocyclina {Discocyclina) eamesi sp. nov. From Simsong River section near Siju Artheka, Garo Hills. 1, External view of holotype, CUGD SD61, X 4. 2, Equatorial section of paratype, Xl8. 3, Part of an equatorial section, X 36, illustrating the embryonic chambers. 4, Vertical section of paratype, x 20. Palaeontology , Vol. 6 PLATE 94 SAMANTA, Upper Eocene Discocyclina Palaeontology, Vol. 6 PLATE 95 SAMANTA, Upper Eocene Discocyclina B. K. SAMANTA: DISCOCYCLINA FROM THE UPPER EOCENE 663 Equatorial sections Specimen no. 1 2 3 4 Diameter of test (mm.) 4-2 7-0 9-5 6-5 Embryonic chambers; Diameters of initial chamber (fx) . 112x125 100x100 112x 112 100x100 Diameters of second chamber (p) . 300x375 260x300 250x287 212x237 Thickness of outer wall (u) 12 12 8 7 Periembryonic chambers : Number .... 35 31 28 23 Radial diameter ip) 50 50 38 38 Tangential diameter (/x) . 38 25 31 25 Equatorial chambers : Near Centre : Radial diameter (/x) 50 38 38 38 Tangential diameter {p) 25 25 25 31 Near periphery : Radial diameter (p) 81 88 94 75 Tangential diameter (/x) 31 25 25 25 Vertical sections 1 Specimen no. 2 3 4 Diameter of test (mm.) 9-5 + 10-8 10-4 10-8 Thickness at centre (mm.) . 2-0 0-95 1-6 1-4 Diameter of umbo (mm.) . 31 Not 3-5 2-6 Thickness at periphery (mm.) 015 developed 015 015 0-45 Embryonic chambers : Length ip) .... 175 200 287 325 Height (p) . 100 140 106 112 Thickness of outer wall (p) 12 12 8 8 Equatorial chambers : Height at centre (p) 18 19 25 25 Height at periphery (p) . 40 38 31 38 Lateral chambers : Number .... 25 19 25 22 Length (p) . . 100-175 63-163 63-125 75-112 Height (p) . 25-44 19-25 25 13-31 Thickness of roofs and floors ip) 6-12 6-12 6-12 6-12 Surface diameter of pillars (p) . 100-125 100 85-100 75 Occurrence. Holotype and other figured specimens from soft, black mudstone of Kopili Shale formation exposed in Simsong River section near Siju Artheka(25° 20' N.,90° 41' E.); the species also occurs in greyish marly foraminiferal limestone in the Simsong River section, 3 furlongs downstream from the old forest bungalow, in the Rongrenggiri basin. Age. Upper Eocene. Derivation of name. In honour of Dr. F. E. Fames. Remarks. The Assam form with its small bilocular embryonic apparatus can easily be distinguished from the related Indian forms by the characters of its lateral chambers. It resembles closely Discocyclina undulata Nuttall, from which it differs in having rather 664 PALAEONTOLOGY, VOLUME 6 fewer pillars, and the septa of the rosettes are fewer and straight, not markedly sinuous as in Nuttall’s form. It dilfersfrom D. ranikotensis Davies in the broader and less pronounced central boss and also in the much smaller embryonic apparatus. Similar open rectangular chamber cavities between thin roofs and floors, with striking regularity of the tiers of lateral chambers, have been reported also in D.fortisi (d’Archiac), from which the Assam form differs in having quite a different type of embryonic apparatus. REFERENCES BRONNiMANN, p. 1945. Zur Morphologic voD /lA'//V?ocjT//«rt Giimbel 1 868. Eel. geol. Helv. 38, 560-lS, pi. 20. COLE, w. s. 1948. Discocydinidael revised). In J. A. Cushman: Foraminifera, Their Classification and Economic Use, 367-72, Harvard Univ. Press, 4th ed. ELLIS, B. F. and MESSINA, A. R. 1940. Catalogue of Foraminifera. Amer. Mas. nat. Hist. New York. GUMBEL, c. w. 1868. Beitrage zur Foraminiferenfauna der nordalpinen Eociingebilde. Abh. Bayer. Akad. Wiss. 10, 670-730, pi. 1-4. NAGAPPA, Y. 1959. Foraminiferal biostratigraphy of the Cretaceous-Eocene succession in the Indian- Pakistan-Burma region. Micropaleontology, 5, 145-92, pi. 1-11. NEUMANN, M. 1958. Revision des Orbitoides du Cretace et de I’Eocene en Aquitaine occidentale. Mem. Soc. geol. Fr. n.s., 37, fasc. 2-3, 1-174, pi. 1-36. NUTTALL, w. L. F. 1926. The zonal distribution and description of the larger Foraminifera of the middle and lower Kirthar series (Middle Eocene) of parts of Western India. Rec. geol. Surv. India, 59, 115-64, pi. 1-8. VAUGHAN, T. w. 1945. American Paleocene and Eocene larger Eoraminifera. Mem. geol. Soc. Amer. 9, 1-175, pi. 2-46. B. K. SAMANTA Department of Geology, University of Calcutta Manuscript received 13 February 1963 ON GLYPTOGRAPTUS DENTATUS (BRONGNIART) AND SOME ALLIED SPECIES by O. M. B. BULMAN Abstract. The original specimens of BrongniarfsF«co/V/e^f/e/;/ari« can no longer be traced and must be assumed lost. It is therefore proposed that one of the originals of Hall’s Diplograptus pristiniformis be taken as the neo- type, and the species is defined with reference to this and redescribed. A new species believed related to G. dentatus is described from the Arenig rocks of Britain; G. austrodentatus Harris and Keble is redescribed; and four new varieties are recognized from Britain, Scandinavia, and North America. The relations of the G. dentatus and G. austrodentatus groups are discussed, and three other British species are described (two new and one new to Britain) which, though previously identified as G. dentatus, are here assigned to dififerent genera. The earliest true diplograptids are of outstanding importance in graptolite phytogeny, and their sudden appearance has been described as one of the three major events in graptolite evolution. In North America and north-west Europe the majority of such forms have been referred to Glyptograptus dentatus (Brongniart); from Australia the species G. austrodentatus Harris and Keble has been described; and in China both these species have been recorded, together with G. sinodentatus Mu and Lee, and G. eiirvi- thecatus Mu and Lee. The stratigraphical importance of the incoming of the diplograptids, as distinct from the much later ‘Diplograptid Launa’ of Elies (1922, p. 194), was recognized by Harris and Thomas (1938, p. 66) and Bulman (1958, p. 162), but their use in detailed correlation is handicapped by a lack of exact knowledge of the material and of its stratigraphic range. With the co-operation of several friends and colleagues, I have assembled representa- tives of as much of this early material as possible, and it is a great pleasure to acknowledge the ready and generous help I have received. Dr. D. J. McLaren (Canadian Geological Survey) arranged for the loan of the types of Hall’s D. pristiniformis and placed at my disposal collections from the Levis Shales, and Dr. L. M. Cumming specially collected and sent material from Begin’s Hill, Levis; Dr. D. E. Thomas (Department of Mines, Melbourne) sent a large collection of G. austrodentatus, and other specimens were col- lected and sent me by Dr. F. C. Beavis (University of Melbourne); Dr. D. E. Jackson sent specimens from the Yukon, and allowed me to draw upon his collection from the Skiddaw Slates; some sectionable specimens of G. austrodentatus were included in a collection from Marathon, Texas, sent some years ago by Professor C. O. Dunbar and Dr. W. N. Berry; Dr. Isles Strachan (Birmingham University), Mr. J. D. D. Smith (Geological Survey and Museum), Dr. H. W. Ball (Brit. Mus. (Nat. Hist.)), Mr. A. G. Brighton (Sedgwick Museum), and Professor T. S. Westoll (University of Newcastle) have assisted with the loan of material in their charge. Dr. D. Skevington has given me access to his observations on diplograptids from the Orthoceras Limestone; and Pro- fessor Pruvost and Professor Boureau kindly undertook the laborious, but unavailing, search for the types of Fucoides dentatus Brongniart. Finally, my assistant Mrs. C. M. Clarkson has rendered invaluable help in the sectioning of material, construction of wax models, photography, and the preparation of illustrations. [Palaeontology, Vol. 6, Part 4, 1963, pp. 665-89, pi. 96-97.] 666 PALAEONTOLOGY, VOLUME 6 Upon examination it becomes apparent that, though often badly preserved and some- times distorted, a number of recognizably different species and varieties is present, and these appear to fall into two groups. These two groups are typified by G. dentatus and G. austrodentatus. They are also, it appears, represented by the two forms described (Bulman 1936) as G. dentatiis-teretiusculus transient and G. dentatus respectively, from the Holm Collection in Stockholm. This Scandinavian material, augmented by recent preparations of Dr. D. Skevington, provides important information concerning the mode of development of the rhabdosome and the three-dimensional form of the thecae, which can be used in the interpretation of less well-preserved material from elsewhere. The G. dentatus group centres around forms with a slender, somewhat pointed, proxi- mal end; thU has a dominantly upward direction of growth and the apertures of the first four thecae tend to be markedly alternating. The metathecal portion of the thecae has a glyptograptid curvature, with only an incipent geniculation situated low down on the metatheca, and the apertural rim may show a pair of very weakly developed lappets. G. austrodentatus contrasts with this in having a bluntly terminated rhabdosome, with thU and thU almost symmetrically disposed. The metathecal portion of the thecae has a more strongly marked geniculation situated midway along the metathecal length or even more distally, and the difference in metathecal form is very clearly shown by Skevington’s preparations (cf text-figs. Ic, 8^/). The apertural margin in G. austrodenta- tus may possess a vertical or even somewhat introverted ventral lip. The possible relation of the variants and the interpretation of seeming intermediates is discussed below; and three British species are described which, though identified as G. dentatus in the past, are here assigned to different genera. MODE OF DEVELOPMENT OF EARLY DIPLOGRAPTIDS {a) G. dentatus group Exact knowledge of the mode of development of the rhabdosome in this group derives from serial (microtome) sections of a specimen in the Holm Collection already described under the name G. dentatus-teretiusculus transient (Bulman 1936, pp. 57-61), supplemented by an excellent transparent proximal end of another specimen from Oland prepared and figured by Dr. D. Skevington. It is unnecessary to repeat the description already published, but it was pointed out to me by Mrs. Clarkson that in constructing the original wax model I unwittingly exag- gerated the vertical scale by about one-sixth, and she has prepared a corrected drawing depicting this model in true proportion (text-fig. \a). The second theca, thl'*^, originates comparatively low down on thU and grows mainly across the sicula and upwards on the opposite side. Th2^ has a slight initial downward growth before turning upwards. Dr. Skevington has kindly allowed me to reproduce (text-fig. Ic) his figure of his recent preparation (now in the Palaeontological Institute at Uppsala), the growth-lines of which conclusively demonstrate that (as hinted in Bulman 1936, p. 59) the dicalycal theca is th2“. Nevertheless, it seems probable that members of the dentatus group from a lower stratigraphical level would have th2^ dicalycal, and it would be impracticable (as regards shale material) to attempt any specific subdivision on this basis. Details are obscure in shale material, but development of this general type would account for and accord O. M. B. BULMAN: GLYPTOG RAPTUS DENTATUS (BRONGNIART) 667 TEXT-FIG. 1. Glyptograptus dentatus (Brongniart). a. Corrected drawing of wax model in the form of an internal cast of the thecal cavities of the proximal end made from 10|U. serial sections through specimen 419 (Holm Collection, Riksmuseum, Stockholm). Reverse (above) and obverse (below) aspect; only a small portion of the sicular cavity is shown where it communicates with thH; x40 approx, b. Thecal diagram, c. Specimen prepared from Orthoceras Limestone of Hallunden, Oland, by Dr. D. Skevington. 5l. 1232, Palaeontological Institute, Uppsala; x35 approx. 668 PALAEONTOLOGY, VOLUME 6 with the outline visible in shale material belonging to G. dentatus, G. shelvensis, and a considerable number of later Ordovician diplograptids. (b) G. austrodentatus group Development in this group is known from serial (microtome) sections and some early growth-stages of a form unfortunately described as G. dentatus from Oland (Bulman 1936, pp. 49-57), and incomplete or damaged serial sections (ground at ^ mm. intervals) of two limestone specimens from Texas (text-fig. 3). Here again the original wax model was unintentionally distorted and is redrawn by Mrs. Clarkson along with the new Texas models in text-fig. 2. ThH arises relatively high up on thH and grows initially upward to form a prominent hoodlike structure before turning downwards and ultimately outwards and upwards; and it therefore comes to trace a distinctive on-shaped course of growth. It gives rise to th2^ at an early stage, and this theca also has a conspicuous downward element in its early portion and in fact presents a somewhat flattened replica of the same sinuous growth as thH. Th2^ has also initially a downward direction of growth. ThHand thPare almost symmetrically disposed around the sicula, and there is even some degree of sym- metry in the disposition of th2^ and th2^. The proximal end (when preserved in relief) commonly shows an inverted V-shaped swelling at the base of the median septum sepa- rating th2^ and th2^ (text-fig. 8a, b). There is some variation in detail, but all three models agree in the main essentials — the on-shaped configuration of thl- with its conspicuous downwardly directed middle portion, and the initially downward growth of th2^ and even th2^. Such development is known to occur in G. austrodentatus oelandicus and americanus and would account for and accord with what is at present known of G. austrodentatus austrodentatus and angli- cus. To judge from the configuration of the proximal end shown in Mu and Lee’s figures (1958, text-fig. 9e-g) the same type of development characterizes G. sinodentatus and (1958, pi. 2, fig. 17) G. curvithecatus (as well as other forms mentioned below). (c) Developmental relationships The recognition of these two types of proximal end side by side in what is probably the I. gibberulus Subzone of the D. extensus Zone (Shelve Church Beds, see p. 678) and their occurrence in a number of varietal forms at widely scattered localities in Europe, North America, Australia, and Asia, raises interesting possibilities concerning their relationship. The emergence of the biserial glyptograptid rhabdosome is a most important evolu- tionary event, but these geographically distributed variants are not necessarily strictly contemporaneous nor confined to a narrow horizon; and the exact horizon is probably more reliably assessed by general faunal characteristics than by the mere occurrence of one or other of the dentatus or austrodentatus group of glyptograptids. Thus it seems probable that G. austrodentatus anglicus and G. shelvensis originate in the /. gibberulus Subzone (though both may also occur in the D. hirundo Zone); G. austrodentatus oelandi- cus is of upper D. hirundo Zone age; and G. austrodentatus austrodentatus probably lies somewhere between. Secondly, it is (at present at least) impossible to deny that the dentatus and austro- TEXT-FIG. 2. Glyptograptiis austrodentatiis Harris and Keble. a-d. G. aiistrodeiitatiis oelandicits var. nov. a. Thecal diagram ; b, c, d. Reverse, obverse, and lateral view of corrected wax-model in the form of an internal cast of the thecal cavities of the proximal end, made from 10 p, sections through specimen 1638 (Holm Collection, Riksmuseum, Stockholm); x40 approx, e-i. G. aiistrodentatus americanus var. nov. e. Thecal diagram ; /, g. Reverse and obverse views of wax model constructed from serial sections ground at jg mm. intervals through specimen SM A54407 (Marathon, Texas); h, i. Reverse and lateral views of wax model constructed from serial sections ground at jg intervals through specimen SM A 53803 (Marathon, Texas) ; x 40 approx. Shaded areas indicate regions where crumbling of the encasing plaster resulted in complete or partial loss of sections. See also fig. 3. 670 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 3. Glyptograptus aiistrodeutatiis americaniis var. nov. Selected sections from the series through specimen SM A53803 (Marathon, Texas), represented in the upper part of text-fig. 2 h, i. Con- formation of the early parts of thP and th2^ is clearly shown and compares closely with microtome sections of G. austrodentatus oekimlicus shown in Bulman 1936, text-fig. 21. X 32 approx. dentatus groups may have evolved independently; and obviously independent scandent forms (including the cryptograptids, the scandent dichograptids, and probably also Trigonograptus) did indeed evolve independently at approximately the same period. Nevertheless, the dipleural rhabdosome form and generally similar thecal characters would seem to point to a monophyletic origin. This is supported also by the occurrence of forms apparently intermediate between shehensis and augUcus in the Shelve Church O. M. B. BULMAN; GLYPTOG RAPTUS DENTATUS (BRONGNIART) 671 Beds. Some doubt is now thrown, however, upon my suggestion (Bulman 1936, p. 52) that the broader and narrower forms of what is here called auslrodentatiis oelandicus represent transients to deutotus, for it would now appear that the separation of the true dentatus type had already occurred. Having regard to the thecal changes as well as proximal-end changes involved in the separation of the dentatus and austrodentatus groups, it is unlikely that the process was operative over a long period of time. Until more accurate and detailed morphological (and stratigraphical) evidence is available, it it proposed as a working hypothesis to regard the true diplograptids as a monophyletic group, rapidly undergoing a primary division into two groups {dentatus and austro- dentatus) each giving rise to a number of descendants in the D. hinmdo and D. bifidus zones. It is becoming apparent that the sequence of proximal-end development in diplograp- tids is more complicated than was originally supposed, and it is doubtful whether much is to be gained by any attempt to classify it into ‘stages’. What has until now been called the "dentatus stage’, with its atavistic features (in the strong downward component of thU and th2\ &c.) was originally conceived as having only a limited occurrence in ‘the archaic diplograptid ’ of the lower Ordovician. Subsequent work on well-preserved material has added steadily to the number of species known to possess features of this development to a greater or lesser degree, and has extended its upward range certainly as high as the G. teretiusculus Zone. It is now suggested that the development of this whole group might be referred to under the more general term of the streptoblastic type of diplograptid development, in contrast to the prosoblastic type exemplified by the true dentatus and its descendants. This streptoblastic type of development is best shown in the austrodentatus group of glyptograptids, but is well displayed also in Lasiograptus bystrix (Bulman 1936) and (with dicalycal th2^) in Gynmograptus retioloides (Urbanek 1959). It appears also to be a characteristic of those climacograptids for which Pfibyl (1947) proposed the name Pseudoclimacograptus, and combined with the septal features stressed by Pfibyl and the thecal characters emphasized by Jaanusson (1960) may be held to justify the erection of this genus. P. scharenbergi, the genotype, has been shown to possess a modified version of this proximal end, and so may some other species at present included (Jaanusson 1960) but imperfectly known. It is, however, very well displayed in such geologically early species as P. forrnosus Mu and Lee, and P. cumbrensis sp. nov. The modifications seen in P. scharenbergi and the relationship postulated by Urbanek between G. retio- loides and G. linnarssoni suggests that the prosoblastic type of proximal end does from time to time independently develop, and also that the dicalycal theca may be deferred to th2^ even within the strictly streptoblastic type. We arrive therefore at the supposition that the streptoblastic type of development characterizes perhaps the greater part of the early diplograptids, conservative as to their development but variable (in Arenig times) in thecal characters. From it is quickly derived the prosoblastic G. dentatus group, which appears to be ancestral to most of the later Ordovician glyptograptids, orthograptids, and climacograptids. From it also is derived, by thecal differentiation, the Pseudoelimacograptus group, and probably hallo- graptids and lasiograptids. Modification of the mode of development approaching or leading to a prosoblastic type may occur within any of these groups, but a fuller discus- sion of these relationships is outside the scope of this article. 672 PALAEONTOLOGY, VOLUME 6 THE HOLOTYPE OF GLYPTOGRAPTU S DENTATUS The species now known as Glyptograptus dentatus was described (under the name Fucoides dentatus) by Brongniart in 1828 from ‘le calcaire de transition, Pointe Levi pres Quebec dans le Canada (Coll, de M. Stockes)’ and illustrated by four lithographed figures (pi. vi, figs. 9-12). In none of these is the proximal end represented, but the glyptograptid character of the thecae is clearly shown in the two enlarged figures, 1 1 and 12. The diagnosis reads: ‘F. fronde membranacea, lineari (an simplici?), pinnatifido- dentata, enervi, dentibus triangularibus subacutis, apice obtuso’; and the relevant part of his description is contained in the first paragraph (p. 70): ‘Nous ne connaissons que des fragmens de cette petite espece de Fucoide; ce sont des portions, longues de 2 centi- metres environ, d’une fronde lineaire, d’une largeur tres-uniforme et egale a peu pres a 2 ou 3 millimetres. Ces frondes, qui paraissent minces et membraneuses, sont pro- fondement dentees sur les bords et presque pinnatifides; les lobes ou dentelures sont triangulaires, tres-reguliers, d’une forme aigue, mais a pointe mousse.’ Apart from mention by Sternberg (1838) and Bronn (1849) I can find no further cita- tion in the literature till 1865, when James Hall (‘Graptolites of the Quebec Group') added a footnote to the description of his species Diplograptus pristinifonnis (p. 110) suggesting that it was identical with Fucoides dentatus; and on p. 84 (footnotes) he writes of both pristinifonnis and bryonoides: ‘It is only since [Brongniart’s] descriptions have been in print, and published references made to them, that I have discovered this identity, or I would have proposed to substitute the specific names of Brongniart [dentatus and serra respectively] for those given by me in 1857. 1 take the first opportunity of making the correction.’ Hopkinson and Lapworth (1875) use the name dentatus, under which the species has always been described and recorded since. A large number of Brongniart’s types are preserved in the Museum national d’His- toire naturelle in Paris; but a search for the dentatus originals has been unavailing, and they must presumably be regarded as lost. I am greatly indebted to Professor Pierre Pruvost and Professor Ed. Boureau for the trouble they have taken in trying to locate them. Since Brongniart’s originals can no longer be traced, it would seem natural to propose the selection of a neotype from among the material figured by James Hall. The speci- mens are probably from the same locality, and the identity has been accepted for nearly a century. I am most grateful to Dr. Digby McLaren, of the Canadian Geological Survey, for the loan of the specimens concerned. Proposed neotype. Geol. Survey Canada 943 (Hall 1865, pi. 13, figs. 16, 17). Plate 96, fig. 1 ; text-fig. Aa, Description. Rhabdosome length 2 cm., with 1 cm. of virgula preserved. Width M mm. at the level of thH, widening to 1-2 mm. at th3^ and 1-5 mm. at 5 mm. length, reaching a maximum of 2-1 mm. Seven thecae occur in the first 5 mm., 10-1 1 in 10 mm. distally. Thecal overlap is not clearly shown, but seems to be slightly less than half at about 5 mm. from the proximal end and to increase distally. The shape of the ventral wall of the thecae (compressed) is smoothly but not uniformly curved, showing a stronger curvature just above the aperture of the preceding theca and becoming straighter distally; i.e. there is an incipient geniculum situated low down on the metatheca, and the extreme proximal part of the metatheca is slightly impressed. The aperture is denticulate, the apertural O. M. B. BULMAN: G LY PTOG RAPTU S DENTATUS (BRONGNIART) 673 margin slightly inclined proximally inwards. The median septum is not visible; a stout virgula is, however, clearly impressed through the lateral periderm throughout the length of the rhabdosome. The sicula is not visible, but it possesses a stout virgella 2-5 mm. long which is apparently coated with some secondary cortical tissue. Specimen 943a (Hall’s fig. 15) has in my opinion been incorrectly drawn on Hall’s plate and comprises not one but two separate but aligned rhabdosomes. The lower one (Plate 96, fig. 2) is complete proximally and attains a length of 3 cm. with a maximum breadth of 1-8-1 -9 mm. The upper one is incomplete at the proximal end and is inter- rupted in the middle; the total length is 1-5 cm., with a maximum breadth of 1-9 mm. Thecae number about ten in 10 mm. distally, but preservation is too poor to admit the observation of much detail; they appear to be similar in all essentials to those of the proposed neotype. Horizon. Subsequent work on the Levis Shales, particularly that of Raymond (1914), strongly suggests that the specimens must have come from the upper part of the Levis Shales, and probably from Raymond’s Zone Dg; but nothing objective is known of the horizon and locality beyond Hall’s statement ‘Quebec Group; Point Levis’ (and the same applies to Brongniart’s specimens). No special significance attaches to the associa- tion which for the neotype comprises dichograptid stipes including a specimen of Tetragraptus quadribrachiatus (s.L), and (on 943a) a figured dichograptid stipe labelled Graptolithus denticulatus. SYSTEMATIC DESCRIPTIONS Ghptograplus deulalus (Brongniart) Plate 96, figs. 1-5; text-figs. 1, 4 1828 Fucoides den tat us Brongniart, p. 70, pi. vi, figs. 9-12. 1838 Sphaerococcites dentaius; Sternberg, p. 29. 1848 Fucoides dentatus\ Bronn, p. 505. 1848 Sphaerococcites dentatus; Bronn, p. 1162; 1849, p. 8. 1858 Graptolithus pristiniformis J. Hall, p. 133. 1865 Diplograptiis pristiniformis', J. Hall, pp. 84 (footnote), 110, pi. 13, figs. 15-17. 1868 Diplograptiis pristiniformis', Nicholson, p. 140, pi. 5, figs. 14, 15. 1875 Diplograptus dentatus', Hopkinson and Lapworth, p. 656, pi. xxxiv, figs. 5a-h, ? 5/, k. 1904 Diplograptus dentatus', Ruedemann, p. 719, pi. 17, figs. 10-13. 1907 Diplograptus (Glyptograptus) dentatus', Elies and Wood, p. 253, pi. xxxi, fig. 4a-d. 1907 Diplograptus (Glyptograptus) dentatus appendiculatus', Elies and Wood, p. 255, pi. xxxi, fig. 5. 1931 Glyptograptus dentatus', Bulman, p. 55, pi. 6, figs. 7-11, pi. 7, figs. 1, 2. 1936 Glyptograptus dentatusiteretiiisculus transient; Bulman, p. 57, pi. 3, figs. 1-4, 8-11 ; pi. 4, figs. 1-3. 1947 Glyptograptus dentatus', Ruedemann, p. 404, pi. 69, figs. 1-8. Revised diagnosis. Rhabdosome 2-3 cm. in length, widening gradually from a pointed proximal end furnished with conspicuous virgella. Width at thH about 1 mm., increasing to 1-7 or 1 -8 mm. at 10 mm., and attaining a maximum of 2-0-2-2 mm. Thecae 7-9 in the first 5 mm., 10-13 in 10 mm. distally, with glyptograptid curvature of the ventral wall of the metathecae, slightly impressed just above the aperture of the preceding theca; apertural margin slightly concave in profile view with very shghtly developed lappets. 674 PALAEONTOLOGY, VOLUME 6 TEXT-FIG. 4. Glyptograptus dentatus (Brongniart). All figures X 5 approx, a. Proposed neotype, GSC 943, Point Levis, Quebec; figured by Hall 1865, pi. 13, figs. 16, 17. b, GSM 23073, D. bifidus Zone, Clynderwen Station, S. Wales; figured EUes and Wood, pi. 31, fig. 4e and text-fig. 1746. c, BMHN, Q1173, D. bifidus Zone, Ellergill, Milburn, Westmorland, d, Riksmuseum, Stockholm, 419 (Holm Collection), Orthoceras Limestone, Halludden, Oland; figured Bulman 1936, text-fig. 226. O. M. B. BULMAN: GLYPTOGRAPTUS DENTATUS (BRONGNIART) 675 Development prosoblastic, with dominantly upward growth of thD, the aperture of which is always at a higher level than that of thD. Median septum complete, practically straight. Proposed neotype. Original of Diplograptus pristiniformis Hall, 1865, pi. 13, figs. 16, 17: CGS, no. 943 : Levis Shales, Levis, Quebec. Here refigured as text-fig. 4a and Plate 96, fig. 1. Description. The chief characteristics of this species are the gradual widening of the rhabdosome from a pointed proximal end, and the glyptograptid form of the thecae, the curvature being greatest near the base of the metatheca which becomes progressively straighter towards the aperture. The form of the proximal end results from a mode of development originally described from serial sections of limestone material from Oland (see text-fig. la). The aperture of thP lies well above that of thD, and at the proximal end the two series of thecal aper- tures are markedly alternate (a feature which may persist throughout the rhabdosome) and the whole proximal end tends to present a rather ‘ drawn-out ’ appearance in contrast to the truncated proximal end of the austrodentatus group of species. According to Skevington (MS.) the sicula has a length of only 1-3 mm. with an apertural width of 0-32 mm. : it is provided with a prominent virgella which may be 4 mm. or even more in length and sometimes shows secondary thickening. A subapertural spine is carried by thD and a similar slender spine is usually visible on thP. The form of the thecae is well seen in three-dimensional specimens from the Holm Collection (Bulman 1936) and in Skevington’s figures (Plate 96, fig. 3, and text-fig. 1). The glyptograptid curvature is strongest low down on the wall of the metatheca, and in its distal (apertural) part the wall becomes nearly straight : on compression this may give the appearance of an incipient geniculation. There is a pair of rudimentary lappets on the apertural margin, produced (Skevington MS.) by uneven development of the final fusellus, and the apertural margin is concave in profile view, and straight (or very slightly concave) in true ventral view — a feature which may cause some variation in the appear- ance of the thecae on compression. The amount of overlap cannot be determined with any accuracy: in the proximal part of the rhabdosome it is rather less than one-half (Bulman 1936, fig. 32). According to Skevington’s observations, the interthecal septum is apt to be incompletely developed even when it is indicated externally on the lateral walls of the rhabdosome. The median septum is here straight and originates between the bases of th3^ and th32. There is generally a well-developed virgula extending a centimetre or so beyond the distal end of the rhabdosome. Rarely this is expanded into an elongate vesicle, and such specimens were described by Elies and Wood (1907) under the name G. dentatus appendi- culatus. It is doubtful whether this name is worth maintaining, particularly as a closely similar modification may occur in G. shelvensis. Synrhabdosomes have been figured by Bulman (1931) and Ruedemann (1947). Occurrence. Widely distributed in the D. bifidus Zone — e.g. Canada — Zone D (? con- fined to Dg), Levis Shales, Quebec. Britain — D. bifidus Zone (South Wales, Shropshire, Lake District) ; 1 D. hirundo Zone; ? D.murchisoni Zone. Scandinavia — Glaukonithaltig gra Vaginatumkalk, Oland {D. hi r undo j bifidus Zone). 676 PALAEONTOLOGY, VOLUME 6 Glyptograptus shelvensis sp. nov. Plates 97, figs. 1-3, 14; text-fig. 5 Diagnosis. Rhabdosome small, about 1-| cm. long, widening gradually from less than 1 mm. (usually 0-8-0-9 mm.) at thP to a maximum of 1-3-14 mm. (rarely 1-5 mm.). Thecae number about six in 5 mm. (maximum range 11-14 in 10 mm.), in more distal part of the rhabdosome, of the type of G. denlalus. Proximal end with virgella about 1 mm. in length, and delicate subapertural spines on thl^ and thP; orientation of thl^ and thT^ and presumably mode of development as in G. dentatus. Holotype. Sedg. Mus. Camb., A40788a, b (text-fiig. 5fl); Shelve Church Beds (upper part of D. extensiis Zone), Shelve, Salop. Description. The species resembles G. dentatus in the characters of the thecae and the general appearance of the proximal end, but dilfers in its smaller size and more parallel- sided rhabdosome. The sicula is virtually unknown though one specimen (from Shelve) shows some indication of a sicula about 1-4 mm. in length. All the material is compressed and often slightly sheared, so that even thecal characters are not easy to determine ; in the better preserved of the specimens, however, the thecae have an almost even glyptograptid curvature, with a slight indication of an incipient geniculation low on the metatheca as in G. dentatus. In place, however, of the incipient lappets of G. dentatus there seems to be a broad, low projection or ‘lip’ on the ventral portion of the edge of the aperture; in some preservations this may give an appearance of slight apertural ‘isolation’, but it is presumably due to the deposition of incomplete or uneven fuselli at the apertural margin. The metathecal portion of thP has a dominantly upward direction of growth and its aperture characteristically lies well above that of thP. Partly because of poor preserva- tion and partly as a consequence of this orientation of thP, it is difficult to determine reliably the width of the rhabdosome at the level of the thl^ aperture, but it appears to be always less than 1 mm. and may be as little as 0-75 mm. The mode of development is EXPLANATION OF PLATE 96 Magnification X 5 approx., except where otherwise stated. Figs. 1-5. Glyptograptus dentatus (Brongniart). 1, Proposed neotype, GSC 943 (figd. Hall 1865, pi. 13, figs. 16, 17). Quebec Group (probably Da of Raymond); Point Levis, Quebec. 2, GSC 943a (figd. Hall 1865, pi. 13, fig. 15), x3 approx.; same locality. 3, Qbverse (a) and reverse (h) view of rhab- dosome in full relief, X 10 approx. Riksmuseum, Stockholm, Holm Collection 1280 (figd. Bulman 1936, pi. 3, figs. 1, 2). Vaginatumkalk (D. hirundolbifidus Zone); Halludden, Oland. 4, GSM 23073 (figd. Elies and Wood 1907, pi. xxxi, fig. 4c). D. bifidus Zone; Rhyd-y-behan Cottage, near Clynder- wen, S. Wales. 5, GSM 23072 (figd. Elies and Wood 1907, pi. xxxi, fig. 4b). D. bifidus Zone; Gilfach- wen near Clynderwen, S. Wales. Fig. 6. Climacograptus ci. bifonnis M\x diwAGee. GSM JR4141. Skiddaw Slates, ? D. iurundo Zone; NE. of Kirkland Church, Cumberland. Figs. 7-8. Pseudoclimacograptus cumbrensis sp. nov. Skiddaw Slates, ? D. hirundo Zone; Bassen- thwaite Sandbeds, Cumberland. 7, Holotype, SM A53044 (latex mould). 8, SM A18134 (latex mould) (distal part of rhabdosome). Fig. 9. Diplograptus ellesisp. nov. Holotype, BMNH, Q1 174. D. Zone; Ellergill, near Milburn, Westmorland. Palaeontology, Vol. 6 PLATE 96 BULMAN, Glyptograptiis dentatus and allied species ■■rf- ■ii O. M. B. BULMAN; GLYPTOGRAPTUS DENTATUS (BRONGNIART) 677 assumed to be generally similar to that of G. dentatiis, but there is no evidence here that the dicalycal theca is other than th2^. The median septum, when recognizable, seems to be straight and complete at least from th2^. Some specimens from the D. hinmdo Zone (Skiddaw Slates) possess an expanded virgula as in ‘G. dentatus appendicidatus\ f TEXT-FIG. 5. Glyptograptus shelvensis sp. nov. All figures X 5 approx, a-e. Shelve Church Beds (?/. gibberiilus Subzone), Shelve Church, Shropshire, a, SM A40788b, holotype. b, SM A40798. c, SM A40791b. d, SM A45891b. e, SM A40796, abnormal specimen in respect of thP and thl". /, Skiddaw Slates (?i). hinmdo Zone), Outerside, Cumberland; BMNH, Q1178. Remarks. There appears to be some variation, particularly in the proximal end, which cannot be entirely attributed to preservation, but which is difficult to evaluate. One specimen (A40796, text-fig. 5e) has the general characters of shelvensis, but the initial thecae (P and P) are more or less symmetrical, and the thecae remain paired rather than alternate throughout the rhabdosome (as they are also in A40791, text- fig. 5c). Such variations suggest the occurrence of structural intermediates between shelvensis and anstrodentatus anglicus. In general, however, the species is readily distinguishable from anstrodentatus anglicus by its narrower rhabdosome, more pointed proximal end, type of theca, and smaller thecal number. In the collections from Shelve at the Sedgwick Museum, shelvensis outnumbers the others by about three to two. Yy C 1713 678 PALAEONTOLOGY, VOLUME 6 The species is probably ancestral to G. dentatus, and some specimens from the D. hinmdo Zone of the Skiddaw Slates appear to be transitional; but the relationship is difficult to substantiate. There is some resemblance to Hsii’s G. dentatus intermedius, but G. sfielvensis lacks the peculiar basal spines of that species. Occurrences. This is one of the characteristic fossils of the Shelve Church Beds, Shelve Church, Salop. According to Professor Whittard (in litt., 1962) the Shelve Church Beds occur some 500 feet below the base of the Hope Shales (D. bifidus Zone) ; the D. hirundo Zone, represented by the Tankerville Flags, is some 200 feet in thickness, and it is reasonable stratigraphically to assign the Shelve Church Beds to the upper part of the D. extcnsus Zone, probably equivalent to the I. gibberulus Subzone of the Skiddaw Slates (see Jackson 1962, p. 305). The species also occurs in the Skiddaw Slates in the I. gib- berulus Subzone and in the overlying D. hirundo Zone ; and in the upper part of the G. dentatus Zone (T. ensiformis Subzone) of the Arenig of Yukon (Jackson and Lenz 1962). Glyptograptus austrodentatus Harris and Keble Remarks. G. austrodentatus has proved to be a species of widespread occurrence repre- sented in four continents by a number of slightly differing forms. Despite the poor preservation of much of the material, and the occurrence of some variation in any one EXPLANATION OF PLATE 97 Magnification X 5 approx., unless otherwise stated. Fig. 1. Glyptograptus cf. sbelveusis sp. nov. SM A54388 Arenigian, Zone of G. dentatus (T. ensiformis Subzone); Road River, 2530, Yukon. Figs. 2, 3, 14. Glyptograptus shelvensis sp. nov. 2, BMNH, Q1178. Skiddaw Slates (? D. hirundo Zone) ; Outerside, Cumberland. 3, SM A40784. Shelve Church Beds, D. extensus Zone (II. gibberulus Subzone); Shelve Church, Shropshire. 14, BMNH, Q1176, Skiddaw Slates {ID. hirundo Zone); Outerside, Cumberland. Fig. 4. Glyptograptus austrodentatus anglicus var. nov. SM A40786. Shelve Church Beds, D. extensus Zone ( ? /. gibberulus Subzone); Shelve Church, Shropshire. Fig. 5. G. austrodentatus nuitabilis war. nov. Holotype; SM A45906b. Shelve Church Beds, D. extensus Zone (? /. gibberulus Subzone); Shelve Church, Shropshire. Figs. 6-9, 12, 15. G. austrodentatus austrodentatus Harris and Keble. 6, SM A51691. 7, SM A51692. 8, SM A51693. 9, SM A51689. All from Darriwil, D1 (zone of G. austrodentatus)', loc. A310, parish of Sedgwick, Victoria, Australia. 12, SM A51698, X 10. Darriwil, D1 (zone of G. austrodentatus)', loc. 2076, parish of Wellsford, Victoria, Australia. 15, Specimen sheared, but in high relief, showing form of thecae and interthecal septa. SM A51688 (latex mould), X 10. Darriwil, D1 (zone of G. austrodentatus)', loc. A3 10, parish of Sedgwick, Victoria, Australia. Figs. 10-11, 13. G. austrodentatus cf. americanus var. nov. 10, SM A54398. Arenigian, Zone of G. dentatus (T. ensiformis Subzone); Road River, 2540, Yukon. 11, SM A54405. Arenigian, Zone of G. dentatus (I. caduceus Subzone); Road River, 2565, Yukon. 13, SM A54406. Arenigian, Zone of G. dentatus (/. caduceus Subzone); Road River, 2580, Yukon. Figs. 16-17. G. austrodentatus oelandicus var. nov.; Vaginatumkalk (D. hirundo! bifidus Zone); Hallud- den, Oland. 16, Reverse (a) and obverse (b) view of rhabdosome in full relief. Riksmuseum, Stock- holm, Holm Coll. 307 (figd. Bulman 1936, pi. 3, figs. 5, 6). 17, Reverse view in full relief, X 10. Holm Coll. 307 (figd. Bulman 1936, pi. 3, fig. 7). Figs. 18-19. G. austrodentatus americanus var. nov. 18, SM A53803. Fort Pena formation. Zone 9 (Hallograptus etheridgei)', 1 miles SW of Marathon, Texas. 19, GSC 8074, Shumardia limestone (Di); Point Levis, Quebec. Palaeontology, Vol. 6 PLATE 97 BULMAN. Glyptograptus O. M. B. BULMAN: GLYPTOGRAPTUS DENTATUS (BRONGNIART) 679 area, these local forms seem to express genuine differences, to ignore which would be to extend unduly the range of variation in the species. There is evidence suggesting that these are not strictly contemporaneous, but it is not at present possible to recognize with certainty any distinctive chronological element in these variants, and their taxo- nomic status may be essentially that of geographical races. For convenience in ter- minology they are referred to as varieties. TEXT-FIG. 6. Glyptograptiis aiistrodentatus austrodentatus Harris and Keble. a, SM A51689. b, SM A51691. c, SM A51692. d, SM A51687, mould of proximal end in relief, e, SM A51688, latex cast of natural mould of proximal end in relief. All specimens from G. austrodentatus Zone (Darriwill, Dl), loc. A310, Parish of Sedgwick, Victoria, Australia. X 5 approx. Glyptograptiis austrodentatus austrodentatus Harris and Keble Plate 97, figs. 6-9, 12, 15; text-fig. 6 1932 Diptograptus (Glyptograptiis) austrodentatus Harris and Keble, p. 39, pi. v, figs. 4, 5, text-figs. 1-4. 1935 Diptograptus (Glyptograptiis) austrodentatus^ Harris and Thomas, p. 295, fig. 3, nos. 1-5. 1938 Glyptograptiis austrodentatus', Harris and Thomas, pi. 2, fig. 51. 1960 Glyptograptiis austrodentatus', Thomas, pi. vi, fig. 85. Revised diagnosis. Rhabdosome small, rarely exceeding 1 cm. in length, widening rapidly from an initial width of about 1-3 mm. at thH to a maximum of 1-6-1 -9 mm. (mainly 1-7-1 -8 mm.) which is uniform over most of the rhabdosome. Thecae 6-6| in 5 mm. distally, initially with a fairly pronounced geniculation one-third to halfway along the metatheca, becoming less prominent distally. Overlap nearly two-thirds. The first two thecae (H and H) are small, slender, almost symmetrically disposed with slender sub- apertural spines. Median septum complete, slightly undulating. Sicula 2-2-3 mm. long with slender virgella. Lectotype. With the approval of Dr. D. E. Thomas I designate the original of Harris and Keble 1932, text-fig. 3 (specimen no. 31365, Geol. Surv. Viet. — cited in error as 3165). Description. This is a well-characterized subspecies and amongst some thirty specimens sent me from Victoria there is very little variation that cannot be attributed to preservation. 680 PALAEONTOLOGY, VOLUME 6 Rhabdosomes rarely exceed 1 cm. in length and are almost parallel-sided throughout; the width at the first thecal pair is about 1-3 mm., and what is practically the maximum width is attained by the second thecal pair. In a few specimens the maximum width is as little as 145 mm. (? distortion) or as much as 1-9 mm., but the range is generally between 1-6 and 1-8 mm. The form of the thecae is well shown in two specimens preserved as natural moulds in high relief. In profile view the wall of the metatheca is initially concave, becoming strongly convex about halfway along its length, and is very slightly introverted at the apertural end ; this introversion appears to affect principally a vertical ‘ lip ’ of incomplete fuselli (see description of G. shehensis). There seems to be some thickening at the sides of the aperture, and the ventral wall is somewhat flattened. A well-marked sinuous interthecal groove (text-fig. 6d, e) reveals a thecal overlap of three-fifths. In normal compressed rhabdosomes the interthecal groove is not seen, and the thecae show a fairly pronounced geniculation, becoming less pronounced distally and situated nearer the aperture of the theca. Unfortunately the specimens preserved in relief provide no evidence of the mode of development of the rhabdosome, and flattened early growth stages cannot be interpreted in any detail. The symmetrical distribution of thU and thT^ suggests that development is of the streptoblastic type. The sicula is a little over 2 mm. in length and the apex extends to between the level of the third and fourth thecal pair. Remarks. One specimen from loc. 2206 (parish of Wellsford) represents a form more like G. shelvensis; the proximal end is damaged, but thU appears to open at the level of th2^ (resulting in a much narrower proximal end) and the whole rhabdosome is nar- rower and the thecae more suggestive of shelvensis. Another specimen from loc. ^ (parish of Lancefleld) shows a number of rhabdosomes which also differ in having a more pointed proximal end and seemingly more denlatus-Wkc thecae, but the preserva- tion is poor. The bulk of the material, however, conforms closely to the norm. As Harris and Keble observe, however, there is variation in thecal appearance as an effect of com- pression, and as pointed out above there is a slight progressive change in the thecae along a rhabdosome in the position where the ‘geniculation’ occurs on the metatheca and the extent to which it is developed. Occurrence. The best material used in the above description was from loc. A3 10, parish of Sedgwick, allotment 2, section 1 1 . The subspecies characterizes a well-marked zone (Dl) at the base of the Darriwilian in Victoria. Glyptograptus austrodentatus var. anglicus var. nov. Plate 97, fig. 4; text-fig. la-d Diagnosis. Rhabdosome small, 1-1 i cm. long, widening from an initial width of about 1-2 mm. (lT-1'3 mm.) at the level of thH aperture to a maximum of about 1-7 mm. (1-6- 1-8 mm.). Thecae numbering 7-7| in 5 mm. (maximum range 13-17 in 10 mm.) in the more distal part of the rhabdosome, of the type of G. austrodentatus. Sicula 1 4 mm. in length, the apex extending to the level of the third thecal pair with conspicuous virgella; thH and thl'^ symmetrically disposed, with delicate subapertural spines. Holotype. SM A53381 (text-fig. 7b): Shelve Church Beds, Shelve, Shropshire. O. M. B. BULMAN: GLYPTOGRAPTUS DENTATUS (BRONGNIART) 681 Description. This variety resembles G. austrodentatiis in the characters of the thecae, with an incipient geniculation about halfway along the metatheca, and the blunt proximal end with thl^ and thP almost symmetrically disposed; differences are enumerated below. The sicula is known from an early growth stage (text-fig. Id) where it is seen to be only 14 mm. in length and to extend to the level of the third thecal pair. It is provided TEXT-FIG. 7. a, Glyptograptiis austrodentatus cf. auglicus. Skiddaw Slates (? D. hinimlo Zone), NW. of Lingside, Cumberland, SM A18162. b-d, G. austrodentatus auglicus var. nov. b, Holotype, SM A53381. c, SM A45892. d. Early growth stage (and counterpart), SM A40789a and b. e-g, Glyptograptiis austrodentatus miitabilis var. nov. e, SM A40799. /, Holotype, SM A45906a. g, SM A40786. b-g. Shelve Church Beds (? I. gibberuliis Subzone), Shelve Church, Shropshire. All figures X 5 approx. with a conspicuous virgella which may show some secondary cortical thickening. The thecae number 14-16 in 10 mm. (exceptionally 13 or more than 16); they show a pronounced glyptograptid curvature due to the presence of a slight geniculation about halfway along the metatheca; the infragenicular portion is concave (in some preserva- tions strongly so) and the supragenicular portion more gently curved and slightly inclined outwards. A slight ventral apertural ‘lip’ (similar to that described in G. shelvensis) appears to be present, but there is no introversion. The visible portions of thH and thl^ are more or less symmetrically developed and it is inferred that the mode of development of the rhabdosome is of the streptoblastic type. The resulting square-cut termination of the rhabdosome is a conspicuous feature. 682 PALAEONTOLOGY, VOLUME 6 It differs from austrodentatus austrodentatus in the possession of more numerous thecae; the possession of a smaller sicula, with a more pronounced virgella; and a less conspicuous difference in size between the thecae of the first and second thecal pairs. The austrodentatus thecae typically show a slight but distinct introversion, which is lack- ing in angUcus. Occurrence. The variety occurs most abundantly in the I. gibberulus Subzone (p. 678) of the Shelve Church Beds in Shropshire. Poorly preserved specimens, which appear to be referable to this form, occur in Skiddaw Slates in the D. hirundo Zone, but have not so far been noted from undoubted I. gibberulus Subzone there. Glyptograptus austrodentatus var. mutabilis var. nov. Plate 97, fig. 5 ; text-fig. le-g Diagnosis. Rhabdosome rarely exceeding 1 cm. in length, with an almost uniform breadth of 1-5-1 -7 mm. (TO-1-2 mm. at thP). Thecae 14-15 (rarely 13) in 10 mm., resembling in general character those of var. anglicus. Proximal end asymmetric, with thP upwardly directed and opening near the level of the aperture of th2^. Holotype. SM A45906a, b (Plate 97, fig. 5 ; text-fig. 7 /). Remarks. This form differs from anglicus in possessing a somewhat more slender rhab- dosome, rather fewer thecae in 10 mm., and principally in the character of the proximal end, which is not so bluntly truncated and approximates in appearance to that of the prosoblastic group. It would seem to be transitional between the two large groups, but since details of the proximal end are unknown it is included in the austrodentatus group because of the characters of the thecae and the greater breadth of the rhabdosome at the proximal end. Occurrence. This form occurs not infrequently in the Shelve Church Beds (/. gibberu- lus Subzone), Shropshire. Glyptograptus austrodentatus var. oelandicus var. nov. Plate 97, figs. 16, 17; text-figs. 2a-d, 8 1936 G. Bulmanp. 49, pi. 3, figs. 5-7, 12-13, 14-21, and pi. 4, figs. 4-6, text-figs. 18-21. Diagnosis. Rhabdosome 1 cm. or more in length, widening from an initial width of about 1 mm. at thC to a maximum of about 1-8 mm. Eight thecae occur in the first 5 mm. of the rhabdosome length, but distally number about 14-15 in 10 mm. Thecae with pronounced sigmoid curvature, the geniculation situated about midway on the meta- theca proximally; and about two-thirds along the metatheca distally; overlap rather more than one-half proximally, half distally. Median septum undulating, markedly so at the proximal end. Sicula T3-T8 mm. in length, with prominent virgella; development pronouncedly streptoblastic in type, with ‘truncated’ proximal end and subsymmetrical disposition of thE and thl^. Holotype. G. dentatus Bulman 1936, pi. 3, figs. 5-7 (Holm Collection, no. 307, Riksmuseum, Stock- holm); Glaukonithaltig gra Vaginatumkalk, Hallunden, Oland. Here reproduced as Plate 97, figs. 16, 17. O. M. B. BULMAN: GLYPTOGRAPTUS DENTATUS (BRONGNIART) 683 Description. The variety has been described in Bulman (1936, pp. 49-57) and the mode of development described on pp. 5-6 and shown in text-fig. 2 is based on an excellent series of microtome sections and an incomplete series of growth stages. The form of the thecae in undistorted three-dimensional material is best seen in one of Dr. Skevington’s transparencies on which text-fig. 8^/ is based. As in G. clentatus, the interthecal septum is sometimes not developed, or is extremely short, even though its course is indicated externally on the lateral walls of the rhabdosome. TEXT-FIG. 8. Glyptograptus austrodentatiis oelandiciis var. nov. a. Reverse and obverse views of specimen 1647. b. Reverse and obverse views of specimen 1646. c. Distal portion of rhabdosome, specimen 1641. All specimens from Orthoceras Limestone, Halludden, Oland, Holm Collection, Riksmuseum, Stockholm; from Bulman 1936, text-figs. 19 and 18a. x5 approx, d. Distal end of rhabdosome showing form of thecae and growth-lines. Halludden, Oland; D. Skevington prep. (Palaeontological Institute, Uppsala, Ol. 1230). x 15 approx. Remarks. In the original description (1936, p. 52) reference was made to the existence of broad and narrow forms, the narrow forms having a more pointed proximal end, rather less overlap, and a more distally sited geniculum. It may be that these narrower forms have a similar relation to typical var. oelandiciis that the forms here described as var. mutabilis bear to var. angliciis, but it seems unlikely that they represent real transients to G. dentatus at this rather high horizon when the dentatus type had already been differentiated. The variety differs from aiistrodentalus in having a less ‘truncated’ proximal end, a less parallel-sided rhabdosome, and more numerous thecae. ThU and thD are not so symmetrically disposed, and the thecae are more alternating and lack any ventral apertural ‘lip’. From var. angUcus it differs principally in its rather more pointed proximal end, less symmetrically disposed basal thecae, and in lacking the vertical apertural ‘lip’. Occurrence. Glaukonithaltig gra Vaginatumkalk, Oland {D. hirundolbifidus Zone). Glyptograptus austrodentatus var. americanus var. nov. Plate 97, figs. 18, 19; text-figs. 2e-i, 3, 9 1960 Glyptograptus cf. austrodentatus'. Berry, p. 87, pi. 13, fig. 2, 3. Diagnosis. Rhabdosome 1-1 1 cm. long, widening from an initial width of about 1 -3 mm. at thU to a maximum of 1-6-1 -9 mm. Thecae numbering 5-6 in 5 mm. in more distal part of the rhabdosome, of the general type of G. austrodentatus, but with a prominent 684 PALAEONTOLOGY, VOLUME 6 ventral apertural ‘lip’ (as in G. shelvensis), overlapping one-half proximally and some- what more distally. Median septum complete, slightly or conspicuously undulating; proximal end with thl^ and thl^ subsymmetrically disposed, a prominent virgella, and delicate subapertural spines on thP and thP. TEXT-FIG. 9. Glyptograpliis austrodentatus ainericamis var. nov. a, SM A53803, Marathon, Texas, obverse view of speci- men before sectioning of proximal end. b, SM A23025a. c, SM A23024. d, SM A53801. b-d. From Shumardia Lime- stone (?), Point Levis, Quebec (T. C. Nicholas Coll.). All figures X 5 approx. Holotype. G. cf. austrodentatus'. Berry 1960, pi. 13, fig. 2, 3; Fort Pena formation. Marathon (XPM 20334). Description. This variety differs from the typical form in its slightly larger rhabdosome, larger and fewer thecae, and more prominently developed ventral apertural ‘lip’. The initial thecae, thl^ and thP, tend to be relatively larger and less con- spicuously symmetrical in their disposition than in austrodentatus, though the development is of streptoblastic type. The development is known from an imperfect section series from the limestone material of Marathon, Texas. Crumbling of the encasing plaster resulted in the loss of nearly all sections below the level of the origin of th2^, but the restora- tion shown in text-fig. 2 reveals a sufficiently close approximation to the Holm material from Oland. The initially upward growth and hood formation of thP, and the downward growth of the initial part of th2^ are well displayed here and in the section series, text-fig. 3. Occurrence. Two specimens used for sectioning, from the Fort Pena formation of Marathon, Texas. Three specimens (two of immature rhabdosomes) from Levis, Quebec (T. C. Nicholas Coll., Sedgwick Museum). There is some doubt as to the horizons of these, but they are probably from the lower portion (D^) of the dentatus Zone, and additional specimens from the Shumardia limestone (D^) of Begins Hill have been col- lected by Dr. Cumming. The austrodentatus from the Yukon {T. ensiforrnis and /. gibberulus Subzones of Jackson and Lenz 1962) seems most nearly related to this variety, but the rhabdosome is more slender and the apertural lip of the thecae is not so conspicuously shown. A form possibly related to this has recently been described by Blake (1962) as G. dentatus, from what is probably the D. hirundo Zone of the Trondheim area. Pseudoclimacograptus cumbrensis sp. nov. Plate 96, figs. 7, 8; text-fig. 10 Diagnosis. Rhabdosome 2|-3 cm. in length, widening from 1-2 or T3 mm. at thH to a maximum of T 8-2-0 mm., parallel-sided for the greater part of its length. Thecae 7 in the first 5 mm., 11-12 in 10 mm. distally, of climacograptid type but lacking pro- i O. M. B. BULMAN: GLYPTOGRAPTUS DENTATUS (BRONGNIART) 685 TEXT-FIG. 10. Pseudoclimacograptus cumbrensis sp. nov. a, Holotype, latex cast of natural mould, SM A53044. b. Proximal end, obverse, latex cast of SM A18139. c. Proximal end, reverse, latex cast of SM A18133. d. Distal thecae, flattened, latex cast of SM A53041. e. Distal thecae in partial relief, latex cast of SM A18134. /, SM A18134. All specimens from Skiddaw Slates (? D. hirimdo Zone), Bassenthwaite Sandbeds, Cumberland, x 5 approx. nouncedly angular geniculation, and with gently convex supragenicular ventral wall; length about 2-2 mm., overlapping two-thirds. Development of streptoblastic type, giv- ing a blunt proximal end (with strong virgella), but with thecal apertures alternating. Holotype. SM A53044. Plate 97, fig. 7, text-fig. lOu; Skiddaw Slates, Bassenthwaite Sandbeds, Cumberland. 686 PALAEONTOLOGY, VOLUME 6 Description. This species shows considerable resemblance to Ps. formosus Mu and Lee in its thecal characters, but is an altogether larger rhabdosome. Many specimens are preserved as natural moulds in strong relief and are best studied as latex casts. The development is of streptoblastic type, and an inverted V-shaped swelling at the base of the median septum, separating the initial portions of th3^ and th3^, reveals that there must be an important downward-growing portion of th2h ThL and thP are not, however, symmetrically developed, and the thecae are in the main alternating through- out the rhabdosome. Mature thecae are over 2 mm. in length and overlap for two-thirds of their length. When undistorted, they are seen to be strongly sinuous; but the genicula- tion is not sharply angular and in some preservations (especially when completely flattened) may appear almost glyptograptid. The aperture is slightly introverted, produc- ing what is in effect a slight ventral apertural lip, but it is doubtful whether this is due to the formation of any incomplete fuselli (cf. Jaanusson’s figure, 1960, pi. iv, fig. 5, of Ps. angulatus sebyensis). The median septum is somewhat undulating proximally, becoming straight distally. A stout virgula may project distally, and there is a strong virgella about 1 mm. in length. Remarks. The species differs from Keller’s Pseudoclimacograptus romanovskyi (probably G. teretiusculus zone) in its much greater thecal overlap, and probably (though this is difiicult to determine from Keller’s figures) in the characters of the proximal end. Hsu (1959) has described a variety of that species from, probably, Llanvirn strata under the name sinicus, but this is characterized by prominent basal spines on thH and thP and a gradually widening rhabdosome. In thecal characters and the nature of the proximal end, it comes nearest to Mu and Lee’s Ps. formosus, from a comparable horizon, but the rhabdosome is two or three times as long and nearly twice as broad. Occurrence. Skiddaw Slates — probably D. hirundo Zone — at Bassenthwaite Sandbeds, Cumberland. It is represented in Dr. Jackson’s collections at Newcastle from the D. hirundo Zone of Ling How and Halls Fell. Climacograptus cf. biformis Mu and Lee Plate 96, fig. 6; text-fig. 11a Cf. Climacograptus biformis Mu and Lee, 1958, p. 423, pi. v, figs. 1-5. Remarks. This form is known from a single immature specimen and its counterpart (GSM, JR4140, 4141) from the Skiddaw Slates (? D. hirundo Zone) near Kirkland Church, Cumberland. It agrees in general character and in all dimensions with Mu and Lee’s species from the A. confertus Zone (? = bifidus) of the Ningku Shale, except in possessing rather more numerous thecae (about 7 in 5 mm. compared with 11-12 in 10 mm. in the proximal region of biformis). The distal thecae of CL biformis become glyptograptid (hence the specific name) and there is a strong suggestion of this in the Skiddaw specimen, which was in fact originally identified as G. dentatus. The proximal end suggests derivation from the dentatus group and this would appear to be the earliest true climacograptid. Its generic position in relation to Diplograptus s. str. is, however, somewhat uncertain. O. M. B. BULMAN; GLYPTOG RAPTUS DENTATUS (BRONGNIART) 687 Diplograptus ellesi sp. nov. Plate 96, fig. 9; text-fig. 1 16, c Diagnosis. Mature rhabdosome about 3 cm. in length, widening from an initial width of 0-8-0-9 mm. at thP to a maximum of 2-0-2-1 mm., thereafter typically narrowing again to about 1-8 mm. Thecae 7-8 in the first 5 mm., 12-13 in 10 mm. distally, overlapping about two-thirds their length; those at the proximal end are of amplexograptid type, with apertural excavations about one quarter the rhabdosome width and one-third the length of the free ventral wall; becoming glypto- graptid distally. Periderm somewhat attenuated. There is a strong wiry virgula, commonly projecting well beyond the distal extremity, but the median septum is invisible. Virgella conspicuous. Holotype. BMNH, Q1174. Plate 96, fig. 9, text-fig. lie. Skid- daw Slates (D. bifidiis Zone), Ellergill, near Milburn, Westmor- land. Description. The full-grown rhabdosome is some 3 cm. in length, widening steadily from a narrow but rounded proximal end (0-8-0-9 mm. at the level of the aperture of thP) to a maximum of 2-0-2T mm. Typically it narrows again slightly towards the distal end (1-7-1 -8 mm.) giving it a distinctive, slightly fusiform shape. The periderm is evidently somewhat attenuated, so that shale specimens, strongly compressed, show traces of a wiry virgula throughout the length of the rhabdosome. No definite traces of a median septum have been dis- cerned, but its presence is probable. The proximal end possesses a stout virgella with some indication of secondary cortical thickening. The initial thecae appear to be small and are rela- tively inconspicuous; details of development are un- known. The proximal thecae are of amplexograptid type, with rounded and only slightly inclined apertural text-fig. 11. a, CUmacogmptus cf. excavations, which are about one quarter the rhabdo- some width and about one-third the length of the free ventral wall, and may show a slight selvage. The supra- genicular wall is almost parallel to the rhabdosome axis and the apertural margins are gently concave in true profile. Distally the thecae become glyptograptid in character and have an overlap of nearly two-thirds. In much of the material, the transition from one type to the other occurs well within the first centimetre of rhabdosome length in the region of the twelfth thecal pair; but in biformis Mu and Lee, Skiddaw Slates (? D. hirimdo Zone), Kirk- land Church, Cumberland; GSM, JR4141. b, c, Diplograptus ellesi sp. nov., D. bifidiis Zone, Ellergill, Milburn, Westmorland, b, SM A18129. c, Holotype, BMNH, Q1174. X 5 approx. 688 PALAEONTOLOGY, VOLUME 6 one specimen the amplexograptid thecae persist distally to about the eighteenth thecal pair. Remarks. In general character this species resembles D. decoratus Harris and Keble. No specimen has yet been observed, however, with the distinctive heart-shaped vesicle at the distal termination of the virgula, all the dimensions of the rhabdosome are con- siderably smaller, and the thecae are more closely spaced. It may prove to be only varietally distinct ; but to record it here under the name decoratus might lead to insecure stratigraphical correlations. Occurrence. Skiddaw Slates — D. bifidus Zone; Ellergill and Thornship Beck, near Shap, Westmorland. REFERENCES BERRY, w. B. N. 1960. Graptolitc faunas of the Marathon region, Wcst Texas. University of Te.xas, Publ. 6005. BLAKE, D. H. 1962. A new Lower Ordovician graptolite fauna from the Trondheim region. Norsk, geol. Tidsskr. 42, 223-38. BRONGNiART, A. 1828. Histoire des vegetaiix fossiles. Paris. BRONN, H. G. 1848, 1849. Index Palaeontologiciis A. Nonienclator Palaeontologicus; B. Enumerator Palaeontologicus. Stuttgart. BULMAN, o. M. B. 1931. South American graptolites. Ark. Zool. 22\, 3. 1936. On the graptolites prepared by Holm, pt. 7. Ibid. 28a, 17. • 1958. The sequence of graptolite faunas. Palaeontology, Loud. 1, 159-73. ELLES, G. L. 1922. The graptolite faunas of the British Isles. Proc. Geol. Ass., Lond.33,\6%-2Q(3. and WOOD, E. M. R. 1907. A monograph of British graptolites, pt. 6. Palaeontogr. Soc. [Monogr.]. HALL, J. 1858. Geol. Survey Canada, Kept, for 1857. 1865. Canadian Organic Remains, dec. 2: Graptolites of the Quebec Group. Montreal. HARRIS, w. j. and keble, r. a. 1932. Victorian graptolite zones, with correlations and description of species. Proc. roy. Soc. Viet. 44, 25-48. and THOMAS, D. E. 1935. Victorian graptolites (New Series) part 3. Ibid. 47, 288-313. 1938. A revised classification and correlation of the Ordovician graptolite beds of Victoria. Min. geol. J. 1, 62-72. HOPKiNSON, J. and lapworth, c. 1875. Descriptions of the graptolites of the Arenig and Llandilo Rocks of St. David’s. Quart. J. geol. Soc. Lond. 31, 631-72. HSi), s. c. 1959. A new graptolite fauna from the Lower Ordovician Shale of Tsaidam, Chinghai Province. Acta palaeont. sin. 1, 161-91. JAANUssoN, V. 1960. Graptoloids from the Ontikan and Viruan (Ordov.) Limestones of Estonia and Sweden. Bull. geol. Instn. Univ. Uppsala, 38, 289-366. JACKSON, D. E. 1962. Graptolite zones in the Skiddaw group in Cumberland, England. J. Paleont. 36, 300-13. and LENZ, A. c. 1962. Zonation of Ordovician and Silurian graptolites of northern Yukon. Bull. Amer. Ass. Petrol. Geol. 46, 30-45. MU, A. T. and LEE, c. K. 1958. Scandent graptolites from the Ningkuo Shale of the Kiangshan- Changshan Area, western Chekiang. Acta palaeont. sin. 6, 391-427. NICHOLSON, H. A. 1868. The graptolites of the Skiddaw Series. Quart. J. geol. Soc. Lond. 24, 125-45. pribyl, a. 1947. Classification of the genus Climacograptus Hall 1865. Bull. Acad, tcheque sci. 48, 2, 1-12. ruedemann, R. 1904. Graptolites of New York; 1. Graptolites of the Lower Beds. New York State Museum Mem. 7. 1947. Graptolites of North America, Geol. Soc. Amer., Mem. 19. O. M. B. BULMAN; GLYPTOG RAPTUS DENTATUS (BRONGNIART) 689 STERNBERG, K. 1838. Versuck einer . . . Darstelliaig der Flora der Vorwelt, fasc. 7. Leipzig and Prague. THOMAS, D. E. 1960. The zonal distribution of Australian graptolites. J. roy. Soc. N.S. W. 94, 1-58. URBANEK, A. 1959. On the development and structure of the graptolite genus Gymnograptiis Bulman. Acta palaeont. polon. 4, 279-338. O. M. B. BULMAN Sedgwick Museum, .Manuscript received 22 January 1963 Cambridge THE SYSTEMATIC STATUS OF OPPEL’S SPECIMENS OF BELEMNITES GERARDI by G. R. STEVENS Abstract. Oppel’s original specimens of Belenmites gerardi from Kalabagh (Punjab) are described and refigured. The name is restricted to the specimens figured as Oppel’s pi. 88, figs. 1 , 2, and the original of fig. 3 is assigned to a new species, Belemnopsis uhligi, based on one of Uhlig’s specimens from the Spiti Shales (northern India). Synonymy and diagnoses are provided for B. gerardi and B. uhligi and their stratigraphic ranges discussed. It is shown that Belemnopsis kimtkotensis (Waagen) is a synonym of B. gerardi and that the majority of belemnites hitherto assigned to B. gerardi should be reassigned to B. uhligi. The species usually known as Belemnopsis gerardi (Oppel), e.g. as interpreted by Uhlig (1903-10, pp. 386-8), is a common belemnite of the Kimeridgian of the Indo-Pacific region, having been recorded from numerous localities in Indonesia (Kruizinga 1921; Stolley 1929, 1934, 1935) and northern India (Oppel 1863; Uhlig 1903-10). Related species have been recorded from East Africa (Tate 1867), northern Australia (Teichert 1940), New Caledonia (Avias 1953), and New Zealand (Marwick 1953). The name was introduced in 1863 (Oppel 1863, p. 273), but the species was not described and figured until 1865 (Oppel 1865, pp. 296-7). Study of the species has been hampered by lack of adequate figures and descriptions of Oppel’s original specimens (1865, pi. 88, figs. 1-3). Through the courtesy of Dr. K. Werner Barthel of the Bayerische Staatssammlung fiir Palaontologie und historische Geologic, Munich, the writer has been able to examine these specimens. The species was based on two complete specimens (Oppel 1865, pi. 88, figs. 1, 2), and a third incomplete specimen (fig. 3) which was only tentatively referred to the species by Oppel. All three specimens are preserved in the Schlagintweit Collection in Munich and are labelled ‘ Macroccp/irz/w^-schichten, Kalabagh im Ob. Punjab’. The accession numbers are as follows: 1872. xv. 502 (Oppel 1865, pi. 88, fig. 1); 1872. xv. 501 (fig. 2); 1872. XV. 46 (fig. 3). Kalabagh (32° 58' S., 71° 36' E.) is on the right bank of the Indus, some 70 miles south of Peshawar. Previous workers have agreed that more than one species is represented by Oppel’s specimens, but there has been no agreement on the identity of these species. In 1929 Stolley examined Oppel’s specimens, and based his concept of the species (1929, pp. 147, 151) on Oppel’s fig. 3 (Oppel 1865, pi. 88). He identified Oppel’s fig. 1 as Belemnopsis aucklandica (Hauer) (1929, pp. 151, 168) and fig. 2 as B. alfurica (Boehm) (op. cit., pp. 151, 172). Spath (1927-33, pp. 661-2) supported Stolley’s identification of Oppel’s fig. 2 as B. alfurica, but did not agree that fig. 1 represents a specimen of B. aucklandica. Spath rejected Stolley’s choice of the original of Oppel’s fig. 3 as the type (properly a lectotype) for B. gerardi, as this specimen was only tentatively referred to the species by Oppel. Spath took two of Uhlig’s specimens (1903-10, pi. 93, figs. 5, 7) to be typical B. gerardi of the Spiti Shales, though Stolley disagreed with this selection (Spath 1927-33, pp. 4, 661). Teichert (1940, p. 114) designated Oppel’s fig. 1 as type (lectotype) for (Palaeontology, Vol. 6, Part 4, 1963, pp. 690-8, pis. 98-99.] G. R. STEVENS: OPPEL’S SPECIMENS OF BELEMNITES GERARDI 691 B. gerardi. This appears to be the first valid selection of lectotype for Oppefs nominal species and is accepted here. The original of Oppefs pi. 88, fig. 3 (PI. 98, figs. 9-13; PI. 99, fig. 4) is a fragment of the alveolar region, 52 mm. in length. Measurements are as follows (in mm.): Posterior end Transverse diameter (dt): 21-5 Sagittal diameter (ds): 21-5 Anterior end: dt:23 0 ds:22-0 At the approximate position of the protoconch: dt:220 ds:21-5 The guard is massive and non-hastate, with a well-developed ventral groove. The groove is 7 mm. wide where dt = 22-0 mm., and 3 mm. deep where ds = 21-5 mm. Lateral lines and a dorsal groove appear to be absent. The originals of Oppefs pi. 88, figs. 1 and 2 (PI. 98, figs. 1-8) are reasonably complete and their measurements (in mm.) are given below (for explanation of abbreviations see Avias 1953, pp. 158-9). Oppel I li V dtM dtm dsM dsin 11 jv Ht Hs A Rs Fig. 1 65 26 39 9-5 8-5 8-5 90 0-66 111-7 94-4 111-7 0-27 Fig. 2 72 27 45 11-5 10-5 10-5 110 0-6 109-5 95-4 109-5 0-3 In the original of fig. 2 (PI. 98, figs. 5-8) the guard is elongate and sharply pointed. The outline and profile are symmetrical, and in both the sides gradually taper towards the apex. The outline of the guard is slightly hastate. The ventral groove reaches almost to the apex, and is broad and moderately shallow in relation to the dimensions of the guard. The groove is 1 mm. deep where ds = 10-5 mm., and 3-5 mm. wide where dt = 11-5 mm. Lateral lines are not visible, and a dorsal groove is absent. The cross- section of the guard is depressed in the apical and stem regions, but becomes com- pressed in the alveolar region. The original of fig. 1 (PI. 98, figs. 1-4) is a juvenile and is more hastate than that of fig. 2, but otherwise shows the same characters. Traces of lateral lines are present on the flanks. This specimen appears to have been slightly distorted; the apical region has been bent in a dorsal direction, producing an apparent downwarping of the posterior portion of the guard. Notwithstanding this distortion, it appears that the original specimens of Oppefs figs. 1 and 2 are of the same species, and distinct from Oppefs fig. 3. Boehm (1907, p. 55) refigured the cross-sections of Oppefs pi. 88, fig. Ic (ibid., fig. 20), and pi. 88, fig. 3b, c (ibid., figs. 21<7, b)\ and as may be seen from these, the cross-section of fig. 1 (compressed) is quite different from that of fig. 3 (circular). The characters of the originals of Oppefs figs. 1 and 2 are identical with those of Belenmopsis kimtkotensis (Waagen) recorded from Upper Oxfordian localities in Kachh by Waagen (1873, pp. 3-5) and Spath (1927-33, pp. 8, 9). Waagen’s description of B. kimtkotensis applies perfectly to the originals of Oppefs figs. 1 and 2, except that he mentions the presence of a dorsal alveolar groove which can- not be identified in Oppefs specimens because of their incompletely preserved alveolar regions. This feature, however, is usually impersistent in Belenmopsis and has no signifi- cance (cf Spath 1927-33, pp. 662-3). Like many species of Belenmopsis, B. kimtkotensis becomes less hastate with maturity. 692 PALAEONTOLOGY, VOLUME 6 Whereas the juvenile is hastate (as noted by Waagen 1873, p. 4) the adult is almost non- hastate (e.g. Waagen’s type of B. kimtkotensis; Waagen 1873, pi. 1, figs. 3a-e; Spath 1927-33, pi. 1, fig. lb). The type of B. kimtkotensis is evidently a more mature stage than the originals of Oppel’s figs. 1 and 2, with fig. 2 representing a later stage than fig. 1. In the table below the measurements of the juvenile specimen identified by Spath (1927- 33) as B. aff. kimtkotensis (ibid., pi. 2, fig. 4) are given, along with the measurements of Waagen’s holotype, taken from Waagen (1873, p. 4) and Spath’s figure (1927-33, pi. 1, fig. 1/)). The original of Spath’s pi. 1, fig. la (BMNH C19921) is too crushed to provide accurate measurements. Spath 1927-33, pi. 2, fig. 4. BMNH, 1 u V dtM dtni dsM dsm ujv Ht Hs A Rs Cl 9943. Waagen’s holotype; measurements taken from Waagen 1873, p. 4, and Spath 1927- 46 16 30 8-5 70 6-7 70 0-53 121-4 95-7 126-8 0-33 Juvenile 33, pi. 1, fig. \b 78 20 Est. 58 Est. 11-5 Est. 110 110 Est. 120 0-34 104-5 91-6 104-5 0-29 The original of Oppel’s fig. 3 conforms in its characters to those of specimens usually identified by previous workers as Belemnopsis gerardi. But as fig. 3 is not a specimen in Oppel’s type series, its designation by Stolley (1929) as lectotype for gerardi is invalid and Teichert’s designation (1940, p. 114) of the original of fig. 1 must be accepted. Therefore the name gerardi must take precedence over kimtkotensis and a new species proposed for belemnites resembling the original of Oppel’s fig. 3, previously identified as B. gerardi. Belemnopsis gerardi is redefined below and a new species, Belemnopsis iddigi, proposed. SYSTEMATIC DESCRIPTIONS Belemnopsis gerardi (Oppel) Plate 98, figs. 1-8 1863 Belemnites gerardi Oppel, p. 273 (nom. nudum). 1865 Belemnites gerardi (partim) Oppel, pp. 296-7, pi. 88, figs. 1, 2 (non fig. 3). 1873 Belemnites kimtkotensis Waagen, pp. 3-5, pi. 1, fig. 3. 1879 Belemnites grantianus d’Orbigny; Medlicott and Blanford, pi. 12, fig. 2. 1893 Belemnites grantianus d’Orbigny; Oldham, pp. 222-4. EXPLANATION OF PLATE 98 Figs. 1-4. Belemnopsis gerardi (Oppel). Lectotype. Upper Jurassic of Kalabagh, Upper Punjab. Accession No. 1872. xv. 502, Schlagintweit Collection, Munich. Original of Oppel 1865, pi. 88, fig. 1. 1, Ventral view. 2, Dorsal view. 3, Left lateral view (i.e. ventral groove facing left). 4, Right lateral view. All X 1 . Figs. 5-8. Belemnopsis gerardi (Oppel). Upper Jurassic of Kalabagh, Upper Punjab. Accession No. 1872. XV. 501, Schlagintweit Collection, Munich. Original of Oppel 1865, pi. 88, fig. 2. 5, Ventral view. 6, Dorsal view. 7, Left lateral view. 8, Right lateral view. All X 1 . Figs. 9-13. Belemnopsis uhligi sp. nov. Upper Jurassic of Kalabagh, Upper Punjab. Accession No. 1872. XV. 46, Schlagintweit Collection, Munich. Original of Oppel 1865, pi. 88, fig. 3. 9, Cross- section, stem region. 10, Cross-section, alveolar region. 1 1 , Ventral view. 12, Dorsal view. 13, Left lateral view. All X 1. See also Plate 99, fig. 4. Palaeontology, Vol. 6 PLATE 98 STEVENS, Belemnopsis G. R. STEVENS: OPPEL’S SPECIMENS OF BELEMNITES GERARDI 693 1907 Belemnites gerardi Oppel (partim) \ Boehm, p. 55, fig. 20 (non figs. 21a, b). 1927-33 Belemnopsis kuntkotensis (Waagen), B. aff. kiiiitkotensis; Spath, pp. 8-9, pi. 1, figs, la, b; pi. 2, fig. 4. cf. 1929 Belemnites tangauensis Futterer; Weir, p. 18, pi. 2, fig. 23; pi. 5, figs. 19, 20. cf. 1930 Belemnopsis tangauensis (Futterer); Weir, p. 90, pi. 10, fig. 3. cf. 1933 Belemnopsis tangauensis (Futterer) (paiiim); Stefanini, pp. 47-50, pi. 4, figs. 3-5 (non figs. 2, 6-17). ? cf. 1934 Belemnopsis tangauensis (Futterer); Spath, pp. 21, 22. 1935 Belemnopsis kuntkotensis (Waagen); Spath, p. 207, pi. 24, fig. 4. 1939 Belemnopsis gerardi (Oppel) (partim)', Spath, pp. 110, 1 1 1, pi. 24, fig. 12 (non figs. 11, 13). 1951 Belemnopsis aff. kuntkotensis (Waagen); Nicolai, pp. 33, 34. ? var. 1953 Belemnopsis kuntkotensis (Waagen) var. puenensis Avias, pp. 156-7, 164-5, pi. 14, figs. 4, 9, 12, 13, 15; pi. 15, figs. 24, 30, 31, 34, 36-39; pi. 16, figs. 5, 11, 17. 1956 Belemnopsis kuntkotensis (Waagen); Hunt, p. 12. non 1845 Belemnites grantianiis d’Orbigny, p. 307, pi. 58. non 1873 Belemnites gerardi Oppel; Waagen, pp. 13, 14, pi. 2, fig. 3 (see Spath 1927-33, pi. 1, fig. 3fl). non 1894 Belemnites tangauensis Futterer, pp. 30-32, pi. 5, figs. 2, 3. }wn 1935 Belemnopsis kuntkotensis (Waagen); Spath, p. 218 (original of Spath 1927-33, pi. l,fig. 3b). Diagnosis. A Belemnopsis with an elongate and tapering guard. Outline and profile symmetrical, sides and ventral and dorsal surfaces gradually tapering towards the apex. Outline slightly hastate, more so in sub-mature forms. Apex not inflated, tapering. Depressed cross-sections in apical and stem regions, compressed in alveolar region. Ventral groove deep and moderately broad, extending from alveolar region and almost reaching apex. Groove deepest in alveolar region, gradually shallowing towards apex. Poorly developed lateral lines and dorsal alveolar groove may or may not be present. Lectotype. Original of Oppel 1865, pi. 88, fig. 1, accession number 1872. XV. 502, Schlagintweit Collection, Bayerische Staatssammlung fiir Paliiontologie und historische Geologic, Munich. Desig- nated Teichert 1940, p. 114 (see also Glaessner 1945, p. 155). Type locality. Kalabagh, Upper Punjab. Localities and stratigraphic range. Upper Oxfordian-Middle Kimeridgian. The species occurs in the Punjab (Pakistan) and Kachh (India), Somalia, Mada- gascar, and probably in Kenya and New Caledonia. In the Kachh sequence it is known from the Kantkot (Kuntkote) Sandstone (Upper Oxfordian) and in Somalia from the lower part of the Daghani Shales (Middle Kimeridgian). The Madagascar and New Caledonia occurrences have been dated as Upper Oxfordian to Lower Kimeridgian. The precise age of the beds at Kalabagh is unknown. Spath (1927-33, pp. 661-2), influenced by Stolley’s work on Indonesian belemnites, favoured an Upper Oxfordian age for B. gerardi from Kalabagh. His opinion changed, however, and he later stated (Spath 1939, p. 110; see also 1927-33, p. 802) that there was no evidence for an Upper Oxfordian age for Oppel’s specimens of B. gerardi, and implied that they came from beds nearly equivalent to the Chidamu Beds of Spiti (i.e. Lower Tithonian). Both B. gerardi (Oppel’s figs. 1 and 2) and B. nhligi (fig. 3) occur at Kalabagh, but their stratigraphic relationship is unknown. In New Zealand species allied to B. uhligi first appear in the Middle Kimeridgian and range up to Lower or Middle Tithonian. Therefore if all three of Oppel’s specimens came from approximately the same horizon, a Middle Kimeridgian age is favoured. z z C 1713 694 PALAEONTOLOGY, VOLUME 6 Belemnopsis uhligi sp. nov. Plate 98, figs. 9-13; Plate 99, figs. 1-9 1833 Belemnites; Everest, pi. 1, fig. 17. 1863 Belemnites sulcatus Miller; Blanford, p. 125, pi. 1, figs. 1, la-c. 1865 Belemnites gerardi (partim) Oppel, pp. 296-7, pi. 88, fig. 3 (non figs. 1, 2). 1866 Belemnites canaliculatus Schlotheim; Stoliczka, pp. 111-12. 1889 Belemnites gerardi Oppel (partim)-, Neumayr, pp. 52-56 (only references to Oppel’s fig. 3). ? aff. 1892 Belemnites gerardi OppQl (partim)-, Rothpletz, pp. 104-5, pi. 13, fig. 10 (non figs. 6-8, 12). 1903-10 Belemnites (Belemnopsis) gerardi Oppel (partim)-, Uhlig, pp. 386-8, pi. 93, figs. 5, 7, 9 (non pi. 93, figs. 1, 2, 10-13; pi. 93a, figs. 1, 2, 4, 5). 1907 Belemnites gerardi Oppel (partim); Boehm, p. 55, figs. 2la, b (non fig. 20). 1920 Belemnopsis gerardi (Oppel) (partim); Biilow-Trummer, pp. 129-30 (non occurrences in S. Africa, N. Alps, and France). 1921 Belemnopsis gerardi (Oppel); Kruizinga, pp. 163-6, pi. 1, figs. 1-4; pi. 2, fig. 11. ? 1927-33 Belemnopsis cf. gerardi (Oppel); Spath, pp. 706-7. 1929 Belemnopsis gerardi (Oppel) (partim); Stolley, pp. 151-7, pi. 1, figs. 18, 20, 22, 24-29; pi. 2, figs. 1-3 (non pi. 1, figs. 16, 17, 19, 21, 23, 30-32). 1931 Belemnopsis gerardi (Oppel); Kruizinga, p. 368. 1931 Belemnopsis gerardi (Oppel); Wanner, pp. 585-95. 1934 Belemnopsis gerardi (Oppel) typ. ; Belemnopsis aff. gerardi; Stolley, pp. 470-86. 1935 Belemnopsis gerardi (Oppel); Stolley, pp. 49, 50. aff. 1939 Belemnopsis gerardi (Oppel) (partim); Spath, pp. 110-11, pi. 24, figs. 11, 13 (non fig. 12). 1945 Belemnopsis gerardi (Oppel); Glaessner, pp. 155-6, pi. 6, figs. 8, 9. 1956 Belemnopsis gerardi (Oppel); Marks, pp. 199, 200. ex.gr. 1956 Belemnopsis gerardi (Oppel); Wanner, p. 135. 1956 Belemnopsis gerardi (Oppel); Arkell, pp. 408, 447. 1957 Belemnopsis gerardi (Oppel); Holland et a!., pp. 86, 160, 247. Jion 1820 Belemnites canaliculatus Schlotheim, p. 49. non 1823 Belemnites sulcatus Miller, p. 59, pi. 8, figs. 3-5. non 1873 Belemnites gerardi Oppel; Waagen, pp. 13, 14, pi. 2, fig. 3. non 1895 Belemnites gerardi Oppel; Kihan, p. 673. non 1914 Belemnites sp., related to B. gerardi; Spitz, pp. 222-3, pi. 19, figs. 11, 12. no)i 1924 Belemnites gerardi Oppel; Broili, pp. 8, 9; pi. 2, fig. 9. non 1951 Belemnopsis cf. gerardi (Oppel); Brunnschweiler, p. 8. non 1958 Belemnopsis cf. gerardi (Oppel); McWhae et ah, p. 90. Diagnosis. A Belemnopsis with a short robust guard, cyhndrical or cylindro-conical. Outline and profile symmetrical, both non-hastate. Apex not inflated, usually tapering. Apical, stem, and alveolar cross-sections usually circular. Ventral groove broad and deep. EXPLANATION OF PLATE 99 Figs. 1-3,5. Belemnopsis uhligi sp. nov. Holotype. Upper Jurassic (Spiti Shales) of the Niti Pass area, northern India. Accession No. 10161, Geological Survey of India Museum, Calcutta. Original of Uhlig 1903-10, pi. 93, figs. 9a, b. 1, Ventral view. 2, Dorsal view. 3, Right lateral view (ventral groove facing right). 5, Cross-section at anterior end. All x 1. Fig. 4. Belemnopsis uhligi sp. nov. Upper Jurassic of Kalabagh, Upper Punjab. Accession No. 1872. XV. 46, Schlagintweit Collection, Munich. Original of Oppel 1865, pi. 88, fig. 3. Right lateral view, X 1. See also Plate 98, figs. 9-13. Figs. 6-9. Belemnopsis uhligi sp. nov. Upper Jurassic (Uppermost Fatjet Shales) of Fatjet, Misol Archipelago, Indonesia. Accession No. 116/1, Weber Collection (Locality M.36, see Stolley 1934, p. 484), Naturhistorisches Museum, Basel. 6, Ventral view. 7, Dorsal view. 8, Left lateral view. 9, Right lateral view. All x 1 . Palaeontology, Vol. 6 PLATE 99 STEVENS, Belemnopsis G. R. STEVENS: OPPEL’S SPECIMENS OF BELEMNITES GERARDI 695 extending from alveolar region to apex. Groove deepest in alveolar region, gradually shallowing towards apex. Lateral lines not observed. Holotype. Original of Uhlig 1903-10, pi. 93, figs. 9a, b. No. 10161, Geological Survey of India collec- tions, Calcutta. Type locality. Spiti Shales, Jandu, Hundes (Niti Pass area, Tehri Garhwal and Garhwal provinces, northern India). Localities and stratigraphic range. Middle Kimeridgian to Middle (or Upper?) Tithonian. Belenmopsis uhligi occurs in the Himalayas (Spiti and Niti regions) and Kachhin India, the Salt Range (Kalabagh) in West Pakistan, Indonesia (Timor, Roti, Sula Islands, Misol, Celebes, Buton), and the Central Highlands of New Guinea. B. uhligi is present in the Ganjansar Beds (Middle Tithonian), which form the upper- most subdivision of the Katrol Series of the Jurassic succession in Kachh (see Nath 1932, p. 167; Holland et al. 1957, p. 86), and Spath (1927-33, pp. 662, 706) recorded B. cf. uhligi (as B. cf. gerardi) from the Umia Ammonite Beds (Upper Tithonian), immediately above the Ganjansar Beds. In the Himalayas B. uhligi occurs in the Lower and Middle Spiti Shales, the main occurrence being in the Lower Spiti. Arkell (1956, p. 407) has dated the Middle Spiti Shales (Chidamu Beds) as Lower Tithonian, but assigned an Upper Oxfordian age to the Lower Spiti Shales. In his dating of the Lower Spiti he was probably influenced by the agreement of previous workers on an Upper Oxfordian age for B. uhligi (i.e. B. gerardi auctt.) and by the work of Uhlig (1903- 10) and Spath (1927-33), whose age determinations he quoted. Species closely related to B. uhligi first appear in the New Zealand belemnite succession in the late Middle Kimeridgian and B. uhligi probably appeared at the same time in the Indian succession. Holland et al. (1957, p. 247) state that the Lower Spiti Shales are equivalent to the Kantkot Sandstone of the Kachh sequence, containing B. gerardi (= B. kuntkotensis), but their belemnite assemblages have nothing in common; B. uhligi appears to be absent in the latter and B. gerardi in the former. A possible explanation of this is that B. uhligi occurs only in the upper portion of the Lower Spiti Shales, and that this portion is not represented in the Kantkot Sandstone. So too B. gerardi may be restricted to the basal portion of the Kantkot Sandstone and this zone is not represented in the Lower Spiti Shales. Stolley (1929) assigned an Upper Oxfordian age to the Indonesian B. uhligi and this age has been accepted by later workers (e.g. Wanner 1931 ; Wandel 1936; Teichert 1940; Vogler 1941; Glaessner 1945) though in later papers Stolley (1934, p. 473; 1935, p. 50) stated that B. uhligi may range into the Kimeridgian. Correlations between Indonesia and New Zealand may be established on the basis of their belemnites, Buchia and Inoceramus, and ages obtained from the ammonite sequence determined by Arkell in New Zealand (Fleming and Kear 1960). These have necessitated a revision of the ages commonly accepted for some of the Upper Jurassic guide fossils in the Indonesian and Australasian regions (Fleming 1960), including the Upper Oxfordian age for B. uhligi (B. gerardi auctt.). In Taliabu (Sula Islands, Indonesia) B. uhligi is associated with Inoceramus galoi Boehm, which is of Lower-Middle Kimeridgian age in New Zealand. In Misol B. moluccana (Boehm), a predecessor of B. uhligi, occurs in the Lilinta Marls (Stolley 1934), which have been correlated with the Middle Kimeridgian of New Zealand, and B. uhligi occurs in the Fatjet Shales, which have been correlated with Lower, and 696 PALAEONTOLOGY, VOLUME 6 probably Middle, Tithonian. In the Kuabgen Group of the Central Highlands of New Guinea (Glaessner 1945) B. uhligi occurs with Buchia malayomaorica (Krumbeck) and Inoceranms sp. resembling I. haasti Hochst., /. subhaasti Wandel and I. galoi Boehm, which in New Zealand indicate Lower-Middle Kimeridgian. In New Zealand species closely related to B. uhligi first appear in the late Middle Kimeridgian, and the occur- rences of B. uhligi in New Guinea and Taliabu mentioned above are thought to be of a similar age. To sum up, B. uhligi probably appeared in the Middle Kimeridgian and ranged through to at least Middle or Upper Tithonian. In northern India, Indonesia, and New Guinea B. uhligi is associated with a number of allied species (the 'gerardi group ’ of Uhlig, Spath, Stolley, &c.) which range into Australia and New Zealand, Iran, Arabia, and East Africa. These allied species differ from B. uhligi in characters such as degree of elongation and hastation, nature of apex and cross- sections of the guard. They will be described in a forthcoming publication (Stevens, in press). Migrations of Hibolithes temporarily replaced Belemnopsis uhligi and allied species in Indonesia and New Zealand during parts of Middle Kimeridgian and Lower Tithonian time. The Belemnopsis uhligi group was finally replaced throughout the Indo-Pacific by migrations of Hibolithes and Duvalia in Upper Tithonian and Neocomian times. Acknowledgements. This paper is part of a study undertaken at the University of Cambridge. The writer gratefully acknowledges the opportunities provided for this research by the award to him of a Shell Post-Graduate Scholarship in Science for 1956-9. He also wishes to thank Mr. A. G. Brighton (Sedgwick Museum, Cambridge), Mr. R. V. Melville (H.M. Geological Survey, London), Dr. J. Mar- wick and Dr. C. A. Fleming (New Zealand Geological Survey) for advice and encouragement and for reading the manuscript critically. Dr. K. Werner Barthel kindly sent Oppel’s specimens on loan and assisted the writer in Munich, and Mr. L. Bairstow assisted in the examination of collections at the British Museum (Natural History). Dr. E. Gasche (Naturhistorisches Museum, Basel) kindly sent on loan many Indonesian belemnites from the Weber Collection and assisted the writer during a visit to Basel. Dr. B. C. Roy, Director of the Geological Survey of India, kindly arranged for photographs of some of UhUg's figured specimens to be sent to the writer. The photographs used for Plate 99, figs. 1-3, 5, were supplied by Dr. Roy and the remainder were taken by Mr. A. Barlow of the Sedgwick Museum, Cambridge. REFERENCES ARKELL, w. j. 1956. Jurassic Geology of the World. Edinburgh and London. AViAS, J. V. 1953. Contribution a 1’ etude stratigraphique et paleontologique de la Nouvelle-Caledonie centrale. Sci. Terre, 1 (1-2). BLANFORD, H. F. 1863. On Dr. Gerard’s collection of fossils from the Spiti Valley, in the Asiatic Society’s Museum. J. Asiat. Soc. Beng. 32, 124-38. BOEHM, G. 1907. Beitrage zur Geologic von Niederlandisch-Indien. 1 ; Die Sudkiisten der Sula-Inseln Taliabu und Mangoli. Pt. 2; Der Fundpunkt am Oberen Lagoi auf Taliabu. Palaeontographica, Suppl. 4, Abt. 1, Abs. 2, 47-58, pi. 8. BROiLi, F. 1924. Zur Geologic des Vogelkop (N. W. Neu-Guinea). Wet. Meded. Dienst Mijnb. Ned.-O.- Ind. 1, 1-13. BRUNNSCHWEiLER, R. o. 1951. Notcs on the gcology of Dampicr Land, N.W. Australia. Aust.J.Sci. 14, 6-8. BULOW-TRUMMER, E. von. 1920. Cephalopoda Dibrancluata. Fossilium Catalogus, 1 {W). Berlin. d’orbigny, A. 1845. Paleontologie Universelle des Coqidlles et des Mollusques. Paris. EVEREST, R. 1833. Memorandum on the fossil shclls discovcrcd in the Himalayan Mountains. Asiat. Researches, 18, 107-14. G. R. STEVENS: OPPEL’S SPECIMENS OF BELEMNITES GERARDI 697 FLEMING, c. A. I960. The Upper Jurassic sequence at Kawhia, New Zealand, with reference to the ages of some Tethyan guide fossils. Kept. 21st sess. hit. geol. Cong. 21, 264-9. and KEAR, D. 1960. The Jurassic sequence at Kawhia Harbour, New Zealand. Bull. geol. Siirv. N.Z., N.S., 67. FUTTERER, K. 1894. Bcitragc ZLir Kcnntniss dcs Jura ill Ost-Afrika. Z. dtsch. geol. Ges. 46, \-49. GLAESSNER, M. F. 1945. Mcsozoic fossils from the central highlands of New Guinea. Proc. roy. Soc. Viet. 56, 151-68. HOLLAND, T. H., et ol. 1957. India, Pakistan, Nepal, Bhutan. Lexiqiie Strut. Inteniat. 3 (8n). HUNT, J. A. 1956. British Somaliland. Lexiqiie Strut. Internut. 4 (5u). KiLiAN, w. 1895. Environs de Sisteron et contributions a la connaissance des terrains secondaires du sud-est de la France. Bull. Soc. geol. Fr., ser. 3, 23, 659-803. KRUiziNGA, p. 1921. Die Belemniten uit de Jurassische afzettingen van de Soela-Eilanden. Juurh. Mijnw. Neci.-Oost-Inil. 49 (2), 161-89. 1931. Depalaeontologieen stratigraphic van NederlandschOost-Indie. 1 : Onze palaeontologische kermis van Nederlandsch Oost-Tndie in 1930. 10: Cephalopoda. Lehl. geol. Meclecl. 5, 297-389 (Martin Feestbundel). MCWHAE, J. R. H., PLAYFORD, P. E., LINDNER, A. W., GLENLSTER, B. F. and BALME, B. E. 1958. The Strati- graphy of Western Australia. J. geol. Soc. Aust. 4 (2). MARKS, p. 1956. Indonesia. Lexiqiie Strut. Internut. 2 (lu). MARWICK, J. 1953. Divisions and faunas of the Hokonui System (Triassic and Jurassic). Puleont. Bull. geol. Surv. N.Z.ll. MEDLicoTT, H. B. and BLANFORD, w. T. 1879. Muniiul of the Geolog}' of IiicUu. Calcutta. MILLER, J. s. 1823. Observations on belemnites. Truns. geol. Soc. Loud., 45-62. NATH, RAJ. 1932. A contribution to the stratigraphy of Cutch. Quurt. J. geol. Soc. Indiu, 4, 161-74. NEUMAYR, M. 1889. Ucbcr einige Belemniten aus Central-Asien und Siidafrika und iiber den Canal der Belemniten. Verli. geol. ReichsAnst. (St Anst.) Wien, 52-56. NICOLAI, M. 1951. Etude de quelques gisements fossiliferes du sud-ouest de Madagascar. Ann. Puleont. 37, 1-46. OLDHAM, R. D. 1893. Gcologv of Iiidiu. Calcutta. OPPEL, A. 1863. Ueber Ostindische Fossilreste aus den Secundaren Ablagerungen von Spiti und Gnari-Khorsum in Tibet. Puluont. Mitt. Mus. Buyer. Stuutes, 4, 267-88. 1865. (Continuation of oppel 1863.) Ibid. 289-304. ROTHPLETZ, A. 1892. Die Perm-, Trias- und Jura-Formation auf Timor und Rotti im Indischen Archipel. Pulueontogruphicu, 39, Lief. 2, 57-106. SCHLOTHEiM, E. F. voii. 1820. Die Petrefuktenkunde uiifilirein jetzigen Stundpiinkte, diircli die Besclirei- bung seiner Sunvnlung versteinerter und fossiler Uber-reste des Tier- und Pfiunzenreiclies der Vorwelt erluutert. Gotha. SPATH, L. F. 1927-33. Revision of the Jurassic cephalopod fauna of Kachh (Cutch). Pulueont. Indicu, N.S., 9, Mem. 2. 1934. The Jurassic and Cretaceous ammonites and belemnites of the Attock district. Ibid. 20 (4). 1935. The Mesozoic pulueontology of British Soinulilund. 10, The Cephulopodu. Hargeisa. 1939. The cephalopoda of the Neocomian Belemnite Beds of the Salt Range. Pulueont. Indicu, N.S., 25, Mem. 1. SPITZ, A. 1914. A Lower Cretaceous fauna from the Himalayan Gieumal Sandstone. Rec. geol. Surv. Ind. 44, 197-224. STEFANiNi, G. 1933. Mollusclii del Giuralias della Somalia. Pulueontogr. itul. 31,Sux>\A. \-53. STEVENS, G. R. Ill prcss. The Jurassic and Cretaceous belemnites of New Zealand and a review of the Jurassic and Cretaceous belemnites of the Indo-Pacific region. Puleont. Bull. geol. Surv. N.Z. 36. STOLiczKA, F. 1866. Geological sections across the Himalayan Mountains from Wangtu Bridge on the River Sutlej to Sungdo on the Indus. Mem. geol. Surv. Indiu, 5(1). STOLLEY, E. 1929. Uber Ostindische Jura-Belemniten. Puluont. v. Timor, Lief. 16, Abh. 29. 1934. Beitrage zur Paliiontologie des Ostindischen Archipels. 1 1 : Zur Kenntnis des Jura und der Unterkreide von Misol. Stratigraphischer Teil. Neues Jb. Min. Geol. Puluont. 71 (B-B), Abt. B, 47-86. 698 PALAEONTOLOGY, VOLUME 6 STOLLEY, E. 1935. Zut Kenntnis des Jura und der Unterkreide von Misol. Palaontologischer Teil. Neites Jb. Mill. Geol. Paldoiit. 73 (B-B), Abt. B, 42-69. TATE, R. 1867. Secondary fossils from South Africa. Quart. J. geol. Soc. Loud. 12>, TEICHERT, c. 1940. Marine Jurassic of East Indian affinities at Broome, North-western Australia. J. roy. Soc. fV. Aust. 26, 103-19. UHLiG, V. 1903-10. The fauna of the Spiti Shales (Cephalopoda). Palaeoiit. Indica, Ser. 15, 4. VOGLER, j. 1941. Beitrage zur Geologie von Niederlandisch-Indien: Ober-Jura und Kreide von Misol. Palaeoiitographica, Suppl. 4, Abt. 4, Lief. 4, 243-93. WAAGEN, w. 1873-5. The Jurassic fauna of Kutch: the Cephalopoda. Palaeoiit. Indica, Ser. 9, 1. WANDEL, G. 1936. Beitrage zur Palaontologie des Ostindischen Archipels. 13 ; Beitrage zur Kenntnis der jurassischen Molluskenfauna von Misol, Ost. -Celebes, Buton, Seran und Jaindena. Neues Jb. Min. Geol. Paldont. 75 (B-B), Abt. B, 447-526. WANNER, J. 1931. Die stratigraphie van Nederlandsch Oost-Indie: Mesozoicum. Leid. geol. Meded. 5, 567-610 (Martin Feestbundel). 1956. Zur Stratigraphie von Portugiesisch Timor. Z. dtscli. geol. Ges. 108, 109-40. WEIR, J. 1929. Jurassic fossils from Jubaland, East Africa. Monogr. geol. Dep. Hunter. Miis. 3. 1930. Reports on geological collections from the coastlands of Kenya Colony. 4: Mesozoic brachiopoda and mollusca from Mombasa. Monogr. geol. Dep. Hunter. Mus. 4, 77-102. G. R. STEVENS New Zealand Geological Survey, P.O. Box 368, Lower Hutt, Manuscript received 23 January 1963 New Zealand NEW ZEALAND SPECIES OF THE PERMIAN BIVALVE ATOMODESMA BEYRICH by J. B. WATERHOUSE Abstract. Eight species of Atomodesma are described from New Zealand, ranging from basal to topmost Permian. The following new species are described; A. obliqualmu, A. woodi, and A. trabecidiim. The genus Atomodesma Beyrich is the most abundant and widespread fossil known from the New Zealand Permian. It occurs at almost every fossil locality in the Permian of the marginal and shelf facies of the Southland and Nelson synclines, and fragments of its shell are scattered throughout the otherwise barren greywackes of Canterbury and north Otago. Much of the widespread Wooded Peak and correlative limestones seems to be made up of Atomodesma prisms, and lenses of limestone at the base of the Stephens Formation are of a similar composition. Only two species have been named hitherto: Atomodesma trechmamii (Marwick 1934) and A. marwicki Waterhouse 1958. A fragment was also described by the writer in a note included in a paper by Bruce (1962). A further six species are recorded, and three named in the present work. Most of the Lower Permian species are found in western Southland in the Takitimu Mountains and foothills, and the Upper Permian species are found in Nelson, north-west Southland, and south Otago (text-fig. 1, table 1). The fossil localities are described in an appendix : each locality is numbered serially with the prefix GS. As mentioned in the text, some specimens are kept at the British Museum (Natural History), and some at Otago University, Dunedin. The remaining specimens are stored at the New Zealand Geological Survey, Lower Hutt, and are registered serially by num- bers with the prefix TM. SYSTEMATIC PALAEONTOLOGY Genus atomodesma Beyrich 1864 Type species. Atomodesma exaratiim Beyrich, subsequently designated by Wanner 1922, p. 63. Diagnosis. Large biconvex shells, with anterior prosogyrous or orthogyrous beaks, a weak byssal gape in many species, a small anterior ear in some, and generally a large posterior wing. Ornament of low concentric wrinkles, with or without radial plicae. Ligament area largely or entirely posterior, edentulous, striated by growth-lines. An umbonal septum present in both valves (text-fig. 2). Anterior adductor scar uncertain. Posterior adductor scar large, pallial line integripalliate, pitted. Shell prismatic. Per- mian of Africa, Salt Range, Karakorum, Russia, Australia, Indonesia, New Caledonia, and New Zealand. Discussion, {a) Subgenera. Several genera have been erected for forms that were not [Palaeontology, Vol. 6, Part 4, 1963, pp. 699-717, pi. 100-5.] 700 PALAEONTOLOGY, VOLUME 6 originally recognized as being closely allied to, or congeneric With, Atomodesma exaratwn. These include Aphanaia de Koninck 1877, from Australia, and Maitaia Marwick 1934, from New Zealand. In 1956 Dickins showed that Aphanaia is probably congeneric with TEXT-FIG. 1. Distribution of the Kazanian Wooded Peak Limestone and Tramway Sandstone of the Nelson Syncline and correlative beds of the Southland Syncline, with Atomodesma species. The width of the rock belts is exaggerated. Atonwdesma, and in 1958 the writer synonymized Maitaia with Atomodesma. The Rus- sian workers B. K. Likharev and U. N. Popov, on the other hand, subdivided Atomo- desma, erecting the genera Kolymia Likharev 1941 (see Likharev and Einor 1941), and lutomodesma Popov 1958. Dickins (1961u, p. 123) and the writer (1959) preferred to emphasize the closeness of these forms to Atomodesma exaratwn, and proposed that the genera be relegated to subgenera. Atomodesma {Atomodesma) has radial plicae, Kolymia Likharev has anterior ears and is not plicate, and Aphanaia de Koninck is said to be inequi- valve and lacks plicae or anterior ears. Maitaia Marwick and Intomodesma Popov are J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 701 close to Aphanaia in that they lack anterior ears and are not plicate. Although they appear to be equivalve, they were tentatively regarded as synonymous with Aphanaia. The writer now considers that plication is not of generic or even subgeneric significance. The anterior ear of Kolymia is probably an important feature and forms, with the rest of Table I. Occurrence of New Zealand species of Atomodesma the umbonal region, the crucial part of the shell. But the umbonal region has scarcely ever been figured or described in detail in Russian and Australian shells. As a result the type species of Kolymia, Intomodesma, and Aphanaia are so inadequately known that the writer proposes to refer all these related taxa to Atomodesma, at least until the vital structures are fully described. The umbonal region varies in different New Zealand species, and is consistent within each species. In several species the ligament area ends just in front of the beak, the umbonal septum is below the area, and is bounded in front by the anterior wall of the shell, which is in these forms normal to the commissure. This part is slightly but sharply depressed below the rest of the anterior face of the shell (the ‘depression’ in text-fig. 2). The depression is related to a narrow byssal gape, and passes upwards towards the liga- ment area. Such an arrangement is seen in New Zealand species from the Letham 702 PALAEONTOLOGY, VOLUME 6 Formation, from the widespread Upper Permian Wooded Peak limestone, and in the specimen of A. variabilis figured by Wanner (1940, text-fig. 1, p. 9), of which a mould was kindly sent to the writer by Prof. MacGillavry from the Geological Instituut of Amsterdam. The depression is feebly developed also in a species from the upper Tram- way Sandstone of New Zealand. Several species from New Zealand have a small anterior ear. This ear is of different structure in different species. In one from the Upper Permian Stephens Formation, Nel- son, the parts of the anterior walls which lie between the umbonal septa and commissure of the two valves are inclined forward, instead of lying normal to the commissure. The beak TEXT-FIG. 2. Dorsal anterior part of a right valve of Atomodesma, showing the ligament plate and umbonal septum. depressed area next to the commissure is well defined in this species, as in variabilis. In the oldest Atomodesma from New Zealand, the anterior walls between the septa lie flush with the tops of the septa and are scarcely depressed below the rest of the shell. In A. marwicki Waterhouse the entire septum lies in front of the beak, judging from an internal mould, and the anterior wall next to the septum does not seem to be depressed. Here then are three kinds of anterior ‘ear’, and it is not certain that the three species are more closely related to one another than to forms without ears. Yet another arrangement is present. A. trechmamii (Marwick) has no ear, and the depressed region in front of the ligament curves laterally well away from the commissure. {b) Characters of specific importance. The umbonal region is of foremost importance in circumscribing the specific characteristics of the New Zealand Atomodesma. The shape of the umbonal septum and the angle of the septum (measured between its anterior and dorsal edges, i.e. the inner edge of the anterior wall, and the inner edge of the hinge), the nature of the anterior margin immediately in front of the septum, and the anterior end of the ligament area, are consistent within each species and differ considerably from species to species. The definition of the posterior wing, the height and curvature of the umbo, and the shape of the shell, are of secondary importance, tending to vary laterally at the same horizon. The strength of the concentric wrinkles varies even more, as does the diameter of prisms in the shell. The striations of the ligament area and the muscle scars do not appear to differ significantly from species to species. (c) Structure of the shell. The shell structure comprises hexagonal or polygonal prisms normal to the shell surface, ranging in diameter from 0-025 to 0-05 mm., and of the same J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 703 diameter on each valve. Under crossed nicols the prisms extinguish for their complete length, although interrupted by weak and strong growth-lines. In a few specimens the prisms taper towards a strong growth junction, and a few interstitial trigonal prisms appear. But as a rule the prisms are parallel-sided. In 1958 the writer reported that a fibrous outer layer was present on a fragment from GS 6323. When forwarded to be sectioned, it shattered and the shell was destroyed. As a result the structure has not been verified, and the writer is inclined to consider that the layer may have been recrystallized, or composed of fine prisms, for no fibrous or lamellar structure has been seen in other sections. The shell where a muscle scar is situated was sectioned; no hypostracum is visible; the structure remains prismatic. However, Wanner (1922, p. 61, text-fig. 12) recorded a thin inner layer in A. exaratum. hwceramus, another prismatic shell common in New Zealand, has much thicker prisms, and its inner non-prismatic layer is frequently preserved. Atomodesma sp. A Plate 100, figs. 1, 3 Maitaia sp. Mutch 1957, p. 501. Atomodesma sp. nov. Waterhouse 1958/?, p. 605. Material. Two specimens are available, one a fragmentary internal mould, the other an external mould, both of the beak region. Other fragments and prisms are also present. Diagnosis. Beak weakly prosogyrous, ligament area almost completely open at anterior end, anterior part of septum apparently lying within a small anterior ear. Horizon and locality. Takitimu Group, Permian; GS 5885. Dimensions (in mm.). A right valve, from a growth-line. Specimen Length Height Width Umbonal angle TM 3786 17-5 19-5 6 75° Description. External. The specimens are so fragmentary that only a few details can be described. The beak is very weakly prosogyrous, and the anterior face is steep and incurved. The shell is smooth apart from very low growth wrinkles. Internal. The ligamental area is concave under the beak, and flatly concave posteriorly, and is weakly striated by longitudinal growth-lines and weak vertical lines. The area does not extend in front of the beak, and is almost completely open at its anterior end except for a very short ridge above the umbonal septum. Between the ligament and the anterior commissure is a concave umbonal septum, with the angle between its anterior and dorsal margins measuring about 50°. The anterior part of the septum seems to lie at the level of the commissure and to extend beyond the beak within a small ear, although this is possibly due to crushing. A narrow groove is present on the outer surface, parallel to and placed 1 mm. from the commissure. Shell structure. Prisms are long and about 0-03 mm. in diameter. Affinities. Maitaia trechmanni Marwick 1935, pi. 34, figs. 1-3, from the Upper Permian Tramway beds of Nelson, has a wider umbonal angle (70°), stronger concentric growth 704 PALAEONTOLOGY, VOLUME 6 wrinkles, and no anterior ear. The ligament area beneath the beak of trechmanni is more constricted by the umbonal septum, and the septum is set well below the commissure. The umbonal fragment of Atornodesma recorded by Waterhouse {in Bruce 1962, p. 166) from the Brook Street Volcanics is like the Takitimu form in its beak and anterior end of the ligament area, but apparently lacks an anterior ear, and has a septum set well below the commissure. Atornodesma rnanvicki Waterhouse 1958 Plate 100, figs. 2, 4 Maitaia sp. Marwick in Wood 1956, p. 38, fig. 18. Atomodesnm rnanvicki Waterhouse 1958u, p. 173, fig. 2a. Atornodesma (Kolymia) marwicki Waterhouse 1959, p. 260. Material. More specimens have been collected from the type locality of marwicki, but add little to the description of the species. Horizon and locality. Waipahi Group, Permian; GS 5082. Holotype. TM 2035. Diagnosis. Septum placed anterior to the beak, and largely anterior to the inner face of the anterior wall. Description. The species is described by Waterhouse (1958u). Atornodesma obliquatwn sp. nov. Plate 100, figs. 5-7 Material. A large left valve, a crushed specimen with valves conjoined, an umbonal fragment, and two small left valves. Specimens etched in acid to show internal details. It is not known whether the shells are equivalve or not. Horizon and locality. Productus Creek Group, Letham Formation, Permian; GS 7344. Holotype. TM 3366 (PI. 100, fig. 5). Paratypes. TM 3365, 3426, 3427. Diagnosis. Oblique shell, posterior wing small, ligament constricted by a ridge which passes upward from the anterior end of the umbonal septum, septum thin and plate-like, and the adjoining anterior wall between the septum and anterior commissure is depressed below the rest of the outer face of the anterior wall. Dimensions (in mm.), (left valves) Specimen Length Height Width Hinge length Umbonal angle a Angles b c TM 3426 30 38 ? 11 19 85° 55° 40° 90° TM 3366 90 100 21 27 60° 50° 40° 85° EXPLANATION OF PLATE 100 Figs. 1-3. Atornodesma sp. A. PVC cast of umbonal region, TM 3364, GS 5885, x2. Figs. 2, 4. Atornodesma marwicki Waterhouse. 2, Rubber latex cast of umbonal region. 4, Internal mould of holotype TM 2035, GS 5082, X 1. Figs. 5-7. Atornodesma obliquatiim sp. nov. 5, Internal mould of holotype, left valve TM 3366. 6, Internal mould and hinge of paratype TM 3426. 7, Rubber latex cast of umbonal region of TM 3427 ; the lower right outline of the figure defines the lower edge of the umbonal septum; GS 7344, X 1. Figs. 8-10. Atornodesma sp. aff. A. mitchelli M‘Coy. 8, PVC cast of crushed I’ight valve TM 3537, GS 6072. 9, PVC cast of TM 3535, GS 6072. 10, External rubber latex cast of TM 3540, GS 7350. Xl. Palaeontology, Vol. 6 PLATE 100 WATERHOUSE, Atomodesma J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 705 Description. External. These are the best-preserved specimens of Atomodesma from the Lower Permian of New Zealand. The valves are prosocline, with an anterior weakly prosogyrous umbo, steep anterior wall, and a short hinge. Between the septum and commissure the anterior wall is depressed like a flange. A small posterior wing is developed, with a narrow angle of only 25°. Irregular concentric wrinkles cover the shell. Internal. The ligament area is concave and striated, with longitudinal ridges in the larger specimens but not in the small TM 3426. The area is thickened moderately, and is con- stricted at the anterior margin by a low ridge that extends from the dorsal anterior end of the umbonal septum. The umbonal septum is thin and plate-like, with an angle close to 50°, and has a curved dorsal margin. The under- side of the septum and the ligament form one continuous surface. In the small specimen TM 3426 the septum is very small. The posterior adductor muscle scar is visible in TM 3366. It lies close to mid-height, below the posterior end of the hinge, and is bean- shaped, with a small lobe adjoining a large ventral lobe. No pallial line or anterior scar is visible. Shell structure. Prisms are 0-05 mm. in dia- meter, and the shell as a whole is normally 0-8 to 1 mm. thick. TEXT-FIG. 3. Right valve oi Atomodesma, show- ing how the angles are measured. The umbonal angle is measured between the anterior and dorsal margins of the first formed part of the shell. Angle a is measured between the dorsal margin of the shell behind the umbo and the umbonal ridge (as in Newell 1942, p. 22). Angle b is measured between the dorsal margin and the dorsal part of the posterior margin (Newell 1942). Angle c is measured between the dorsal margin and the anterior margin. Affinities. This species has more concentric wrinkles and a narrower posterior wing than in the Takitimu Atomodesma. The part of the anterior wall between the septum and commissure is depressed and normal to the com- missure, whereas it is apparently almost continuous with the rest of the exterior and points forward like an ear in the Takitimu specimen. Also the ligament of A. obliquatum is more constricted under the umbo. A. trechmanni (Marwick) from the Upper Permian of New Zealand has a larger pos- terior wing, a produced rather than depressed anterior commissure, with subterminal beaks, and a larger, more bulbous ridge protruding above the umbonal septum into the ligament area. Inoceramus mitchelli M‘Coy 1847 from Eastern Austraha has similar ornament and a similar narrow posterior wing, but is less oblique and has a longer hinge. Perhaps the closest specimens are those figured as A. cf. timorense Wanner by Dickins (1963, pi. 9, figs. 8, 10-12) from the Artinskian Callytharra Formation and Nura Nura Member of Western Australia. The Australian shells have a high tumid beak and prominent posterior wing, as in Wanner’s specimens from the Artinskian beds of Niloelet, Timor. Unfortunately the details of the umbonal region of timorense are not available for comparison. A. timorense has a higher beak and greater inflation than the associated 706 PALAEONTOLOGY, VOLUME 6 species A. mytiloides Beyrich 1864, from the Bitauni beds of Timor, and the Callytharra Formation, Poole Sandstone, Fossil Cliff Formation, Noonkanbah Formation, and Byro Formation of Western Australia, but the umbonal structures must first be known in order to determine the significance of these differences : they may be only subspecific or varietal. Atomodesma sp. aflf. A. mitchelli M‘Coy 1847 Plate 100, figs. 8-10; Plate 101, fig, 1 ; Plate 105, figs. 8, 9 Material. Two right valves and a left valve showing the hinge, a damaged specimen with valves con- joined, and large and small fragments from GS 6072. Natural external moulds of two right valves from GS 7350 and fragments from GS 3616 come from the same formation and so might be conspecific. All specimens incomplete and crushed. Horizon and localities. Mangarewa Formation, Permian; GS 6072, ? 3616, ? 7350. Diagnosis. Small well-rounded shell with terminal weakly prosogyrous beak and a very narrow posterior wing, concentric wrinkles usually well developed. Dimensions (in mm.), estimated from broken specimens. GS 6072. Hinge Umbonal Angles Specimen Length Height Width length angle a b C Right valve TM 3535 45 40 15 31 00 0 o c/1 1 o ' 30° 105° Left valve TM 3536 45 38 ? 13 27 — 45° O O 90° Description. External. The shells are small and moderately inflated. with a subquadrate outline, and weakly prosogyrous terminal beaks. The anterior wall is unusually wide. It is flattened on the beak, and below the hinge lies normal to the commissure. In each specimen a furrow 2 or 3 mm. wide arises at the commissure just below the umbonal septum, and passes dorsally almost to the anterior end of the ligament. The furrow might be due to crushing but looks like a byssal groove, marked as it is by several wrinkles. Only a very narrow posterior wing is present. Apart from the first formed part of the shell, which is almost smooth, the specimens from GS 6072 have strong concentric growth rugae which are occasionally overturned on the ventral side. One to three lower concentric wrinkles lie on each prominent ruga. The two valves from GS 7350 have only low wrinkles. Internal. The face of the ligament area is inclined in section at 20° to 30° dorsally away from the commissure. The area is not supported by heavy thickening. Under the beak it is strongly concave and is constricted by a low convexity which lies above the EXPLANATION OF PLATE 101 Fig. 1. Atomodesma sp. aff. A. mitchelli M‘Coy. Part of PVC cast of umbonal region, TM 3535, GS 6072, X 2. Figs. 2-5. Atomodesma woodi sp. nov. 2, Rubber cast showing long umbonal septum, OU 2436, GS 4652, X 1. 3, Internal mould of right valve TM 2037 from GS 4586, showing posterior muscle scar, X 1. 4, Holotype, PVC cast, TM 3645, GS 4652, X 1. 5, Rubber latex cast of umbonal region with the valves conjoined and opened out to show the umbonal septa separated by the ligament area, TM 3643, GS 4652, X2. Palaeontology, Vol. 6 PLATE 1 01 WATERHOUSE, Atomodesma J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 707 umbonal septum; posteriorly it widens and becomes flatter. About three longitudinal grooves, separated by sharp ridges, are present on the area, the dorsal grooves passing back on to the body of the shell and the inner one commencing at the convexity. The area is also marked by fine longitudinal ridges (3 or 4 in 1 mm.) and by even finer vertical striae, corresponding with the prismatic structure of the shell. The umbonal septum is set 2 mm. below the commissure, and is prolonged posteriorly for half the length of the hinge as a narrow ledge. In TM 3537 it is acutely angular at the apex under the beak, with an angle of 45° between its anterior and dorsal margins. In the other two specimens, TM 3535 and 3536, the apex is rounded, and the angle of the sep- tum as a whole is only 30° or 40°. Muscle scars and pallial line are uncertain. Shell structure. The shell away from the umbonal region is up to 1-5 mm. thick. Thin sections show that the prisms are usually 0-025 mm. in diameter in specimens from both GS 6072 and 3616. In some specimens from GS 3616 the prisms are 0-05 mm. in diameter. The prisms are usually hexagonal, or almost so, in cross-section, but are sometimes square. Affinities. The Mangarewa shells lack the anterior ear of the Takitimu Atomodesma sp. A, and also have a more constricted ligament area under the beak, and slightly more prosogyrous beaks. The Letham species A. obliquatum sp. nov. is more oblique and has a better defined posterior wing, and a more thickened ligament area. The anterior wall of A. obliquatum is depressed close to the commissure to form a narrow flange, and lacks any signs of a broad byssal groove. A. trechmaimi (Marwick) has a much larger posterior wing and is a larger, less inflated shell, with subterminal beaks. The umbonal septum and anterior wall are moderately similar, but the anterior wall of trechmaimi bulges forward, and the ligament area is thickened more posteriorly and is less inclined from the commissure. Closest species in shape is Inoceramus mitchelH M‘Coy 1847 from beds of approxi- mately the same age at Glendon and Wollongong, New South Wales. Plaster duplicates of the holotype and two paratypes have been kindly sent to the writer from the Sedg- wick Museum by Mr. A. G. Brighton, and other specimens have been kindly loaned from the Australian Museum by Mr. H. O. Fletcher. The specimens vary in shape and ornament. Closest to the New Zealand specimens are the holotype, and M‘Coy’s para- type E 10733. These are subquadrate in outline with a wide anterior wall. Strong rugae and a very narrow posterior wing occur in the holotype, as in the New Zealand specimens. The second paratype, E 10732, is a higher shell, with slightly more regular wrinkles, but is otherwise close. The specimens sent to the writer from the Australian Museum have wide anterior walls but vary a little in shape and ornament. Internal details are preserved in specimen E 48885, figured by Dickins 1963, pi. 9, figs. 14, 15. The ligament area is narrow and con- cave, as in the small shell TM 3540 from GS 7350, and the umbonal septum is small, and a little thicker than in the New Zealand forms. A septum is also visible in F 2170, figured by Dickins 1963, pi. 9, figs. 16, 17. (Dickins suggested that an anterior ear is possibly present, but this is very doubtful.) In F 48885 the inner side of the anterior wall below the septum bulges inwards where the supposed byssal groove is situated in the New 708 PALAEONTOLOGY, VOLUME 6 Zealand specimens. Unfortunately the external mould is not available, so that further details, necessary to estabUsh or disprove the identity of the Austrahan and New Zealand shells, are not available: similarity in shape and ornament might prove to be misleading. Alomodesma woodi sp. nov. Plate 101, figs. 2-5; Plate 102, figs. 1, 4; Plate 105, fig. 12 Maitaia trechmanni, non Marwick, Wood 1956, p. 44, figs. 23a, d. Material. A right valve, two left valves, and three specimens with valves conjoined from GS 4652, four right valves, two left valves, and a specimen with valves conjoined from GS 4586. A left valve from GS 4562 is kept at Otago University. Best preserved of New Zealand Atomodesma, showing muscle scars and colour bands as well as other details. About thirty single valves from GS 5077 might be con- specific but do not show the hinge. Horizon and localities. AG 4 Limestone, Arthurton Group, Permian; GS 4652, 4586, ? 5077. Holotvpe. TM 3645, GS 4652 (PI. 101, fig. 4). Paratvpes. TM 3641-4, 3646, OU 2436, GS 4652; TM 2037, 3636-40, GS 4586. Name. After Mr. B. L. Wood, N.Z. Geological Survey, Dunedin. Diagnosis. Large shell with tiny anterior ear in small specimens, posterior wing usually scarcely defined and wrinkles usually very low. Umbonal septum large, with a wide angle of 80° to 90° ; anterior edge of septum almost straight. Dimensions (in mm.). Specimen TM Length Height Width L. valve R. valve Hinge length Umbonal angle a Angle b c GS 4586 3638 59 59 — 8 33 70° 48° 40° 80° 2037 85 ?85 ?13 — — 70° 60° 20° 100° 3637 100 88 — 1L5 — 85° 50° 40° 90° 3639 103 96 11 — — 80° 50° 30° 100° GS 4652 3644 gr 35 40 4-5 — . ■ — 90° 55° 40° 110° 3645 115 no — 16 50 70° 50° 20° 85° GS 5077 3674 gr 11-5 9-5 4 — 8 55° 50° 55° 65° 3675 12-5 12 4-5 — 9 60° 50° 45° 75° 3671 17 22 8 — 11 60° 50° 40° 80° 3672 17 24 — 5-5 12 65° 50° 55° 70° 3676 18 17-5 6 — ?12 60° 45° 40° — 3673 20 24 10 — 9-5 50° 45° 50° 0 O 00 gr — measured from growth-line. Description. External. The shells are little inflated, probably equivalve, and large, the biggest specimen TM 3646 measuring more than 12 cm. in length. The beak is terminal, and weakly prosogyrous. Maximum height lies close to the anterior margin of the shell, which is gently convex in outline. A very low anterior ear or ledge projects forward for 1 or 2 mm. from the commissure in front of the umbonal septum in specimens up to 80 mm. long. Adjoining the ledge is a sharp depression, which continues from the anterior end of the ligament. As the shells increase in size, additional layers of shell increase the J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 709 width of the ledge, but do not project beyond the first layers, so that the ear-like effect is lost. A posterior wing is usually not developed. The specimens vary in shape. Some, such as TM 3637, are very high; others, such as TM 2037, are more elongated, becoming so after early ontogeny. Some of the shells are smooth, and others have low, well-spaced concentric wrinkles. Internal. The ligament area is supported on a sturdy shelf, which is not greatly thickened. It is concave under the beak and lies roughly in the plane of the commissure until beyond mid-length and then turns outward through 90°. Striations and grooves occur as in other species and persist anteriorly beyond the ligament on to the depressed part of the anterior face. The ligament area is not constricted under the beak by the top of the umbonal septum, recalling the areas of the Takitimu and Brook Street specimens. The septum is very large and thin, with its anterior margin straight rather than curved. The angle of the septum varies, measuring 95° in TM 3644, 90° in TM 3643, and about 70° in TM 3638 and 3645. It extends well back from the umbo, particularly in the speci- men at Otago University. The posterior adductor muscle scar is well defined, bean-shaped in outline, placed well below and behind the posterior extremity of the hinge, and marked by low growth lines. A pitted pallial line extends forward from the scar. Nearer the beak a second scar of similar outline is visible in TM 2037 and 3637 from GS 4586, and another pitted line can be discerned in TM 3637, possibly representing gill suspensories. Shell structure. Radial colour bands of light and dark shell are visible on decorticated specimens from GS 4652 and 4586. The dark bands coincide with the pallial muscle scars, at least in TM 3644. A thin section of a specimen from GS 4586 shows thin prisms only 0-025 mm. across in some fragments and up to 0-03 mm. in diameter in others. Prisms are 0-04 mm. in diameter in fragments from GS 4652. Prisms from the correlative Wooded Peak Lime- stone at Nelson, and D’Urville Island and the Waituna Stream boulder are only 0-025 mm. in diameter, but are 0-05 mm. in diameter in the Wooded Peak Limestone at Pahakorea, just south of D’Urville Island. Variation. Most of the specimens from GS 5077 are tiny mytiliform shells, close to the holotype in shape, or narrower, with a very low b angle between hinge and posterior margin, and narrow c angle. The shells are almost smooth. A few are elongate, with a posterior wing and low wrinkles. Absence of hinge details makes identification uncertain. A ffinities. As noted by Waterhouse (1958u, p. 172) this species is distinguished from A. trechmanni (Marwick) by its wider umbonal septum. The beaks of the new species are terminal, a tiny anterior ‘ear’ rather than a large bulge is present in small specimens, the posterior wing is poorly developed, and the concentric wrinkles are fewer than in A. trechmanni. The tiny projection of the anterior wall under the beak recalls that of Atomodesma sp. A, from GS 5885 of the Takitimu Group. But in the Takitimu species the projection is larger, contains the anterior part of the umbonal septum, and is not distinctly sepa- rated from the rest of the exterior by a deep depression. Also, the umbonal septum is narrower in the Takitimu form, and the beak less prosogyrous. A. obliquatum from the Letham Formation is close in details of the umbonal region, having a similar sharp depression in front of the ligament area. But the anterior wall 3 A C 1713 710 PALAEONTOLOGY, VOLUME 6 next to the depression is normal to the commissure in A. obliquatwn, and the depression is narrower. The Letham species is further distinguished from A. woodi by the narrower, shorter umbonal septum, the presence of a narrow posterior wing, and better developed concentric ornament. Also the ligament is constricted under the beak. Of overseas species the closest appears to be Aphanaia hoibabemis Reed (1935, pi. 5), a species distinguished by its strong rugae, but similar to A. woodi in shape and in lack- ing a posterior wing. A. haibabensis comes from the Upper Dwyka Beds of South-west Africa, considered to be Lower Permian in age (Dickins 1961h). Atomodesma trechmanni (Marwick 1934) Plate 102, figs. 2, 3, 5-7; Plate 103, figs. 1-4 Aphanaia sp. cf. mitchelli Trechmann 1917, p. 56, pi. 4, figs. 1-4, 8. Maitaia trechmanni Marwick 1934, p. 948. Maitaia trechmanni Marwick 1935, p. 295, pi. 34, figs. 1-3. Atomodesma trechmanni Waterhouse 1958o, p. 171, fig. 2b. Atomodesma (Aphanaia) trechmanni Waterhouse 1959, p. 259. Material. The internal mould of a large right valve (holotype), and about twelve other specimens, from GS 143, preserved chiefly as natural moulds, with a wide range in size. A fine collection from this locality is kept at the British Museum (Natural History). Fragments of the hinge and more or less broken specimens come from GS 7410. Crushed specimens from GS 741 1 might be conspecific. A speci- men with conjoined valves from GS 7638, and other fragments. Several incomplete single valves and umbonal fragments from GS 2946. Fragments of two valves from GS 7664. A right valve and two left valves, and fragments from GS 5185. Horizons and localities. Lower Tramway Sandstone, Permian; GS 143, 2946, 7410, ? 7411, 7638. Lower Annear Sandstone, Permian; GS ? 5185, 7664. Holotype. TM 2033, GS 143 (PI. 103, figs. 1, 3). Paratypes. TM 2304, 3678-84, 3712. Diagnosis. Moderately high, little inflated shell, with subterminal beak, a wide anterior EXPLANATION OF PLATE 102 Figs. 1, 4. Atomodesma woodi sp. nov. 1, Internal mould TM 3637, GS 4586, showing posterior muscle scar and obscure pallial line, with ? gill suspensories nearer the beak, X 1. 4, TM 3641, GS 4652, with white and black colour bands, X 1 . Figs. 2, 3, 5-7. Atomodesma trechmanni (Marwick). 2, Internal mould and hinge of TM 3712, GS 143, X 1. 3, Internal mould of TM 3682 with valves conjoined but displaced, and slightly crushed, GS 143, X 2. 5, Part of internal mould TM 3684 with valves conjoined and crushed, GS 143, X 2. 6, Crushed internal mould of TM 3690, GS 2946, X 1. 7, Right valve of TM 3678, with anterior margin crushed to simulate an ear; contrast the small posterior wing and smooth shell with that in specimens from the same locality shown in figs. 2, 3, and 5; GS 143, X 1. EXPLANATION OF PLATE 103 Figs. 1-4. Atomodesma trechmanni (Marwick). 1, 3, Mould and cast of umbonal region of holotype, TM 2033, GS 143, X 1. 2, InternalmouldofTM 3705, GS 5185, xl. 4, Rubber latex cast of umbonal region, TM 3685, GS 7410, x2. Figs. 5-7. Atomodesma trabeculum sp. nov. 5, Holotype, rubber latex cast, showing thick hinge, TM 3708, GS 7637, x2. 6, Mould of holotype, X 1. 7, Rubber latex cast of umbonal region, showing septum unusually well developed for the species, TM 3703, GS 7637, x2. Palaeontology, Vol. 6 PLATE 102 i'MM WATERHOUSE, Atoinodesma Palaeontology, Vol. 6 PLATE 103 WATERHOUSE, Atomodesma J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 711 wall bulging beyond the beak and large posterior wing; umbonal septum with curved anterior margin. Dimensions (in mm). Specimen TM Length Height Width L. valve R. valve Hinge length Umbonal angle a Angle b C GS 143 3682 gr 6 7-5 — 1-8 3-5 ? 90° 65° 70° 95° 3684 gr 8 10 2-2 5 70° 60° 50° 105° 3680 28 32 — 7-5 17-5 70° 65° 80° 100° 3678 63 59 — 15 ?40 50° 45° 35° 70° GS 7638 3689 gr 23-5 23 — 8-5 17-5 ?80° 55° KM o o 60° gr — measured from growth-line. Description. External. The shell is large, usually slightly higher than long, and little inflated, with subterminal weakly prosogyrous beaks, and a large moderately defined posterior wing. Except in the small TM 3682 no byssal gape is visible, perhaps because of imperfect preservation. There is a small gap at the top front of the umbonal septum in the holotype, and in TM 3685 and other specimens. Some specimens appear to be inequivalve, but in all such instances one or other of the valves has been displaced dor- sally (making it seem to be higher than the opposite valve), and has been rucked up by distortion, exaggerating its width. The anterior wall is characteristically wide, and usually bulges forward below the beaks at 120° to 140° from the hinge. The depression that continues on from the ligament area curves well away from the commissure, unlike that of other species. Ornament comprises low wrinkles, except in a few topotypes which are smooth. Internal. The ligament area is described in detail by Waterhouse (1958u, p. 171). The umbonal septum is particularly characteristic. Its upper (dorsal anterior) end interrupts the growth-lines of the ligament and bulges a little beyond the beak. The anterior margin of the septum is set below the commissure, and curves through 40°, so that the sep- tum as a whole has a narrow angle of only 45-50° (see Trechmann 1917, pi. 4, fig. 4). The shape of the septum differs in some of the specimens at the British Museum (Trech- mann 1917, pi. 4, fig. 1), but this is due to crushing. The posterior muscle adductor scar is bean-shaped, as described in Waterhouse { 1958r/). Shell structure. The shell is generally about 1 mm. thick away from the beak region, and in specimens from Hackett Stream is made up of coarse fibres 0-04 to 0-05 mm. in diameter. Variation. The small specimen from GS 7638 is slightly more inflated and elongate than usual, but has the subterminal beaks and thin septum typical of the species. The septum has a wide angle of 70°, and a straight anterior edge as in immature topotypes. Crushed mature specimens from GS 741 1 also appear to have a straight anterior edge to the umbonal septum. The umbonal region in fragments from GS 2946, D’Urville Island, agrees with that of the type trechmanni. Strong concentric wrinkles cover the valves, as in only a few of the specimens of trechmanni from the type locality. The valve TM 3709 from GS 7664 is incomplete, but has a hinge and fragmentary 712 PALAEONTOLOGY, VOLUME 6 septum like that of trechmami, and a second specimen from the same locality has a similar beak, ligament area, and upper part of the septum. Two specimens from GS 5185 have close-set, well-defined wrinkles and a large pos- terior wing. A third, TM 3708, is high with a short wide anterior wall and no posterior wing. The beaks are weakly prosogyrous, and the umbonal septum visible in TM 3706 is typical of the species. A ffinities. Resemblances to other species are discussed in the other descriptions. Atomodesnm trabeculum sp. nov. Plate 103, figs. 5-7; Plate 104, figs. 1-6; Plate 105, figs. 1-5, 10, II New genus of pelecypod, Marwick 1925, p. 362, figs. 6, 7. Maitaia trechmcmui Marwick 1935, p. 296. Atomodesma marwicki (part) Waterhouse 1958u, p. 174. Atomodesma (Kolymia) marwicki Waterhouse 1959, p. 260. Material. Four fragmentary specimens with valves conjoined, a left valve and other fragments from GS 7637, three specimens with valves conjoined and a left valve from GS 3757, two right valves, a left valve and a specimen with valves conjoined from GS 5926, and three specimens with valves conjoined, two left valves and other fragments from GS 9158. Hinge well shown, but muscle scars and pallial line unknown. Two specimens with valves conjoined from GS 7633 and fragments from GS 7666 might be conspecific. Natural internal and a few external moulds of twelve small and two large speeimens with valves conjoined and twelve single valves from GS 1456; another ten large and small internal moulds from the same locality, kept at Otago University. Thirty fragmentary internal and external moulds of small specimens from GS 5060. Horizon and localities. Upper Tramway Sandstone, Permian; GS 5926, 7633, 7637, and derived boulders at GS 3757 and 9158. Upper Annear Sandstone, Permian; ? GS 7666. Arthurton Group, Permian; GS 1456. Kuriwao Group, Pine Bush Formation, Permian; GS 5060. Holotype. TM 3708, GS 7637 (PI. 103, figs. 5, 6). Paratvpes. TM 3692-5, GS 9198; 2036, 3696-7, GS 3757; 3698-9, GS 5926; 3700-4, GS 7637. Diagnosis. High oblique shell with prosogyrous beak, well-defined posterior wing, orna- ment of moderately strong wrinkles. No anterior ear. Umbonal septum very small, and heavily thickened under the umbonal cavity; ligament area supported by heavily thickened ridge. EXPLANATION OF PLATE 104 Figs. \-6. Atomodesma trabeculiimsp. now. Allfigs.X 1. 1, Internal mould TM 3694,GS9158. 2, Internal mould TM 3696, slightly turned to side, GS 3757. 3, Internal mould TM 3698, GS 5926. 4, Internal mould TM 3687, GS 7633. 5, Rubber latex external cast, TM 3694, GS 9158. 6, Lateral anterior view of internal mould, TM 3697, GS 3757; the umbonal septum simulates an ear. EXPLANATION OF PLATE 105 Figs. 1-5. Atomodesma trabeculum sp. nov. X 1. 1, Internal mould TM 3662, GS 1456. 2, TM 2036, GS 3757, wrongly identified asT. marwicki in 1958. 3, 4, Anterior and lateral views of internal mould TM 3657, GS 1456. 5, Internal mould TM 3663, GS 5060. Figs. 6, 7. Atomodesma sp. B. X 2. 6, Anterior view of rubber latex cast of exterior, TM 3718, GS 3758. 7, Rubber latex cast of ligament area and septum of TM 3720, GS 3758. Figs. 8-12. Thin sections of Atomodesma. x40. 8, 9, Atomodesma sp. aflf. A. mitchelli, GS 6072 (oblique) and 3616. 10, 11, A. trabeculum sp. nov., end-on and lateral views in umbonal region, GS 1456. 12, A. woodi sp. nov., GS 4652. Palaeontology, Vol. 6 PLATE 104 WATERHOUSE, Atomodesma Palaeontology, Vol. 6 PLATE 105 WATERHOUSE, Atomodesma J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 713 Dimensions (in mm.) Specimen TM Length Height Width L. valve R. valve Hinge length Umbonal angle a Angle b c GS 9198 3692 36 42 ?15 — 25 o^ o fyi o o 60° o O OO GS 3757 2036 gr 26-5 34 12 — 19-5 o O SO 55° O O o O o GS 5926 3698 gr 29 32-5 — 9 20 o o kO ?60° o O '-J o o GS 7637 3702 gr 32 25-5 11 — 20-5 o o 40° 60° 80° 3700 gr 38 49 11 — 28-5 — 60° 55° 75° GS 7633 3687 14 20-3 — 4-5 14 60° 65° 80° 90° 3688 18-5 26-5 — 9 17 , 55° 55° 70° 70° GS 1456 3660 16-5 15-5 4 5 8 55° 35° 30° ?75° 3653 21 16 4 4 12 55° 35° 30° 70° 3658 23 22-5 5-5 6-2 11 70° 55° 35° 90° 3650 25-5 26-5 8-5 8 — 70° 50° — 70° 3657 27 29 16-3 19-4 11 70° 50° 35° 100° 3649 29 30-5 7-5 6 20 60° 50° 30° 75° 3654 19-5 24 — 7-5 9-5 90° 55° 50° 110° 3661 27 31-5 8 — 14 80° 50° 40° 100° 3662 29-5 40-5 10-3 — 17-5 75° ?60° 50° 95° 3715 65 57 — 23 28 70° 60° 40° 100° GS 5060 3664 22-5 20-3 13 — ?16 — 45° 65° — 3667 23-5 22 3 5 10-5 75° 55° 33° o O OO 3663 24 27-5 ?6 — 10-5 50° 55° ?50° — gr — measured from growth-line. Description. External. The shells are high, apparently equivalve, and typically strongly prosocline in outline, well inflated near the anterior margin, and with a well-defined posterior wing. Beaks are narrow, strongly prosogyrous for the genus, and terminal. The anterior wall is wide and almost normal to the commissure. A weak depression passes ventrally from the anterior end of the ligament area, close to the commissure, as in A. woodi and A. obliquatum, but not so sharply defined as in these species. The anterior wall between the depression and commissure is raised to form a low ridge. A narrow byssal gape, wider in the left valve, is visible in TM 3697 from GS 3757, and in TM 3695 from GS 9198. Some of the specimens from GS 9158 have only a few strong concentric wrinkles, but as a rule the shell is covered by strong wrinkles, sometimes imbricate. Internal. The ligament area is striated and grooved as in other species, and is constricted under the umbo by the top of the umbonal septum, as in ^4. trechmanni. The hinge- plate under the ligament is heavily thickened into a roller-like ridge, which continues under the umbonal septum. The anterior edge of the septum is also supported in this fashion, so that the septum is very sturdy. The anterior and dorsal edges of the septum are comparatively straight, and diverge at 60° to 65° in the holotype and associated specimens, and at 60° to 70° in specimens from GS 3757. The septum itself is usually 714 PALAEONTOLOGY, VOLUME 6 small, and either deeply concave or even V-shaped in profile, as in the holotype or in TM 2036 from GS 3757. A few associated specimens (such as TM 3697) have a better formed septum set in front of the inner face of the anterior wall, as though an ear is developed. But the anterior wall is extremely thick, and the outer surface lies flush with the beaks. Shell structure. Prisms are 0-03 mm. in diameter in specimens from GS 3757, and 0-05 mm. in a specimen from GS 1456. In cross-section the prisms are irregularly hexagonal. The shell is about 1-5 mm. thick. Variation. The identity of the D’Urville specimens from GS 7633 is uncertain because the hinge is somewhat obscure. They agree in shape, ornament, and prosogyrous beaks with A. trabeculum and seem to have a similar thick but shorter hinge. Obscure and distorted fragments from GS 7666 in the Hollyford Valley have strong wrinkles and a posterior wing. One fragment has the umbonal septum set in front of the inner face of the anterior wall, as in TM 3697 from GS 3757, and on this basis is identified with trabeculum. The Clinton specimens are smaller than shells from Nelson, and include a number of juveniles. A few are obliquely elongated, and all tend to have slightly lower wrinkles and an anterior wall more sharply defined from the rest of the shell than in specimens of trabeculum from Nelson, suggesting some geographic dilferentiation. The beaks are prosogyrous, and the posterior wing, thick ligament plate, and small septum are developed as in typical trabeculum. The angle of the umbonal septum between its anterior and dorsal margins varies from 60° to 70°. The specimens from GS 5060 are also small. They vary in outline, and not all have a well-developed posterior wing. However, the prosogyrous beaks, anterior wall sharply defined from the rest of the outer surface, close concentric wrinkling, and strong ridge- like development of the ligament area indicate affinity with A. trabeculum. The umbonal septum is poorly preserved but a few specimens such as TM 3713 seem to have the small septum typical of trabeculum. Affinities. The position of the umbonal septum in front of the anterior wall recalls that of A. marwicki Waterhouse 1958 from the Waipahi Group, and the writer (1958a) erroneously identified a specimen from GS 3757 with A. marwicki. However, the two forms differ in other details. The hinge of marwicki is not thickened as in the new species, concentric ornament is not so well developed, and the umbonal septum is better defined. These differences also apply in comparison with A. trechmanni, which is further dis- tinguished by its less prosogyrous, less terminal beak, lower inflation, and less defined posterior wing. The specimens of A. trechmanni from Hollyford are closer in ornament and the posterior wing, but agree with the trechmanni from Nelson in other details. The Upper Permian specimen figured as Liebea sp. aflf. hausmanni by Diener (1903, pi. 5, figs. 18a, b) and referred to Aphanaia by Reed (1944) has strong concentric wrinkles and a high but less prosogyrous beak, and less developed posterior wing. Kolymia irregularis Popov (1958, pi. 3, fig. 2) is also moderately close in appearance. Atomodesma sp. B Plate 105, figs. 6, 7 Material. Umbonal fragments of a right valve and a specimen with valves conjoined, both showing the J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA 715 hinge and anterior ear, an internal mould of the posterior part of a left valve, and an external mould of the posterior part of a specimen with valves conjoined. Horizon and locality. Boulder in Stephens conglomerate, Permian; GS 3758. Diagnosis. Shell with moderately prosogyrous beak and well-defined posterior wing, umbonal septum set within anterior ear, and pouting anterior margins. Description. External. The shells are of moderate size and apparently equivalve, with weakly prosogyrous subterminal beaks and an umbonal angle of 80°. The anterior wall is wide, and excavated near the commissure by a deep depression which arises just below the anterior end of the ligament area. The anterior wall between the septum and commissure forms a low anterior ear. A possible narrow byssal gape lies below the septum, and below it the anterior edges point forward to form a second prominence. A posterior wing is well developed and the shell is ornamented by a few low wrinkles well away from the beak. Internal. The ligament area is concave and is constricted below the beak by the umbonal septum. The septum is well defined, and lies well below the commissure, with an angle of 55°. The hinge is thickened along the posterior edge of the septum. Shell structure. The shell is about 1 to 1-2 mm. thick, made up of prisms about 0-03 mm. in diameter. Affinities. The beaks are less prosogyrous and the hinge is a little less thickened than in A. trabeculum, and a small anterior ear is developed. In the details of the posterior wing and ornament the two are close, suggesting that this species might be a descendant of trabeculum. The umbonal septum lies farther in front of the beaks in A. marwicki. EVOLUTION Most of the important changes seen in New Zealand specimens of Atomodesma occur in the anterior wall and umbonal septum, and probably involved minor changes in the mode of attachment. Another series of changes involved the posterior wing: the wing gradually decreased in size until it is absent in the lower Kazanian A. woodi, and then, more abruptly, increased again in size in later species. The two outstanding New Zealand species are woodi and trabeculum, and the question whether these evolved in situ from endemic New Zealand faunas or whether they invaded the area de novo cannot be answered until overseas species are fully described. Of the two, woodi fits better into an evolutionary series than does trabeculum. A. woodi is characterized by its long septum, and absence of the posterior wing. The Atomodesma in immediately underlying beds has a very small posterior wing, and a relatively long septum. The species in beds above those containing woodi, namely trechmanni, also has a well-formed septum, though the posterior wing is very well developed. The following species, trabe- culum, is characterized by its thick hinge and small septum. It is close to trechmanni in shape and ornament, but the hinge of trechmanni is little if at all thicker than that of earlier species. The latest species known, Atomodesma sp. B, has a moderately thick hinge, and so may have descended from trabeculum. Acknowledgements. I wish to thank Mr. A. G. Brighton, Curator, Sedgwick Museum, for kindly send- ing plaster duplicates of Inoceramus mitchelli M‘Coy, and Mr. H. O. Fletcher for kindly lending 716 PALAEONTOLOGY, VOLUME 6 specimens of mitcheUi from the Australian Museum. I am also grateful to Prof. H. J. MacGillavry, Geological Instituut, Amsterdam, for a mould of A. variabilis, and to Prof. D. C. Coombs for the loan of New Zealand specimens of Atomodesma from Otago University. Dr. J. M. Dickins kindly sent a manuscript on Australian Atomodesma and discussed the genus with me by correspondence. Photographs are by Mr. S. N. Beatus, thin sections were prepared by Mr. R. Freeman, and the text- figure drafted under the guidance of Mr. C. T. Webb, New Zealand Geological Survey. GS no. 143 1456 2946 3616 3757 3758 4586 4652 5060 5082 5077 5185 5885 5926 6072 7344 7350 7410 7411 7633 7637 7638 7664 7666 9158 Localities Description Dun Mountain Tramway, Wooded Peak, Nelson Ridge 1 mile NE of Clinton railway engine sheds. N.Z. M.S. SI 70, grid ref. 190316 (1943 ed.) Te Horo Stream gravel, Manawakupakupa, D’Urville Island. SIO, 125750 (1945 ed.) ‘Productus Creek’, 118 chains at 220° from Letham Trig. SI 59, 994748 (1945 ed.) Boulders in Whangamoa River, 3|: miles towards Nelson from junc- tion with Collins River. SI 5, 822396 (1944 ed.) Boulders in fallen blocks of conglomerate on beach If miles south of Whangamoa Head. SI 5, 845465 Quarry on W bank of Waipahi River, 144 chains at 43-5° from Trig. V, 887 ft. S170, 026393 (1943 ed.) Less-weathered material from the same quarry Hillside quarry, 90 chains NNW of Pine Bush School. SI 82, 722977 (1942 ed.) Hilltop boulders at The Cone. SI 70, 064457 Small exposure on N bank of stream, 64 chains at 124° from Trig. DD. SI 70, 047419 East shore of Lake Alabaster. SI 13 Scree at E side of the upper portal of Wairaki Gorge. SI 59, 860690 Bluff above stream flowing from Mt. Duppa to road -I mile below foot of Whangamoa Saddle. S15, 817367 Ridge below limestone scarp, 10 chains NNE of junction of Letham Burn and first east branch. SI 59, 989763 South edge of first small N tributary of the first E tributary of Letham Burn. SI 59, 990759 Outcrop on W side of the ridge with GS6072. S159, 993755 5 chains W of the deep cut in Dun Mountain Tramway. S20, 678228 (1956 ed.) South bank of Hackett Stream, where stream turns sharply NE. S20, 573152 Wells Bay, Port Hardy, D’Urville Island. SIO, 235878 (1945 ed.) Prominent knob on leading W ridge of Mt. Duppa, and scree below. S15, 816364 Saddle in forest beyond the knob, towards the summit. SI 5, 822358 Ridge from Pyke Hut to Mt. Barrington. 1 mile beyond Trig. SI 13 Ridge above Swamp Creek, Hollyford River, | mile in forest. SI 13 Boulders in Whangamoa River, IJ miles SW of school. S15, 847421 Collector J. Hector A. McKay M. Ongley G. Bishop H. W. Wellman J. B. Waterhouse H. W. Wellman H. W. Wellman B. L. Wood B. L Wood B. L. Wood J. B. Waterhouse B. L. Wood B. L. Wood J. B. Waterhouse H. J. Harrington A. R. Mutch H. W. Wellman A. R. Mutch J. B. Waterhouse A. R. Mutch J. B. Waterhouse J. B. Waterhouse M. Hall J. B. Waterhouse G. Gibson J. B. Waterhouse I. W. Keyes J. B. Waterhouse I. W. Keyes J. B. Waterhouse I. W. Keyes J. B. Waterhouse J. B. Waterhouse A. R. Mutch J. B. Waterhouse T. Grant Taylor J. B. WATERHOUSE: NEW ZEALAND SPECIES OF ATOMODESMA in REFERENCES BEYRiCH, E. 1864. liber eine Kohlenkalk-Fauna von Timor. Abli. dtsch. Akad. fViss. Berk 61-98. BRUCE, j. G. 1962. The geology of the Nelson City area. Trans, roy. Soc. N.Z. Geol. 1 (11), 157-81. DiCKiNS, J. M. 1961a. Permian pelecypods newly recorded from eastern Australia. Palaeontology, 119-30. \96\b. Eitrydesma and Peruvispira from the Dwyka beds of South Africa. Ibid. 4, 138-48. 1963. Permian pelecypods and gastropods from Western Australia. Bull. Bur. Min. Resourc. Aust. 63. DiENER, c. 1903. Permian fossils of the Central Himalayas. Palaeont. Indica. Ser. 15, I (5). KONiNCK, L. G. dc. 1877. Rccherches sur les fossiles paleozoiques de la Nouvelle-Galles du Sud (Australie). Mem. Soc. Sci. Liege, 2. Aiixapee, B. K., 9nHop, O. (LIKHAREV B. K. and einor, o.) 1941. K Bonpocy o Bospacxe Bcpxne- naACosoiicKiix OTAOHcenHii loro-BOCTOTHofi RacxH KoAbiMCKoro Gaccei'ina. JJokm. Anad. nayK CCCP, 31(2), 150-2. m‘coy, f. 1847. On the fossil botany and zoology of the rocks associated with the coal of Austraha. Ann. Mag. nat. Hist. Ser. 1. 20, 145-57, 226-36, 298-312. MARWICK, J. 1925. Upper Palaeozoic (Permian) fossils at Clinton. N.Z. J. Sci. Tech. B. 7, 362-4. 1934. The sequence of molluscan life in New Zealand. Proc. 5th Pacif. Sci. Congr. 947-60. 1935. Some new genera of the Myalinidae and Pteriidae of New Zealand. Trans, roy. Soc. N.Z. 65, 295-303. MUTCH, A. R. 1957. Facies and thickness of the Upper Palaeozoic and Triassic sediments of South- land. Ibid. 84, 499-511. NEWELL, N. D. 1942. Late Paleozoic pelecypods: Mytilacea. Publ. Univ. Kans., State geol. Surv. Kans. 10 (2). IlonoB, lO. H. (POPOV, u. N.) 1948. Hobbic npe/icxaBiixeAii h3 po^a Kolymia Licharew. JJokji. Axad. HayK CCCP, 61 (4), 697-700. ■ 1958. HcKOTopbie nepMCiaie neAeqnno4bi, racTpoybi ii aMMOHiixbi BepxoHHba. Marepuaim no veoji. u nojiesHUM HCKonaeMUM ceeepo-BoctoKa CCCP? 12, 137-50. REED, F. R. c. 1935. A new lamellibranch from the Upper Dwyka Beds of Southwest Africa. Trans, roy. Soc. S. Afr. 23, 161-3. 1944. Brachiopoda and Mollusca from the Productus Limestones of the Salt Range. Palaeont. Indica, 23 (2). TRECHMANN, c. T. 1917. The age of the Maitai Scrics of Ncw Zealand. Geo/. Mag. Dec. 6, 4, 53-64. WANNER, c. 1922. Die Gastropoden und Lamellibranchiaten der Dyas von Timor. Paldont. Timor, 11 (18). 1 940. Neue Permische Lamellibranchiaten von Timor. Geol. Exped. to the Lesser Siinda Islands, 2, 312-95. WATERHOUSE, J. B. 1958a. The occurrence of Atomodesma Beyrich in New Zealand. N.Z. J. geol. geophys. 1, 166-77. 1958/?. The age of the Takitimu Group of western Southland. Ibid. 1, 604-10. ■ 1959. Note on New Zealand species of Atomodesma Beyrich. Ibid. 2, 262-4. WOOD, B. L. 1956. The Geology of the Gore Subdivision. Bull. geol. Surv. N.Z., N.s. 53. J. B. WATERHOUSE N.Z. Geological Survey, Lower Hutt, New Zealand Manuscript received 19 February 1963 TWO AUSTRALIAN TERTIARY NEOLAMPADIDS, AND THE CLASSIFICATION OF CASSIDULOID ECHINOIDS by G. M. PHILIP Abstract. Two new neolampadids, Pisolampas concinna gen. et sp. nov. and Notolampas flosculus gen. et sp. nov., are described from the Tertiary of south-eastern Australia. The classification of cassiduloid echinoids is discussed, and it is suggested that the order Cassiduloida Claus should include three suborders, viz. Cassidulina S.S., Conoclypina Zittel, Neolampadina subord. nov. Although the Tertiary echinoids of south-eastern Australia are being revised in a series of papers elsewhere (Philip, 1963o, b), it is desirable to present at an early stage aspects of this work of more general significance. In this note two new neolampadid genera are proposed, the phytogeny of the family discussed, and suggested alterations made to the current classification of the cassiduloid echinoids. Up to the present no fossil neolampadids have been described. Living representatives (placed by Mortensen, 1948, in the family Neolampadidae of the order Cassiduloida), although rare, are cosmopolitan, widely distributed through tropical and subtropical seas. The absence of petals and the single or atrophied pores of the adapical ambulacra distinguish the group from other cassiduloids. SYSTEMATIC DESCRIPTIONS Genus pisolampas gen. nov. Generic name. From the Greek ttloos, pea; Aa^uTra?, torch. Type species. Pisolampas concinna sp. nov. Diagnosis. Subhemispherical neolampadids with compact apical system possessing three genital pores. Oculars small, often separated from the central part of the apical disk. Adapical ambulacra with pores rudimentary or lacking. Periproct supermarginal, at the upper end of a shallow groove. Floscelle moderately well developed with well- defined bourrelets, and expanded phyllodes with an occluded plate in each posterior phyllode. Remarks. Pisolampas is similar to Neolampas in the character of the compact apical system which has three genital pores. However, the separation of the oculars from the central part of the apical system is a feature not seen in Neolampas (although present in Tropholampas), and the position of the periproct, with its ventral groove, in particular distinguishes Pisolampas. The floscelle is better developed in Pisolampas, although similar occluded plates have been noted in the posterior phyllodes of some specimens of Neolampas rostellata (Agassiz, 1904, pi. 119, fig. 156). [Palaeontology, Vol. 6, Part 4, 1963, pp. 718-26, pi. 106-7.] G. M. PHILIP: TWO AUSTRALIAN TERTIARY NEOLAMPADIDS 719 Pisolampas conchma sp. nov. Plate 106, Plate 107, fig. 11; text-fig. 1 Echinobrissus concinnus Tate MS, museum label. Material. Holotype A.U.G.D. T269a, the best preserved of thirteen specimens, mounted syntypes of R. Tate’s MS species "Echinobrissus concinnus', in the Tate Collection, Geology Dept., University of Adelaide. A.U.G.D. FI 5748-9 and four other uncatalogued A.U.G.D. specimens. National Museum of Victoria P.19252-3. Locality and horizon. Aldinga, South Australia, St. Vincent Basin, Tortachilla Limestone, Upper Eocene. Description. Test small, usually oval in outline, sometimes with greatest width pos- teriorly; adapical surface usually subhemispherical, rarely with the posterior interam- bulacrum raised, so that the posterior part of the test is flattened; adoral surface flat, sunken around the peristome. Peristome anterior, elongate oval in shape. Periproct mounted high on the adapical surface, narrow, with an introverted rim, and a prominent groove running downward to the ambitus. Test covered with small, irregularly disposed, sunken crenulate perforate primary tubercles between which are small closely spaced granules. Apical system (text-fig. Id, e) slightly anterior, compact, with genital pores 1, 2, 4 developed. Genital pores show marked dimorphism, being, in some specimens, much larger than in others. Oculars very small, and in some specimens well separated from the genital disk (text-fig. \a, b). Usually three or four small, centrally located hydropores, but as many as twelve may be present. Ambulacra narrow adapically with small plates and rudimentary or atrophied pores, seen only rarely on some plates (text-fig. \b, e). Floscelle (text-fig. Ic) with moderately prominent, closely granulated bourrelets and expanded phyllodes with single pores and sunken peripodia, arranged in two irregular series, but with inner occluded plates in each posterior phyllode ; two buccal pores in each phyllode. Remarks. The enlarged genital pores of some specimens are seen in many living echinoids and are a characteristic of neolampadids (e.g. Agassiz, 1883, pi. 22, figs. 1, 13). They are due to sexual dimorphism, the specimens with large pores being females (Agassiz, op. cit.). It is reasonable to give a similar interpretation in these fossil forms. Whether or not the adapical ambulacral pores were absent cannot be decided posi- tively; very small pores (text-fig. IZ)) in some specimens suggests that their seeming absence in others could perhaps be a secondary effect due to fossilization. Genus notolampas gen. nov. Generic name. From the Greek voro?, south-wind; Xa(nrds, torch. Type species. Notolampas flosculus sp. nov. Diagnosis. Somewhat posteriorly elongated neolampadids, with compact apical system possessing three genital pores, and pierced oculars; adapical ambulacral plates with single pores continuing to oculars. Periproct submarginal, overhung by posterior inter- ambulacrum. Floscelle well developed, with prominent bourrelets and wide phyllodes with pores arranged in four irregular series in each. 720 PALAEONTOLOGY, VOLUME 6 Remarks. This combination of characters provides no basis for comparison with pre- viously described genera. In particular, the floscelle is much better developed than in other neolampadids. Notolampas flosculus sp. nov. Plate 107, figs. 1-10; text-fig. 2 Pygorhynchus Vassali Wright; Tate, 1891, Tram. Proc. roy. Soc. S.A. 14 (12), p. 275. {Non) Pygorhynchus Vassali Wright; Duncan, 1877, Quart. J. geol. Soc. Lond. 33, pp. 51, 67; Duncan, 1887, ibid. 43, p. 420. = Catopygus elegans Laube, 1869, Sitz. k. Akad. Wissen. Wien, 59, pp. 190-1, figs. 8-8c. = Studeria eiegans (Laube). {Nec) Pygorhynchus vassaii Wright; H. L. Clark, 1946, Pub. Carnegie Instn. 566, p. 358. = Studeria eiegans (Laube). {Nec) Pygorhynchus vassaii Wright; Ludbrook, 1961, Buii. Geoi. Surv. S.A. 36, p. 44. = Progonoiampas Novae-hoiiandiae Bittner, 1892, Sitz. k. Akad. Wissen. Wien, pp. 357-9, pi. 3, fig. 1. = Echinoiampas novaehoiiandiae (Bittner). {Nec) Pygorhynchus Vassaii Wright, 1864, Quart. J. geoi. Soc. Lond. 20, p. 479, pi. 22, fig. 6a-c. Materiai. Holotype A.U.G.D. F15747 {ex Mines Dept. S.A. Coll.). A.U.G.D. T277a-f, six mounted specimens in the Tate Collection. Locaiity and horizon. Mannum, South Australia, Murray Basin, Mannum Formation, Lower Miocene. Description. Test small, suboval in outline and somewhat pointed posteriorly, greatest width posteriorly, adapical surface gently arched to the greatest height posterior to the apical system. Adoral surface tumid, markedly sunken around the peristome. Peristome anterior, elongate oval in shape. Periproct submarginal, overhung adapically by the posterior interambulacrum, subquadrate in shape, with an introverted rim. Test covered with small, irregularly disposed, sunken primary tubercles and small granules. Apical system (text-fig. 2a) central, compact with genital pores 1, 2, 4 present. Oculars abutting against genital disk, and subtending narrow ambulacra with single pores (text- fig. 2a). Up to six hydropores present. Floscelle (text-fig. 2b) well developed, with prominent granulated bourrelets, and expanded phyllodes, possessing single pores arranged in four very irregular series. Adoral pores with sunken peripodia. Two buccal pores in each phyllode. Remarks. Small specimens of the Australian Miocene species Studeria eiegans (Laube) are similar in shape to Notolampas flosculus, but differ obviously in the presence of petals. On such a specimen (British Museum Nat. Hist. E2375) is based Duncan’s (1877, EXPLANATION OF PLATE 106 Figures X 2 unless otherwise stated. Figs. 1-10. Pisoiampas concinna gQV\. et sp. now. 1 , Adoral view. 2, Posterior view. 3, Anterior view. 4, Adapical view. 5, Left anterior view of holotype A.U.G.D. T269a (cJ). 6, Left lateral view. 7, Adapical view of P.19252 (?). 8, Enlargement of floscelle, P.19253, X 12. 9, Apical system of A.U.G.D. F15748 {S), showing numerous hydropores, X 12. 10, Adapical surface of A.U.G.D. T269b (cJ), showing apical system, sutures, and occasional ambulacral pores, X 12. All specimens from the Upper Eocene Tortachilla Limestone, Aldinga, South Australia. Palaeontology, Vol. 6 PLATE 106 9 iO PHILIP, Tertiary neolampids G. M. PHILIP: TWO AUSTRALIAN TERTIARY NEOLAMPADIDS 721 TEXT-FIG. 1. Pisolampas co«c//w« gen. et sp. nov. a, Adapical plating of A.U.G.D. T269c (^), x5. b, Adapical plating of P.19252 ($), X 5. e, Adoral plating of P.19253, X 5. d. Enlargement of apical system of P.19253 {$), X 10. e. Enlargement of apical system of A.U.G.D. F15749 (?), X 10. p. 51) record of Pygorhynchus vassali Wright from ‘East of the Glenelg River’, Victoria. This record, in turn, gave rise to Tate’s (1891) identification of the present species. P. vassali, currently placed in Pliolampas Pomel, has typical echinolampadid morpho- logy, with prominent ambulacra! petals. 722 PALAEONTOLOGY, VOLUME 6 The nature of the tubercles cannot be seen in the available material. As in PisoJampas coucinna, some specimens show enlarged genital pores, which is interpreted as sexual dimorphism. DISCUSSION Mortensen (1948) included the Neolampadidae in the suborder Cassidulina of the order Cassiduloida. Kier (1962), on the other hand, did not mention the neolampadids in his revision of the order. The two Tertiary genera described above, with their well- developed floscelles, clearly establish the cassiduloid ancestry and affinities of the family. a TEXT-FIG. 2. Notolampas flosculus gen. et sp. nov. a, Adapical plating of holotype, A.U.G.D. F15747 b, Adoral plating of same. Both X 5. The neolampadids are, however, strikingly dissimilar from other cassiduloids in their non-petalloid adapical abulacra, the plates of which possess only single pores. Indeed, not only are petals absent, but also the ambulacral pores (and tube feet) may be atro- phied, and the ambulacral columns and oculars even separated from the genitals (in Tropholampas and Pisolampas). The ambulacra thus are the most degenerate of all irregular echinoids. Because of this, a suborder is here proposed for the reception of the family. Stefanini (1913) has suggested that the neolampadid ambulacra were neotenously derived from the cassiduloid ambulacra, and this view must be fully endorsed. Because the earliest neolampadids possess a floscelle, it seems that the absence of this in some living genera must be regarded as further neotenous simplification of the ambulacra. Mortensen (1948, p. 331) has suggested that, if Aphanophora possessed phyllodes, it G. M. PHILIP: TWO AUSTRALIAN TERTIARY NEOLAMPADIDS 723 would ‘be the most advanced of the Neolampadidae in this respect’. Rather than this, it would now seem that floscelle-bearing neolampadids should be thought of as the least modified and least degenerate. The detailed derivation of the neolampadids is speculative. Because NamwJampas has disparate genital plates, it must be suggested that the family arose from forms such as the Nucleolitidae with a similar apical system. It must be noted that the Upper Eocene Pisolampas possesses advanced neolampadid characters suggesting that even older representatives of the family await discovery. A complete summary of the classification of the Cassiduloida demands the brief introduction of broader issues, which are discussed in more detail elsewhere (Philip, 1963c). Mortensen (1948) recognized two suborders of the Cassiduloida, viz. Cassidulina and Conoclypina. Durham and Melville (1957) removed the Conoclypina to the Holecty- poida, and divided the remaining cassiduloids into two orders, viz. Cassiduloida emend, and Nucleolitoida. Grouped together were forms with unequal pores in the petals and disparate apical systems, as opposed to forms with equal pores in the petals and compact apical systems. Kier (1962, p. 23) rightly rejected this twofold subdivision of the Cassi- dulina, observing that ‘As a matter of fact there are more genera with monobasal apical system and unequal pores than with equal pores’. As noted previously, Kier’s usage of the Cassiduloida excluded the neolampadids. Durham and Melville’s (1957) removal of the Conoclypina to the Holectypoida stems from a particular emphasis on the value of the lantern and girdle in echinoid classifica- tion. They divide the irregular echinoids into two superorders, viz. Gnathostomata (roughly corresponding to the holectypoids and clypeasteroids) and Atelostomata (cassi- duloids and spatangoids). This return to nineteenth-century classification ignores the findings of the last sixty years. It seems that, with the development of the irregular condition and the constriction of the peristome, the lantern became superfluous, and so must be considered as a truly regressive character, lost in different lines of descent. The adult holectypoid Echinoneus lacks a lantern and girdle, although these are present in early growth stages. A similar condition obtains in living cassiduloids such as Apatopygus and Echinolampas. Even among spatangoids MacBride (1918, p. 264) has described what he considered to be the rudiments of dental apparatus in the early larval stages of Echinocardiiim cordatum. Although this has been questioned (Mortensen, 1948, p. 6), the living meridosternous spatangoid Pilematechinus possesses ‘well developed auricles around the peristome, indicating the probable existence in young stages of dental apparatus’ (Mortensen, 1950, p. 116). These facts indicate that no clear-cut division into gnathostomatous and atelostomatous irregular echinoids is possible. It also follows that some so-called atelostomatous irregular echinoids, particularly the cassiduloids, are strictly gnathosto- matous. This clarified, the question of the systematic position of the Conoclypina can be discussed. The Tertiary Conoclypidae are virtually identical externally with some of the more advanced cassiduloids. Indeed, Kier (1957) gives as the distinctions between ConocJypus and Echinolampas the more slit-like outer pores in the petals, and the more longitudinally elongate peristome of the fomer. However, Conoclypus possesses a lantern 724 PALAEONTOLOGY, VOLUME 6 in adult stages, and hence Durham and Melville remove it to the largely gnatho- stomatous order Holectypoida. The features unknown in holectypoids which indicate the manifestly cassiduloid morphology of the test of Conoclypiis are : 1. The presence of petalloid ambulacra. 2. The presence of ‘pseudo-phyllodes’ and bourrelets around the peristome. 3. The disordered interambulacral ornament. 4. The absence of gill-slits, which are present in all holectypoids which have a lantern in adults. 5. The compact apical system. These characters indicate close similarity with advanced cassiduloids. This is also seen in other features such as the presence of single pores in the ambulacra outside the petals. Indeed, these, coupled with the compact apical system, are a combination of features found only in Cainozoic cassiduloids (with a few minor exceptions). As various cassiduloid groups must strictly be considered gnathostomatous, this evidence suggests that it is far more probable that the Conoclypina were a neotenous derivative of advanced cassiduloid stock (such as the Echinolampadidae, which are known to possess a girdle and lantern in young stages) than primary descendants of the holectypoids. This interpretation is supported by the fact that no satisfactory holecty- poid ancestor of the Conoclypina can be distinguished. The alternative view is to regard the Conoclypina as a sudden ‘ saltation ’ of the holectypoids to give a form of advanced cassiduloid morphology and the most remarkable example of homeomorphism in the class Echinoidea. As an outcome of these considerations the following classification of the echinoid order Cassiduloida is advocated. SUMMARY OF CLASSIFICATION Order cassiduloida Claus, 1880 emend. Diagnosis. Atelostomatous irregular echinoids (except Conoclypina) lacking gill-slits; ambulacra adapically petalloid (except Neolampadina) with petals usually all similar; floscelle usually well developed; apical system disparate or compact; plastron undif- ferentiated; fascioles absent. EXPLANATION OF PLATE 107 Figures x2 unless otherwise stated. Figs. 1-10. Notolampasflosciiliis gen. et sp. nov. 1 , Adapical view. 2, Adoral view. 3, Anterior view. 4, Left lateral view. 5, Posterior view of holotype, A.U.G.D. FI 5747 (cj). 6, Enlargement of floscelle of holotype, X 10. 7, Left lateral view. 8, Adapical view of A.U.G.D. T277a (?). 9, Adapi- cal view of A.U.G.D. T277b. 10, Enlargement of floscelle of A.U.G.D. T277c, X 10. All specimens from the Lower Miocene Mannum Formation, Mannum, South Australia. Fig. 1 1 . Pisolanipas conciiina gen. et sp. nov. Enlargement of perforate crenulate primary tubercles at the ambitus of A.U.G.D. T269b, x20. Erom the Upper Eocene Tortachilla Limestone, South Australia. Palaeontology, Vol. 6 9 PLATE 107 PHILIP, Tertiary neolampids G. M. PHILIP: TWO AUSTRALIAN TERTIARY NEOLAMPADIDS 725 1. Suborder cassidulina Claus emend. Diagnosis. Cassiduloids lacking a lantern in adult stages ; ambulacra adapically petalloid. Families. Apatopygidae Kier, 1962; Archiacidae Cotteau & Triger, 1869; Cassidulidae L. Agassiz & Desor, 1847; Clypeidae Lambert, 1898; Clypeolampadidae Kier, 1962; Echinolampadidae Gray, 1851; Faujasidae Lambert, 1905; Galeropygidae Lambert, 1911; Nucleolitidae L. Agassiz & Desor, 1847; Pliolampadidae Kier, 1962. Remarks. For present purposes, the families recognized by Kier (1962) in his restricted usage of the order Cassiduloida are listed as the complement of the Cassidulina. It is to be noted, however, that some of these are poorly differentiated (e.g. Nucleolitidae, Clypeidae; Echinolampadidae, Clypeolampadidae, Pliolampadidae) and probably should be considered at most as subfamilies. 2. Suborder conoclypina Zittel, 1 879 Diagnosis. Cassiduloids possessing a lantern in adult stages; ambulacra adapically petalloid. Families. Conoclypidae Zittel, 1879; ? Oligopygidae Duncan, 1891. Remarks. As a lantern-bearing cassiduloid, Bonaireaster Pijpers finds its place with the Conoclypidae. Whether or not the other genera {Oligopygus, Haimea, Mierolampas, Protolampas, and Ovulechinus), placed with Bonaireaster in the family Oligopygidae by Durham and Melville (1957), rightly belong together is uncertain. Durham and Mel- ville’s interpretation is followed to the extent that the family Oligopygidae is questionably referred to the Conoclypina. 3. Suborder neolampadina subord. nov. Type genus. Neolampas A. Agassiz, 1869. Diagnosis. Cassiduloids lacking a lantern in adult stages; petals lacking, and adapical ambulacral plates with single pores. Family. Neolampadidae Lambert, 1918. Genera. Anoehanus Grube, 1 868 ; Aplianapora de Meijere, 1 902 ; Nannolampas Mortensen, 1948; Neolampas A. Agassiz, 1869; Notolampas gen. nov.; Pisolampas gen. nov.; Tropholampas H. L. Clark, 1923. Acknowledgements. This work was initiated during the tenure of an Australian Commonwealth Scientific and Industrial Research Organization Overseas Studentship at the Sedgwick Museum, Cambridge, and has been continued with the aid of a University of New England Research Grant. I am obliged to Dr. M. F. Glaessner, University of Adelaide, Dr. N. H. Ludbrook, Mines Department of South Australia, and Mr. E. D. Gill, National Museum of Victoria, for the loan of specimens. REFERENCES AGASSIZ, A. 1883. Report on the results of dredging ... by the U.S. Coast Survey Steamer ‘Blake’. Report on the Echini. Mem. Mus. comp. Zool. 10 (1), 1-94, 32 pi. 1904. The Panamic deep sea Echini. Ibid. 31, 1-243, 110 pi. C 1713 3 B 726 PALAEONTOLOGY, VOLUME 6 BITTNER, A. 1892. Uber Echiniden dcs Tertiars von Australicn. Sitz. k. Akad. Wisseii. Wien (Math. Nat. CL), 101 (1), 331-71, pi. 1-4. CLARK, H. L. 1946. The echinoderm fauna of Australia. Its composition and its origin. Pub. Carnegie Instn. 566, 1-567. DUNCAN, p. M. 1877. On the Echinodermata of the Australian Cainozoic (Tertiary) deposits. Quart. J. geol. Soc. Land. 33 (1), 42-71, pi. 3-4. 1887. A revision of the Echinoidea from the Australian Tertiaries. Ibid. 43 (3), 41 1-30. DURHAM, J. w. and MELVILLE, R. V. 1957. A classification of echinoids. J. Paleont. 31, 242-12. KiER, p. M. 1957. Tertiary Echinoidea from British Somaliland. Ibid. 31, 839-902, pi. 103-7. 1962. Revision of the cassiduloid echinoids. Smithsonian Misc. Coll. 144 (3), 1-262, 44 pi. LAUBE, G. c. 1869. Uber einige fossile Echiniden von den Murray Cliffs in Siid-AustraUen. Sitz. k. Akad. Wissen. Wien (Math. Nat. CL), 59 (1), 183-98, 1 pi. LUDBROOK, N. H. 1961. Stratigraphy of the Murray Basin in South Australia. Bull. Geol. Surv. S.A. 37, 1-96, 8 pi. MACBRiDE, E. w. 1918. The development of Echinocardium cordatum. II, The development of the internal organs. Quart. J. Microsc. Sci. 63, 259-82. MORTENSEN, T. 1948. A Monograph of the Echinoidea, 4 (1), 1-363, 14 pi. Copenhagen. 1950. Idem. 5 (1), 1-432, 25 pi. PHILIP, G. M. 1963(7. The Tertiary echinoids of south-eastern Australia. I, Introduction and Cidaridae (1). Proc. roy. Soc. Viet. 76, 181-226, pi. 21-26. 1963fi. Idem., Cidaridae (2). Ibid. 77 (in press). ■ 1963c. The classification of echinoids. J. Paleont. 37 (in press). STEFANiNi, G. 1913. Probabile origine neotenica degfi ambulacri apetali di Neolampas. Padova Atti Acc. ven. treat. 3 (6), 33-41. TATE, R. 1891. A bibliography and revised list of the described echinoids of the Austrahan Eocene with descriptions of some new species. Trans. Proc. roy. Soc. S.A. 14 (2), 270-82. WRIGHT, T. 1864. On the fossil Echinoidea of Malta. Quart. J. geol. Soc. Load. IQ, 310-1, p. 22. G. M. PHILIP Department of Geology, University of New England, Armidale, N.S.W. Manuscript received 15 December 1962 BIOMETRIC STUDY OF BARREMITES SUBDIEEICILIS (KARAKASCH) by R. A. REYMENT and P. SANDBERG Abstract. A study is made of allometric growth in four dimensions of the ammonite Barremites subdifficilis (Karakasch) from the Neocomian of the Crimea. The four dimensions are further analysed simultaneously by the multivariate statistical method of principal component analysis. The results are compared with those found for two species of Triassic ammonites. It is often assumed by quantitative paleontologists that there will be an allometric growth relationship between many of the dimensions of the coiled ammonite shell. Obata (1959) investigated several species of Desmoceratidae and found almost isometric growth for most combinations of the variables studied but in a certain number of cases possibly significant positive and negative allometric growth. In the present analysis the variables studied are maximum diameter of shell (xi), maximum diameter of umbihcus (xg), height of last whorl (xg), breadth of last whorl (xd; all measurements are in millimetres. The measurements were converted to base 10 logarithm values for the bivariate and quadrivariate studies. The material was collected by Dr. V. V. Drushshits, Department of Paleontology, University of Moscow, who also checked the measurements. The specimens are stored in the University of Moscow; all (32) specimens were derived from the same bed in the Barremian sequence of the Crimea, U.S.S.R. The quadrivariate computations were performed on the computers of the Swedish Board for Computing Machinery, who provided free machine time. Part of the study was supported by a grant from the Swedish Natural Science Research Council. This study was completed during the junior author’s tenure of a fellowship from the Southern Fellowships Fund, Chapel Hill, North CaroUna, U.S.A. RELATIVE GROWTH ANALYSIS Standard least squares regression was used in the allometric analyses. The equations between pairs of variables are ; = 7-04x20-^6 (1) Xi = 2-43x3«-93 (2) Xi = 2-99X40-99 (3) = 0-39x3io^ (4) Xg = 0-45x/'i^ (5) X3 = T43X40-99 (6) The exponents of equations (2), (3), (4), and (6) differ so slightly from unity that there can be httle doubt that growth in these cases is isometric. Equations (1) and (5) are indicative of possible allometric growth. The 95 per cent, confidence intervals for the regression coefficients of these two equations were computed. For equation (1) it was found that 0-69 5 — , 2 c/5 Z CO ^ _ 1 LIBRARIES SMITHSONIAN INSTITUTION NOliniliSNI NVINOSHill/ — c/5 ~ CO — CO O ^ XiooiS3ix O pc^ _ O i I ES^SMITHSONIAN“'lNSTlTUTION NOliniliSNI “"NVINOSHilWS S3iyvyan”'LIBRARI E S^ SMITHSONIi 5 CO NviNosHii/js S3iyvyan l 2 CO 2 05 50 > X) m "S^’ z m CO — CO li BR ARIES SMITHSONIAN INSTITUTION NOlinillSNl NVINOSHill CO 2 V < A 2 < ES SMITHSONIAN INSTITUTION NOlinillSNl NVINOSHIIIMS S3iyvyai3 LIBRARIES SMITHSONW ISNI NviNosHims S3iyvyan libraries Smithsonian institution NoiiniiisNi nvinoshii^ T^/ — m CO . _ lES SMITHSONIAN INSTITUTION NOlinillSNl NVINOSHIIIMS S3iyVyail LIBRARIES SMITHSONL CO z , CO Z w 2 . CO 2 > '"W" 2 > * 2^00 ' CO 2 CO * 2 ISNI NVINOSHIIWS S3iyvyai1 libraries SMITHSONIAN INSTITUTION NOlinillSNl NVINOSHill — in “ CO — CO = O C_) -SC'" (_) Q z _j 2 _l 2 _l 2 HES SMITHSONIAN INSTITUTION NOlinillSNl NVINOSHIIINS S3iyvyai3 LIBRARIES SMITHSONI 2 [;; 2: ^,2 r- 2 CO t f; CO _ CO £ CO ISNI NviNOSHiiiMS S3iyvyan libraries Smithsonian institution NoiiniiiSNi nvinoshiii 2 CO 2»CO T'vCO 2 ' CD /y::, 70 > 70 m rn C/) £ CD TITUTION NOIlfUliSNl NVINOSHillNS S3iyVaail L rn ^ CD CD m CD *■■ 7Z CD BRARIES SMITHSONIAN INSTITUTION NOlif ^ Z ,v CD < V. s: X CD I ^ ~Z if) z CD avaan libraries Smithsonian institution NoiiniiiSNi nvinoshiiws saiavaan libr — CD ^ CD X .... CD ~ O _ _ ^ TITUTION NOIini!iSNl“'NVINOSHillNS S3iavaaiT LIBRARIES SMITHSONIAN~^INSTITUTION NOlif 2 I- , z r- 2 1“ z CD m ' X rn X^osv^^ y, m CD _ CD fz CD avaan libraries Smithsonian institution NoiiniiiSNi NVlN0SHil^MS saiavaan li Z CD 2 . CD 2 CD 2 (ife Jl) ‘if’ ‘i? 3<) ^ b r \, > 2 >' 5 xiwiiz >- 2 CD ' z if) ■'- 72. if) Z CD , riTUTION NOIlfUliSNl NVINOSHilWS S3iavaail LIBRARIES SMITHSONIAN INSTITUTION NOlif avaan libraries smithsonian institution NoiiniiiSNi nvinoshihns S3i avaan libr Z r- 2 O O TO > ^ ' r_ ' m ^ m P m CD — CD £ CD TITUTION NOlifliliSNI NVINOSHilWS S3iaVaaiT LIBRARIES SMITHSONIAN INSTITUTION NOlii Z CD 2 ,..■ CD 2 ,v CD CD I i I I I ^ I ^ *^ 2 CD Z CD *Z CD*^2 avaan libraries smithsonian institution NoiiniiiSNi nvinoshihais saiavaan libr X CD X CD 5 V C^ ,if) ^ ^ TITUTION ^NOliOiliSN I “"NVlNOSHillMS^S 3 iavaail”^LIBRARI ES^SMITHSONIAN~'lNSTITUTION ^ NOlif Z (- , Z r- 2 •“ z ..IBRARIES SMITHSONIAN INSTITUTION NOliUliSNI NVINOSHillAIS S3iavaail LIBR CD 2»CD 2 CD Z , < '^ 2 <- rs. S -ilfc-, < < X A z. y/^A/y.*.